2 |
3 | [](https://cran.r-project.org/package=interplot) 
4 |
5 | ------------------------------------------------------------------------
6 | interplot
7 | =========
8 |
9 | `interplot` is a tool for plotting the conditional coefficients ("marginal effects") of variables included in multiplicative interaction terms. The function plots the changes in the coefficient of one variable in a two-way interaction term conditional on the value of the other included variable. The plot also includes simulated 95% confidential intervals of these coefficients.
10 |
11 | To install:
12 |
13 | * the latest released version: `install.packages("interplot")`.
14 | * the latest developing version: `remotes::install_github("sammo3182/interplot")`.
15 |
16 |
17 | More details are available at:
18 |
19 | http://cran.r-project.org/web/packages/interplot/vignettes/interplot-vignette.html
20 |
21 |
22 | Please note that this project is released with a [Contributor Code of Conduct](https://github.com/sammo3182/interplot/blob/master/CONDUCT.md). By participating in this project you agree to abide by its terms.
23 |
--------------------------------------------------------------------------------
/CONDUCT.md:
--------------------------------------------------------------------------------
1 | # Contributor Code of Conduct
2 |
3 | As contributors and maintainers of this project, we pledge to respect all people who
4 | contribute through reporting issues, posting feature requests, updating documentation,
5 | submitting pull requests or patches, and other activities.
6 |
7 | We are committed to making participation in this project a harassment-free experience for
8 | everyone, regardless of level of experience, gender, gender identity and expression,
9 | sexual orientation, disability, personal appearance, body size, race, ethnicity, age, or religion.
10 |
11 | Examples of unacceptable behavior by participants include the use of sexual language or
12 | imagery, derogatory comments or personal attacks, trolling, public or private harassment,
13 | insults, or other unprofessional conduct.
14 |
15 | Project maintainers have the right and responsibility to remove, edit, or reject comments,
16 | commits, code, wiki edits, issues, and other contributions that are not aligned to this
17 | Code of Conduct. Project maintainers who do not follow the Code of Conduct may be removed
18 | from the project team.
19 |
20 | Instances of abusive, harassing, or otherwise unacceptable behavior may be reported by
21 | opening an issue or contacting one or more of the project maintainers.
22 |
23 | This Code of Conduct is adapted from the Contributor Covenant
24 | (http:contributor-covenant.org), version 1.0.0, available at
25 | http://contributor-covenant.org/version/1/0/0/
26 |
--------------------------------------------------------------------------------
/dev/runBeforeRelease.R:
--------------------------------------------------------------------------------
1 | # Check List for A New Release
2 |
3 | ## Vignette building
4 |
5 | devtools::install(build_vignettes = TRUE)
6 |
7 | ## Spell checking
8 |
9 | library(devtools)
10 | library(roxygen2)
11 |
12 | spell_check()
13 |
14 | rhub::validate_email(email = "yuehu@tsinghua.edu.cn", token = "345011f4ca60404abf76007e9ae89e3e")
15 |
16 | check_rhub(email = "yuehu@tsinghua.edu.cn")
17 |
18 |
19 | ## Run the CRAN check
20 | # 1. Under the `Build` tab
21 | # 1. Fixed all the errors and warnings
22 |
23 | ## Documentation
24 |
25 | # 1. Update the version number in `DESCRIPTION`
26 | # 1. Update `NEWS`
27 |
28 | ## Removing the htmls to release space
29 |
30 | library(here)
31 |
32 | ls_html <-
33 | list.files(here("inst/tutorials"),
34 | pattern = "*.html$",
35 | recursive = TRUE,
36 | full.names = TRUE)
37 |
38 | file.remove(ls_html)
39 |
40 | ## Release
41 |
42 | devtools::release()
43 |
44 | ## Change the version number in `DESCRIPTION` back to the dev version (i.e., "9999")
45 |
46 | ## Post the Release note when the package is on CRAN
47 |
48 |
49 | ## (No need if the package already has a website) Package website building
50 | library(devtools)
51 |
52 | usethis::git_vaccinate() #Adds .DS_Store, .Rproj.user, .Rdata, .Rhistory, and .httr-oauth to your global (a.k.a. user-level) .gitignore. This is good practice as it decreases the chance that you will accidentally leak credentials to GitHub.
53 |
54 | use_github_links()
55 | use_github_action_check_standard()
56 |
57 | library(pkgdown)
58 | build_site()
59 |
--------------------------------------------------------------------------------
/docs/bootstrap-toc.css:
--------------------------------------------------------------------------------
1 | /*!
2 | * Bootstrap Table of Contents v0.4.1 (http://afeld.github.io/bootstrap-toc/)
3 | * Copyright 2015 Aidan Feldman
4 | * Licensed under MIT (https://github.com/afeld/bootstrap-toc/blob/gh-pages/LICENSE.md) */
5 |
6 | /* modified from https://github.com/twbs/bootstrap/blob/94b4076dd2efba9af71f0b18d4ee4b163aa9e0dd/docs/assets/css/src/docs.css#L548-L601 */
7 |
8 | /* All levels of nav */
9 | nav[data-toggle='toc'] .nav > li > a {
10 | display: block;
11 | padding: 4px 20px;
12 | font-size: 13px;
13 | font-weight: 500;
14 | color: #767676;
15 | }
16 | nav[data-toggle='toc'] .nav > li > a:hover,
17 | nav[data-toggle='toc'] .nav > li > a:focus {
18 | padding-left: 19px;
19 | color: #563d7c;
20 | text-decoration: none;
21 | background-color: transparent;
22 | border-left: 1px solid #563d7c;
23 | }
24 | nav[data-toggle='toc'] .nav > .active > a,
25 | nav[data-toggle='toc'] .nav > .active:hover > a,
26 | nav[data-toggle='toc'] .nav > .active:focus > a {
27 | padding-left: 18px;
28 | font-weight: bold;
29 | color: #563d7c;
30 | background-color: transparent;
31 | border-left: 2px solid #563d7c;
32 | }
33 |
34 | /* Nav: second level (shown on .active) */
35 | nav[data-toggle='toc'] .nav .nav {
36 | display: none; /* Hide by default, but at >768px, show it */
37 | padding-bottom: 10px;
38 | }
39 | nav[data-toggle='toc'] .nav .nav > li > a {
40 | padding-top: 1px;
41 | padding-bottom: 1px;
42 | padding-left: 30px;
43 | font-size: 12px;
44 | font-weight: normal;
45 | }
46 | nav[data-toggle='toc'] .nav .nav > li > a:hover,
47 | nav[data-toggle='toc'] .nav .nav > li > a:focus {
48 | padding-left: 29px;
49 | }
50 | nav[data-toggle='toc'] .nav .nav > .active > a,
51 | nav[data-toggle='toc'] .nav .nav > .active:hover > a,
52 | nav[data-toggle='toc'] .nav .nav > .active:focus > a {
53 | padding-left: 28px;
54 | font-weight: 500;
55 | }
56 |
57 | /* from https://github.com/twbs/bootstrap/blob/e38f066d8c203c3e032da0ff23cd2d6098ee2dd6/docs/assets/css/src/docs.css#L631-L634 */
58 | nav[data-toggle='toc'] .nav > .active > ul {
59 | display: block;
60 | }
61 |
--------------------------------------------------------------------------------
/docs/docsearch.js:
--------------------------------------------------------------------------------
1 | $(function() {
2 |
3 | // register a handler to move the focus to the search bar
4 | // upon pressing shift + "/" (i.e. "?")
5 | $(document).on('keydown', function(e) {
6 | if (e.shiftKey && e.keyCode == 191) {
7 | e.preventDefault();
8 | $("#search-input").focus();
9 | }
10 | });
11 |
12 | $(document).ready(function() {
13 | // do keyword highlighting
14 | /* modified from https://jsfiddle.net/julmot/bL6bb5oo/ */
15 | var mark = function() {
16 |
17 | var referrer = document.URL ;
18 | var paramKey = "q" ;
19 |
20 | if (referrer.indexOf("?") !== -1) {
21 | var qs = referrer.substr(referrer.indexOf('?') + 1);
22 | var qs_noanchor = qs.split('#')[0];
23 | var qsa = qs_noanchor.split('&');
24 | var keyword = "";
25 |
26 | for (var i = 0; i < qsa.length; i++) {
27 | var currentParam = qsa[i].split('=');
28 |
29 | if (currentParam.length !== 2) {
30 | continue;
31 | }
32 |
33 | if (currentParam[0] == paramKey) {
34 | keyword = decodeURIComponent(currentParam[1].replace(/\+/g, "%20"));
35 | }
36 | }
37 |
38 | if (keyword !== "") {
39 | $(".contents").unmark({
40 | done: function() {
41 | $(".contents").mark(keyword);
42 | }
43 | });
44 | }
45 | }
46 | };
47 |
48 | mark();
49 | });
50 | });
51 |
52 | /* Search term highlighting ------------------------------*/
53 |
54 | function matchedWords(hit) {
55 | var words = [];
56 |
57 | var hierarchy = hit._highlightResult.hierarchy;
58 | // loop to fetch from lvl0, lvl1, etc.
59 | for (var idx in hierarchy) {
60 | words = words.concat(hierarchy[idx].matchedWords);
61 | }
62 |
63 | var content = hit._highlightResult.content;
64 | if (content) {
65 | words = words.concat(content.matchedWords);
66 | }
67 |
68 | // return unique words
69 | var words_uniq = [...new Set(words)];
70 | return words_uniq;
71 | }
72 |
73 | function updateHitURL(hit) {
74 |
75 | var words = matchedWords(hit);
76 | var url = "";
77 |
78 | if (hit.anchor) {
79 | url = hit.url_without_anchor + '?q=' + escape(words.join(" ")) + '#' + hit.anchor;
80 | } else {
81 | url = hit.url + '?q=' + escape(words.join(" "));
82 | }
83 |
84 | return url;
85 | }
86 |
--------------------------------------------------------------------------------
/dev/interplot_test.R:
--------------------------------------------------------------------------------
1 | library(mi)
2 | library(mitools)
3 | library(Amelia)
4 |
5 | p2005 <- read_sav("../Meritocracy_Rep/data/pew/merit/version_a/Dec05/Dec05c.sav")
6 | p2005x <- data.frame(
7 | resp = p2005$resp,
8 | fips = as.numeric(p2005$qfips),
9 | state = floor(as.numeric(p2005$qfips)/1000),
10 | rej_merit = ifelse(p2005$q14k<=2, 0, ifelse(p2005$q14k>4, NA, 1)),
11 | income = ifelse(p2005$income<=9, p2005$income, NA), # 1 to 9
12 | educ = ifelse(p2005$educ<=7, p2005$educ, NA), # 1 to 7
13 | age = ifelse(p2005$age<99, p2005$age, NA),
14 | male = ifelse(p2005$sex==1, 1, 0),
15 | white = ifelse(p2005$race==1 & p2005$hisp!=1, 1, 0),
16 | union = ifelse(p2005$labor<=3, 1, ifelse(p2005$labor==4, 0, NA)),
17 | ideo = 6 - ifelse(p2005$ideo<=5, p2005$ideo, NA), # 1 to 5
18 | attend = 7 - ifelse(p2005$attend<=6, p2005$attend, NA) # 1 to 6
19 | )
20 |
21 | m1 <- glm(rej_merit ~ income*educ + age, data = p2005x, family = binomial(link = "logit"))
22 | interplot(m1, "income", "educ")
23 |
24 | p2005x_mi <- missing_data.frame(p2005x)
25 | p2005x_mi <- change_type(p2005x_mi, y = "resp", to = "irrelevant")
26 | p2005x_mi <- change_type(p2005x_mi, y = "fips", to = "irrelevant")
27 | p2005x_mi <- change_type(p2005x_mi, y = "state", to = "irrelevant")
28 | p2005x_mi <- change_type(p2005x_mi, y = "income", to = "ordered-categorical")
29 | p2005x_mi <- change_type(p2005x_mi, y = "educ", to = "ordered-categorical")
30 | p2005x_mi <- change_type(p2005x_mi, y = "attend", to = "ordered-categorical")
31 |
32 | ptm <- proc.time()
33 | p2005x_im <- mi(p2005x_mi, n.chains=8, n.iter=30, seed=324, max.minutes=60)
34 | proc.time() - ptm
35 |
36 | m2_mi <- pool(formula = rej_merit ~ as.numeric(income) * as.numeric(educ) + age,
37 | family=binomial(link="logit"),
38 | data = p2005x_im)
39 | # class "pooled" should be implemented
40 |
41 | p2005x_mi_list <- complete(p2005x_im, m = 10)
42 | p2005x_mi_list2 <- p2005x_mi_list
43 | for (i in 1:10) {
44 | p2005x_mi_list2[[i]]$income <- as.numeric(p2005x_mi_list[[i]]$income)
45 | p2005x_mi_list2[[i]]$educ <- as.numeric(p2005x_mi_list[[i]]$educ)
46 | p2005x_mi_list2[[i]]$attend <- as.numeric(p2005x_mi_list[[i]]$attend)
47 | }
48 | p2005x_mi_list3 <- imputationList(p2005x_mi_list2)
49 |
50 |
51 | m2_mitools <- with(p2005x_mi_list3,
52 | glm(formula = rej_merit ~ income * educ + age,
53 | family=binomial(link="logit")))
54 | interplot(m2_mitools, "income", "educ")
55 |
56 | m3_mitools <- with(p2005x_mi_list3,
57 | glmer(formula = rej_merit~income * educ + age +
58 | (1+income|state), family=binomial(link="logit")))
59 | interplot(m3_mitools, "income", "educ")
60 |
61 |
62 | a <- amelia(p2005x, m = 5, ords = 4) # lgstc doesn't seem to work
63 | a2 <- imputationList(a$imputations)
64 |
65 | m4_amelia <- with(a2, glm(formula = rej_merit ~ income * educ + age,
66 | family=binomial(link="logit")))
67 | interplot(a4, "income", "educ")
--------------------------------------------------------------------------------
/docs/pkgdown.js:
--------------------------------------------------------------------------------
1 | /* http://gregfranko.com/blog/jquery-best-practices/ */
2 | (function($) {
3 | $(function() {
4 |
5 | $('.navbar-fixed-top').headroom();
6 |
7 | $('body').css('padding-top', $('.navbar').height() + 10);
8 | $(window).resize(function(){
9 | $('body').css('padding-top', $('.navbar').height() + 10);
10 | });
11 |
12 | $('[data-toggle="tooltip"]').tooltip();
13 |
14 | var cur_path = paths(location.pathname);
15 | var links = $("#navbar ul li a");
16 | var max_length = -1;
17 | var pos = -1;
18 | for (var i = 0; i < links.length; i++) {
19 | if (links[i].getAttribute("href") === "#")
20 | continue;
21 | // Ignore external links
22 | if (links[i].host !== location.host)
23 | continue;
24 |
25 | var nav_path = paths(links[i].pathname);
26 |
27 | var length = prefix_length(nav_path, cur_path);
28 | if (length > max_length) {
29 | max_length = length;
30 | pos = i;
31 | }
32 | }
33 |
34 | // Add class to parent
9 |
50 |
51 |
52 |
53 |
82 |
83 |
84 |
85 |
86 |
87 |
88 |
89 |
90 |
--------------------------------------------------------------------------------
/docs/articles/index.html:
--------------------------------------------------------------------------------
1 |
2 |
54 |
70 |
71 |
72 |
73 |
55 |
64 |
65 |
68 |
69 |
56 |
58 |
59 | License
57 |YEAR: 2015 60 | COPYRIGHT HOLDER: Frederick Solt, Yue HU 61 |62 | 63 |
9 |
50 |
51 |
52 |
53 |
79 |
80 |
81 |
82 |
83 |
84 |
85 |
86 |
87 |
--------------------------------------------------------------------------------
/docs/bootstrap-toc.js:
--------------------------------------------------------------------------------
1 | /*!
2 | * Bootstrap Table of Contents v0.4.1 (http://afeld.github.io/bootstrap-toc/)
3 | * Copyright 2015 Aidan Feldman
4 | * Licensed under MIT (https://github.com/afeld/bootstrap-toc/blob/gh-pages/LICENSE.md) */
5 | (function() {
6 | 'use strict';
7 |
8 | window.Toc = {
9 | helpers: {
10 | // return all matching elements in the set, or their descendants
11 | findOrFilter: function($el, selector) {
12 | // http://danielnouri.org/notes/2011/03/14/a-jquery-find-that-also-finds-the-root-element/
13 | // http://stackoverflow.com/a/12731439/358804
14 | var $descendants = $el.find(selector);
15 | return $el.filter(selector).add($descendants).filter(':not([data-toc-skip])');
16 | },
17 |
18 | generateUniqueIdBase: function(el) {
19 | var text = $(el).text();
20 | var anchor = text.trim().toLowerCase().replace(/[^A-Za-z0-9]+/g, '-');
21 | return anchor || el.tagName.toLowerCase();
22 | },
23 |
24 | generateUniqueId: function(el) {
25 | var anchorBase = this.generateUniqueIdBase(el);
26 | for (var i = 0; ; i++) {
27 | var anchor = anchorBase;
28 | if (i > 0) {
29 | // add suffix
30 | anchor += '-' + i;
31 | }
32 | // check if ID already exists
33 | if (!document.getElementById(anchor)) {
34 | return anchor;
35 | }
36 | }
37 | },
38 |
39 | generateAnchor: function(el) {
40 | if (el.id) {
41 | return el.id;
42 | } else {
43 | var anchor = this.generateUniqueId(el);
44 | el.id = anchor;
45 | return anchor;
46 | }
47 | },
48 |
49 | createNavList: function() {
50 | return $('');
51 | },
52 |
53 | createChildNavList: function($parent) {
54 | var $childList = this.createNavList();
55 | $parent.append($childList);
56 | return $childList;
57 | },
58 |
59 | generateNavEl: function(anchor, text) {
60 | var $a = $('');
61 | $a.attr('href', '#' + anchor);
62 | $a.text(text);
63 | var $li = $('
54 |
68 |
69 |
70 |
55 |
67 |
56 |
58 |
59 | Articles
57 |
60 |
66 | All vignettes
61 | 62 | 63 |`, then ``, etc.) that has more than one element. Defaults to 1 (for ``).
76 | getTopLevel: function($scope) {
77 | for (var i = 1; i <= 6; i++) {
78 | var $headings = this.findOrFilter($scope, 'h' + i);
79 | if ($headings.length > 1) {
80 | return i;
81 | }
82 | }
83 |
84 | return 1;
85 | },
86 |
87 | // returns the elements for the top level, and the next below it
88 | getHeadings: function($scope, topLevel) {
89 | var topSelector = 'h' + topLevel;
90 |
91 | var secondaryLevel = topLevel + 1;
92 | var secondarySelector = 'h' + secondaryLevel;
93 |
94 | return this.findOrFilter($scope, topSelector + ',' + secondarySelector);
95 | },
96 |
97 | getNavLevel: function(el) {
98 | return parseInt(el.tagName.charAt(1), 10);
99 | },
100 |
101 | populateNav: function($topContext, topLevel, $headings) {
102 | var $context = $topContext;
103 | var $prevNav;
104 |
105 | var helpers = this;
106 | $headings.each(function(i, el) {
107 | var $newNav = helpers.generateNavItem(el);
108 | var navLevel = helpers.getNavLevel(el);
109 |
110 | // determine the proper $context
111 | if (navLevel === topLevel) {
112 | // use top level
113 | $context = $topContext;
114 | } else if ($prevNav && $context === $topContext) {
115 | // create a new level of the tree and switch to it
116 | $context = helpers.createChildNavList($prevNav);
117 | } // else use the current $context
118 |
119 | $context.append($newNav);
120 |
121 | $prevNav = $newNav;
122 | });
123 | },
124 |
125 | parseOps: function(arg) {
126 | var opts;
127 | if (arg.jquery) {
128 | opts = {
129 | $nav: arg
130 | };
131 | } else {
132 | opts = arg;
133 | }
134 | opts.$scope = opts.$scope || $(document.body);
135 | return opts;
136 | }
137 | },
138 |
139 | // accepts a jQuery object, or an options object
140 | init: function(opts) {
141 | opts = this.helpers.parseOps(opts);
142 |
143 | // ensure that the data attribute is in place for styling
144 | opts.$nav.attr('data-toggle', 'toc');
145 |
146 | var $topContext = this.helpers.createChildNavList(opts.$nav);
147 | var topLevel = this.helpers.getTopLevel(opts.$scope);
148 | var $headings = this.helpers.getHeadings(opts.$scope, topLevel);
149 | this.helpers.populateNav($topContext, topLevel, $headings);
150 | }
151 | };
152 |
153 | $(function() {
154 | $('nav[data-toggle="toc"]').each(function(i, el) {
155 | var $nav = $(el);
156 | Toc.init($nav);
157 | });
158 | });
159 | })();
160 |
--------------------------------------------------------------------------------
/vignettes/vignette.bib.bak:
--------------------------------------------------------------------------------
1 | % Encoding: UTF-8
2 |
3 | @Article{Brambor2006,
4 | author = {Brambor, Thomas and Clark, William Roberts and Golder, Matt},
5 | title = {Understanding Interaction Models: Improving Empirical Analyses},
6 | journal = {Political Analysis},
7 | year = {2006},
8 | volume = {14},
9 | number = {1},
10 | pages = {63--82},
11 | __markedentry = {[Sammo:6]},
12 | publisher = {SPM-PMSAPSA}
13 | }
14 |
15 | @Book{Aiken1991,
16 | title = {Multiple regression: Testing and interpreting interactionsMultiple Regression: Testing and Interpreting Interactions},
17 | publisher = {Sage},
18 | year = {1991},
19 | author = {Aiken, Leona S and West, Stephen G and Reno, Raymond R}
20 | }
21 |
22 | @Article{WrightJr1976,
23 | author = {Wright Jr, Gerald C},
24 | title = {Linear Models for Evaluating Conditional Relationships},
25 | journal = {American Journal of Political Science},
26 | year = {1976},
27 | pages = {349--373},
28 | __markedentry = {[Sammo:]},
29 | publisher = {JSTOR}
30 | }
31 |
32 | @Article{Berry2012,
33 | author = {Berry, William D and Golder, Matt and Milton, Daniel},
34 | title = {Improving Tests of Theories Positing Interaction},
35 | journal = {Journal of Politics},
36 | year = {2012},
37 | volume = {74},
38 | number = {3},
39 | pages = {653--671},
40 | publisher = {Cambridge Univ Press}
41 | }
42 |
43 | @Unpublished{Hainmueller2016,
44 | author = {Hainmueller, Jens and Mummolo, Jonathan and Xu, Yiqing},
45 | title = {How Much Should We Trust Estimates from Multiplicative Interaction Models? Simple Tools to Improve Empirical Practice},
46 | year = {2016},
47 | timestamp = {2017-07-20},
48 | url = {https://ssrn.com/abstract=2739221},
49 | }
50 |
51 | @Article{Clark2006,
52 | author = {Clark, William Roberts and Gilligan, Michael J and Golder, Matt},
53 | title = {A Simple Multivariate Test for Asymmetric Hypotheses},
54 | journal = {Political Analysis},
55 | year = {2006},
56 | volume = {14},
57 | number = {3},
58 | pages = {311--331},
59 | publisher = {SPM-PMSAPSA}
60 | }
61 |
62 | @Article{Berry2016,
63 | author = {Berry, William D and DeMeritt, Jacqueline HR and Esarey, Justin},
64 | title = {Bias and Overconfidence in Parametric Models of Interactive Processes},
65 | journal = {American Journal of Political Science},
66 | year = {2016},
67 | volume = {60},
68 | number = {2},
69 | pages = {521--539},
70 | publisher = {Wiley Online Library},
71 | }
72 |
73 | @Unpublished{Esarey2017,
74 | author = {Esarey, Justin and Sumner, Jane Lawrence},
75 | title = {Marginal Effects in Interaction Models: Determining and Controlling the False Positive Rate},
76 | year = {2017},
77 | timestamp = {2017-07-20},
78 | url = {http://jee3.web.rice.edu/interaction-overconfidence.pdf},
79 | }
80 |
81 | @Article{Braumoeller2004,
82 | author = {Braumoeller, Bear F},
83 | title = {Hypothesis Testing and Multiplicative Interaction Terms},
84 | journal = {International Organization},
85 | year = {2004},
86 | volume = {58},
87 | number = {4},
88 | pages = {807--820},
89 | publisher = {Cambridge University Press},
90 | }
91 |
92 | @Article{Benjamini1995,
93 | author = {Benjamini, Yoav and Hochberg, Yosef},
94 | title = {Controlling the False Discovery Rate: A Practical and Powerful Approach to Multiple Testing},
95 | journal = {Journal of the Royal Statistical Society. Series B (Methodological)},
96 | year = {1995},
97 | pages = {289--300},
98 | publisher = {JSTOR},
99 | }
100 |
101 | @Article{Hanmer2013,
102 | author = {Hanmer, Michael J. and Ozan Kalkan, Kerem},
103 | title = {Behind the Curve: Clarifying the Best Approach to Calculating Predicted Probabilities and Marginal Effects from Limited Dependent Variable Models},
104 | journal = {American Journal of Political Science},
105 | year = {2013},
106 | volume = {57},
107 | number = {1},
108 | pages = {263--277},
109 | issn = {1540-5907},
110 | abstract = {Models designed for limited dependent variables are increasingly common in political science. Researchers estimating such models often give little attention to the coefficient estimates and instead focus on marginal effects, predicted probabilities, predicted counts, etc. Since the models are nonlinear, the estimated effects are sensitive to how one generates the predictions. The most common approach involves estimating the effect for the “average case.” But this approach creates a weaker connection between the results and the larger goals of the research enterprise and is thus less preferable than the observed-value approach. That is, rather than seeking to understand the effect for the average case, the goal is to obtain an estimate of the average effect in the population. In addition to the theoretical argument in favor of the observed-value approach, we illustrate via an empirical example and Monte Carlo simulations that the two approaches can produce substantively different results.},
111 | doi = {10.1111/j.1540-5907.2012.00602.x},
112 | publisher = {Blackwell Publishing Inc},
113 | timestamp = {2017-08-06},
114 | url = {http://dx.doi.org/10.1111/j.1540-5907.2012.00602.x},
115 | }
116 |
117 | @Comment{jabref-meta: databaseType:bibtex;}
118 |
--------------------------------------------------------------------------------
/docs/404.html:
--------------------------------------------------------------------------------
1 |
2 |
3 |
4 |
5 |
6 |
7 |
8 | Page not found (404) • interplot
9 |
10 |
11 |
12 |
13 |
14 |
15 |
19 |
20 |
21 |
22 |
23 |
24 |
73 |
74 |
75 |
76 |
77 |
78 |
79 |
80 | Page not found (404)
81 |
82 |
83 | Content not found. Please use links in the navbar.
84 |
85 |
86 |
87 |
91 |
92 |
93 |
94 |
95 |
96 |
107 |
108 |
109 |
110 |
111 |
112 |
113 |
114 |
115 |
116 |
--------------------------------------------------------------------------------
/man/interplot.lmerMod.Rd:
--------------------------------------------------------------------------------
1 | % Generated by roxygen2: do not edit by hand
2 | % Please edit documentation in R/Interplot_mlm.R
3 | \name{interplot.lmerMod}
4 | \alias{interplot.lmerMod}
5 | \title{Plot Conditional Coefficients in Mixed-Effects Models with Interaction Terms}
6 | \usage{
7 | \method{interplot}{lmerMod}(
8 | m,
9 | var1,
10 | var2,
11 | plot = TRUE,
12 | steps = NULL,
13 | ci = 0.95,
14 | adjCI = FALSE,
15 | hist = FALSE,
16 | var2_dt = NA,
17 | predPro = FALSE,
18 | var2_vals = NULL,
19 | point = FALSE,
20 | sims = 5000,
21 | xmin = NA,
22 | xmax = NA,
23 | ercolor = NA,
24 | esize = 0.5,
25 | ralpha = 0.5,
26 | rfill = "grey70",
27 | stats_cp = "none",
28 | txt_caption = NULL,
29 | facet_labs = NULL,
30 | ...
31 | )
32 | }
33 | \arguments{
34 | \item{m}{A model object including an interaction term, or, alternately, a data frame recording conditional coefficients.}
35 |
36 | \item{var1}{The name (as a string) of the variable of interest in the interaction term; its conditional coefficient estimates will be plotted.}
37 |
38 | \item{var2}{The name (as a string) of the other variable in the interaction term.}
39 |
40 | \item{plot}{A logical value indicating whether the output is a plot or a dataframe including the conditional coefficient estimates of var1, their upper and lower bounds, and the corresponding values of var2.}
41 |
42 | \item{steps}{Desired length of the sequence. A non-negative number, which for seq and seq.int will be rounded up if fractional. The default is 100 or the unique categories in the \code{var2} (when it is less than 100. Also see \code{\link{unique}}).}
43 |
44 | \item{ci}{A numeric value defining the confidence intervals. The default value is 95\% (0.95).}
45 |
46 | \item{adjCI}{Not working for `lmer` outputs yet.}
47 |
48 | \item{hist}{A logical value indicating if there is a histogram of `var2` added at the bottom of the conditional effect plot.}
49 |
50 | \item{var2_dt}{A numerical value indicating the frequency distribution of `var2`. It is only used when `hist == TRUE`. When the object is a model, the default is the distribution of `var2` of the model.}
51 |
52 | \item{predPro}{A logical value with default of `FALSE`. When the `m` is an object of class `glmerMod` and the argument is set to `TRUE`, the function will plot predicted probabilities at the values given by `var2_vals`.}
53 |
54 | \item{var2_vals}{A numerical value indicating the values the predicted probabilities are estimated, when `predPro` is `TRUE`.}
55 |
56 | \item{point}{A logical value determining the format of plot. By default, the function produces a line plot when var2 takes on ten or more distinct values and a point (dot-and-whisker) plot otherwise; option TRUE forces a point plot.}
57 |
58 | \item{sims}{Number of independent simulation draws used to calculate upper and lower bounds of coefficient estimates: lower values run faster; higher values produce smoother curves.}
59 |
60 | \item{xmin}{A numerical value indicating the minimum value shown of x shown in the graph. Rarely used.}
61 |
62 | \item{xmax}{A numerical value indicating the maximum value shown of x shown in the graph. Rarely used.}
63 |
64 | \item{ercolor}{A character value indicating the outline color of the whisker or ribbon.}
65 |
66 | \item{esize}{A numerical value indicating the size of the whisker or ribbon.}
67 |
68 | \item{ralpha}{A numerical value indicating the transparency of the ribbon.}
69 |
70 | \item{rfill}{A character value indicating the filling color of the ribbon.}
71 |
72 | \item{stats_cp}{A character value indicating what statistics to present as the plot note. Three options are available: "none", "ci", and "ks". The default is "none". See the Details for more information.}
73 |
74 | \item{txt_caption}{A character string to add a note for the plot, a value will sending to \code{ggplot2::labs(caption = txt_caption))}.}
75 |
76 | \item{facet_labs}{An optional character vector of facet labels to be used when plotting an interaction with a factor variable.}
77 |
78 | \item{...}{Other ggplot aesthetics arguments for points in the dot-whisker plot or lines in the line-ribbon plots. Not currently used.}
79 | }
80 | \value{
81 | The function returns a \code{ggplot} object.
82 | }
83 | \description{
84 | \code{interplot.mlm} is a method to calculate conditional coefficient estimates from the results of multilevel (mixed-effects) regression models with interaction terms.
85 | }
86 | \details{
87 | \code{interplot.mlm} is a S3 method from the \code{interplot}. It works on mixed-effects objects with class \code{lmerMod} and \code{glmerMod}.
88 |
89 | Because the output function is based on \code{\link[ggplot2]{ggplot}}, any additional arguments and layers supported by \code{ggplot2} can be added with the \code{+}.
90 |
91 | \code{interplot} visualizes the conditional effect based on simulated marginal effects. The simulation provides a probabilistic distribution of moderation effect of the conditioning variable (\code{var2}) at every preset values (including the minimum and maximum values) of the conditioned variable (\code{var1}), denoted as Emin and Emax. This output allows the function to further examine the conditional effect statistically in two ways. One is to examine if the distribution of \eqn{Emax - Emin} covers zero. The other is to directly compare Emin and Emax through statistical tools for distributional comparisons. Users can choose either method by setting the argument \code{stats_cp} to "ci" or "ks".
92 | \itemize{
93 | \item "ci" provides the confidence interval of the difference of \eqn{Emax - Emin}. An interval including 0 suggests no statistical difference before and after the conditional effect is applied, and vise versa.
94 | \item "ks" presents the result of a two-sample Kolmogorov-Smirnov test of the simulated distributions of Emin and Emax. The output includes a D statistics and a p-value of the null hypothesis that the two distributions come from the same distribution at the 0.05 level.
95 | }
96 |
97 | See an illustration in the package vignette.
98 | }
99 |
--------------------------------------------------------------------------------
/vignettes/interplot-vignette.R:
--------------------------------------------------------------------------------
1 | ## ----setup, include=FALSE-----------------------------------------------------
2 | knitr::opts_chunk$set(
3 | message = FALSE,
4 | warning = FALSE,
5 | dpi = 300
6 | )
7 |
8 | ## -----------------------------------------------------------------------------
9 | data(mtcars) #load the data
10 |
11 | ## -----------------------------------------------------------------------------
12 | m_cyl <- lm(mpg ~ wt * cyl, data = mtcars)
13 | summary(m_cyl)
14 |
15 | ## ----out.width="80%"----------------------------------------------------------
16 | library(interplot)
17 |
18 | interplot(m = m_cyl, var1 = "cyl", var2 = "wt")
19 |
20 | ## ----out.width="80%"----------------------------------------------------------
21 | interplot(m = m_cyl, var1 = "wt", var2 = "cyl")
22 |
23 | ## ----out.width="80%"----------------------------------------------------------
24 | interplot(m = m_cyl, var1 = "wt", var2 = "cyl", ci = .9, point = T)
25 |
26 | ## ----out.width="80%"----------------------------------------------------------
27 | interplot(m = m_cyl, var1 = "cyl", var2 = "wt", point = T) +
28 | # changing the angle of x labels for a clearer vision
29 | theme(axis.text.x = element_text(angle=90))
30 |
31 | ## ----out.width="80%"----------------------------------------------------------
32 | interplot(m = m_cyl, var1 = "cyl", var2 = "wt") +
33 | # Add labels for X and Y axes
34 | xlab("Automobile Weight (thousands lbs)") +
35 | ylab("Estimated Coefficient for\nNumber of Cylinders") +
36 | # Change the background
37 | theme_bw() +
38 | # Add the title
39 | ggtitle("Estimated Coefficient of Engine Cylinders \non Mileage by Automobile Weight") +
40 | theme(plot.title = element_text(face="bold")) +
41 | # Add a horizontal line at y = 0
42 | geom_hline(yintercept = 0, linetype = "dashed")
43 |
44 | ## ----out.width="80%"----------------------------------------------------------
45 | interplot(m = m_cyl, var1 = "wt", var2 = "cyl", ercolor = "blue", esize = 1.5) +
46 | geom_point(size = 2, color = "red")
47 |
48 | ## ----out.width="80%"----------------------------------------------------------
49 | m_wt <- lm(mpg ~ wt + I(wt^2), data = mtcars)
50 |
51 | interplot(m = m_wt, var1 = "wt", var2 = "wt")
52 |
53 | ## ----out.width="80%"----------------------------------------------------------
54 | mtcars$gear <- factor(mtcars$gear)
55 | m_gear <- lm(mpg ~ gear * wt, data = mtcars)
56 |
57 | interplot(m = m_gear, var1 = "wt", var2 = "gear")
58 |
59 | ## ----out.width="80%"----------------------------------------------------------
60 | interplot(m = m_gear, var1 = "wt", var2 = "gear", facet_labs = c("4-speed", "5-speed"))
61 |
62 | ## ----out.width="80%"----------------------------------------------------------
63 | interplot(m = m_cyl, var1 = "cyl", var2 = "wt", hist = TRUE) +
64 | geom_hline(yintercept = 0, linetype = "dashed")
65 |
66 | ## ----out.width="80%"----------------------------------------------------------
67 | interplot(m = m_cyl, var1 = "cyl", var2 = "wt", hist = TRUE) +
68 | aes(color = "pink") + theme(legend.position="none") + # geom_line(color = "pink") +
69 | geom_hline(yintercept = 0, linetype = "dashed")
70 |
71 | ## -----------------------------------------------------------------------------
72 | set.seed(313)
73 |
74 | interplot(m = m_cyl, var1 = "cyl", var2 = "wt", stats_cp = "ci")
75 |
76 | ## -----------------------------------------------------------------------------
77 | set.seed(313)
78 |
79 | interplot(m = m_cyl, var1 = "cyl", var2 = "wt", stats_cp = "ci",
80 | txt_caption = "\n Source: Motor Trend Car Road Tests 1973.")
81 |
82 | ## ----out.width="80%"----------------------------------------------------------
83 | stdCI_plot <- interplot(m = m_wt, var1 = "wt", var2 = "wt", adjCI = FALSE) +
84 | ggtitle("Marginal Effects with Standard CIs")
85 | adjCI_plot <- interplot(m = m_wt, var1 = "wt", var2 = "wt", adjCI = TRUE) +
86 | ggtitle("Marginal Effects with Adjusted CIs")
87 |
88 | library(gridExtra)
89 | grid.arrange(stdCI_plot, adjCI_plot, ncol = 2)
90 |
91 | ## ----out.width="80%"----------------------------------------------------------
92 | pew1.w <- read.csv("pew1_w.csv")
93 |
94 | m <- glm(formula=meritocracy~ginicnty+income_i+ginicnty:income_i+income_cnty+black_cnty+
95 | perc_bush04+pop_cnty+educ_i+age_i+gender_i+unemp_i+union_i+partyid_i+
96 | ideo_i+attend_i+survid2006+survid2007+survid2009,
97 | data=pew1.w,family=binomial(link="logit"))
98 |
99 | plot_avg <- interplot(m, var1 = "ginicnty",var2 = "income_i", predPro = FALSE) +
100 | ggtitle("Average Conditional Effects")
101 |
102 | plot_3val <- interplot(m, var1 = "ginicnty",var2 = "income_i", predPro = TRUE, var2_vals = c(min(pew1.w$income_i), max(pew1.w$income_i))) +
103 | ggtitle("Conditional Predicted Probabilities for \nCitizens with Low and High Incomes") +
104 | scale_colour_discrete(guide = guide_legend(title = "Income"), labels = c("Low", "High")) +
105 | scale_fill_discrete(guide = guide_legend(title = "Income"), labels = c("Low", "High")) +
106 | theme(legend.position = c(0, .8), legend.justification = c(0, .5))
107 |
108 | grid.arrange(plot_avg, plot_3val, ncol = 2)
109 |
110 | ## ----out.width="80%"----------------------------------------------------------
111 | # Create a fake dataset of conditional effects
112 | fake <- rnorm(100, 0, 1)
113 | coef1 <- fake * sample(.5:2.5, 100, replace = T)
114 | lb <- coef1 - .5
115 | ub <- coef1 + .5
116 |
117 | df_fake <- data.frame(cbind(fake, coef1, lb, ub))
118 |
119 | # Use interplot directly with the dataset
120 | interplot(df_fake)
121 |
122 | ## ----out.width="80%"----------------------------------------------------------
123 | var2_fake <- fake
124 | # Set `hist` to TRUE is required to superimpose a histogram.
125 | interplot(df_fake, hist = TRUE, var2_dt = var2_fake)
126 | ## The ribbon and histogram do not fit. This is just an illustration
127 |
128 |
--------------------------------------------------------------------------------
/man/interplot.mlmmi.Rd:
--------------------------------------------------------------------------------
1 | % Generated by roxygen2: do not edit by hand
2 | % Please edit documentation in R/Interplot_mlmmi.R
3 | \name{interplot.mlmmi}
4 | \alias{interplot.mlmmi}
5 | \title{Plot Conditional Coefficients in Mixed-Effects Models with Imputed Data and Interaction Terms}
6 | \usage{
7 | \method{interplot}{mlmmi}(
8 | m,
9 | var1,
10 | var2,
11 | plot = TRUE,
12 | steps = NULL,
13 | ci = 0.95,
14 | adjCI = FALSE,
15 | hist = FALSE,
16 | var2_dt = NA,
17 | predPro = FALSE,
18 | var2_vals = NULL,
19 | point = FALSE,
20 | sims = 5000,
21 | xmin = NA,
22 | xmax = NA,
23 | ercolor = NA,
24 | esize = 0.5,
25 | ralpha = 0.5,
26 | rfill = "grey70",
27 | stats_cp = "none",
28 | txt_caption = NULL,
29 | facet_labs = NULL,
30 | ...
31 | )
32 | }
33 | \arguments{
34 | \item{m}{A model object including an interaction term, or, alternately, a data frame recording conditional coefficients.}
35 |
36 | \item{var1}{The name (as a string) of the variable of interest in the interaction term; its conditional coefficient estimates will be plotted.}
37 |
38 | \item{var2}{The name (as a string) of the other variable in the interaction term.}
39 |
40 | \item{plot}{A logical value indicating whether the output is a plot or a dataframe including the conditional coefficient estimates of var1, their upper and lower bounds, and the corresponding values of var2.}
41 |
42 | \item{steps}{Desired length of the sequence. A non-negative number, which for seq and seq.int will be rounded up if fractional. The default is 100 or the unique categories in the \code{var2} (when it is less than 100. Also see \code{\link{unique}}).}
43 |
44 | \item{ci}{A numeric value defining the confidence intervals. The default value is 95\% (0.95).}
45 |
46 | \item{adjCI}{Not working for `lmer` outputs yet.}
47 |
48 | \item{hist}{A logical value indicating if there is a histogram of `var2` added at the bottom of the conditional effect plot.}
49 |
50 | \item{var2_dt}{A numerical value indicating the frequency distribution of `var2`. It is only used when `hist == TRUE`. When the object is a model, the default is the distribution of `var2` of the model.}
51 |
52 | \item{predPro}{A logical value with default of `FALSE`. When the `m` is an object of class `glmerMod` and the argument is set to `TRUE`, the function will plot predicted probabilities at the values given by `var2_vals`.}
53 |
54 | \item{var2_vals}{A numerical value indicating the values the predicted probabilities are estimated, when `predPro` is `TRUE`.}
55 |
56 | \item{point}{A logical value determining the format of plot. By default, the function produces a line plot when var2 takes on ten or more distinct values and a point (dot-and-whisker) plot otherwise; option TRUE forces a point plot.}
57 |
58 | \item{sims}{Number of independent simulation draws used to calculate upper and lower bounds of coefficient estimates: lower values run faster; higher values produce smoother curves.}
59 |
60 | \item{xmin}{A numerical value indicating the minimum value shown of x shown in the graph. Rarely used.}
61 |
62 | \item{xmax}{A numerical value indicating the maximum value shown of x shown in the graph. Rarely used.}
63 |
64 | \item{ercolor}{A character value indicating the outline color of the whisker or ribbon.}
65 |
66 | \item{esize}{A numerical value indicating the size of the whisker or ribbon.}
67 |
68 | \item{ralpha}{A numerical value indicating the transparency of the ribbon.}
69 |
70 | \item{rfill}{A character value indicating the filling color of the ribbon.}
71 |
72 | \item{stats_cp}{A character value indicating what statistics to present as the plot note. Three options are available: "none", "ci", and "ks". The default is "none". See the Details for more information.}
73 |
74 | \item{txt_caption}{A character string to add a note for the plot, a value will sending to \code{ggplot2::labs(caption = txt_caption))}.}
75 |
76 | \item{facet_labs}{An optional character vector of facet labels to be used when plotting an interaction with a factor variable.}
77 |
78 | \item{...}{Other ggplot aesthetics arguments for points in the dot-whisker plot or lines in the line-ribbon plots. Not currently used.}
79 | }
80 | \value{
81 | The function returns a \code{ggplot} object.
82 | }
83 | \description{
84 | \code{interplot.mlmmi} is a method to calculate conditional coefficient estimates from the results of multilevel (mixed-effects) regression models with interaction terms and multiply imputed data.
85 | }
86 | \details{
87 | \code{interplot.mlmmi} and \code{interplot.gmlmmi} are S3 methods from the \code{interplot}. It works on lists of mixed-effects objects with class \code{lmerMod} and \code{glmerMod} generated by \code{mitools} and \code{lme4}.
88 |
89 | Because the output function is based on \code{\link[ggplot2]{ggplot}}, any additional arguments and layers supported by \code{ggplot2} can be added with the \code{+}.
90 |
91 | \code{interplot} visualizes the conditional effect based on simulated marginal effects. The simulation provides a probabilistic distribution of moderation effect of the conditioning variable (\code{var2}) at every preset values (including the minimum and maximum values) of the conditioned variable (\code{var1}), denoted as Emin and Emax. This output allows the function to further examine the conditional effect statistically in two ways. One is to examine if the distribution of \eqn{Emax - Emin} covers zero. The other is to directly compare Emin and Emax through statistical tools for distributional comparisons. Users can choose either method by setting the argument \code{stats_cp} to "ci" or "ks".
92 | \itemize{
93 | \item "ci" provides the confidence interval of the difference of \eqn{Emax - Emin}. An interval including 0 suggests no statistical difference before and after the conditional effect is applied, and vise versa.
94 | \item "ks" presents the result of a two-sample Kolmogorov-Smirnov test of the simulated distributions of Emin and Emax. The output includes a D statistics and a p-value of the null hypothesis that the two distributions come from the same distribution at the 0.05 level.
95 | }
96 |
97 | See an illustration in the package vignette.
98 | }
99 |
--------------------------------------------------------------------------------
/docs/authors.html:
--------------------------------------------------------------------------------
1 |
2 | Authors and Citation • interplot
6 |
7 |
8 |
108 |
109 |
110 |
111 |
112 |
113 |
114 |
115 |
116 |
--------------------------------------------------------------------------------
/man/interplot.plot.Rd:
--------------------------------------------------------------------------------
1 | % Generated by roxygen2: do not edit by hand
2 | % Please edit documentation in R/Interplot_plot.R
3 | \name{interplot.plot}
4 | \alias{interplot.plot}
5 | \title{Plot Conditional Coefficients in Models with Interaction Terms}
6 | \usage{
7 | \method{interplot}{plot}(
8 | m,
9 | var1 = NULL,
10 | var2 = NULL,
11 | plot = TRUE,
12 | steps = NULL,
13 | ci = 0.95,
14 | adjCI = FALSE,
15 | hist = FALSE,
16 | var2_dt = NULL,
17 | predPro = FALSE,
18 | var2_vals = NULL,
19 | point = FALSE,
20 | sims = 5000,
21 | xmin = NA,
22 | xmax = NA,
23 | ercolor = NA,
24 | esize = 0.5,
25 | ralpha = 0.5,
26 | rfill = "grey70",
27 | stats_cp = "none",
28 | txt_caption = NULL,
29 | ci_diff = NULL,
30 | ks_diff = NULL,
31 | ...
32 | )
33 | }
34 | \arguments{
35 | \item{m}{A model object including an interaction term, or, alternately, a data frame recording conditional coefficients. This data frame should includes four columns:
36 | \itemize{
37 | \item fake: The sequence of \code{var1} (the item whose effect will be conditioned on in the interaction);
38 | \item coef1: The point estimates of the coefficient of \code{var1} at each break point.
39 | \item ub: The upper bound of the simulated 95\% CI.
40 | \item lb: The lower bound of the simulated 95\% CI.
41 | }}
42 |
43 | \item{var1}{The name (as a string) of the variable of interest in the interaction term; its conditional coefficient estimates will be plotted.}
44 |
45 | \item{var2}{The name (as a string) of the other variable in the interaction term.}
46 |
47 | \item{plot}{A logical value indicating whether the output is a plot or a dataframe including the conditional coefficient estimates of var1, their upper and lower bounds, and the corresponding values of var2.}
48 |
49 | \item{steps}{Desired length of the sequence. A non-negative number, which for seq and seq.int will be rounded up if fractional. The default is 100 or the unique categories in the \code{var2} (when it is less than 100. Also see \code{\link{unique}}).}
50 |
51 | \item{ci}{is a numeric value inherited from the data wrangling functions in this package. Adding it here is just for the method consistency.}
52 |
53 | \item{adjCI}{Succeeded from the data management functions in `interplot` package.}
54 |
55 | \item{hist}{A logical value indicating if there is a histogram of `var2` added at the bottom of the conditional effect plot.}
56 |
57 | \item{var2_dt}{A numerical value indicating the frequency distribution of `var2`. It is only used when `hist == TRUE`. When the object is a model, the default is the distribution of `var2` of the model.}
58 |
59 | \item{predPro}{A logical value with default of `FALSE`. When the `m` is an output of a general linear model (class `glm` or `glmerMod`) and the argument is set to `TRUE`, the function will plot predicted probabilities at the values given by `var2_vals`.}
60 |
61 | \item{var2_vals}{A numerical value indicating the values the predicted probabilities are estimated, when `predPro` is `TRUE`.}
62 |
63 | \item{point}{A logical value determining the format of plot. By default, the function produces a line plot when var2 takes on ten or more distinct values and a point (dot-and-whisker) plot otherwise; option TRUE forces a point plot.}
64 |
65 | \item{sims}{Number of independent simulation draws used to calculate upper and lower bounds of coefficient estimates: lower values run faster; higher values produce smoother curves.}
66 |
67 | \item{xmin}{A numerical value indicating the minimum value shown of x shown in the graph. Rarely used.}
68 |
69 | \item{xmax}{A numerical value indicating the maximum value shown of x shown in the graph. Rarely used.}
70 |
71 | \item{ercolor}{A character value indicating the outline color of the whisker or ribbon.}
72 |
73 | \item{esize}{A numerical value indicating the size of the whisker or ribbon.}
74 |
75 | \item{ralpha}{A numerical value indicating the transparency of the ribbon.}
76 |
77 | \item{rfill}{A character value indicating the filling color of the ribbon.}
78 |
79 | \item{stats_cp}{A character value indicating what statistics to present as the plot note. Three options are available: "none", "ci", and "ks". The default is "none". See the Details for more information.}
80 |
81 | \item{txt_caption}{A character string to add a note for the plot, a value will sending to \code{ggplot2::labs(caption = txt_caption))}.}
82 |
83 | \item{ci_diff}{A numerical vector with a pair of values indicating the confidence intervals of the difference between \code{var1} and \code{var2}.}
84 |
85 | \item{ks_diff}{A \code{ks.test} object of the effect of \code{var1} conditioned on \code{var2}.}
86 |
87 | \item{...}{Other ggplot aesthetics arguments for points in the dot-whisker plot or lines in the line-ribbon plots. Not currently used.}
88 | }
89 | \value{
90 | The function returns a \code{ggplot} object.
91 | }
92 | \description{
93 | Graph based on the data frame of statistics about the conditional effect of an interaction.
94 | }
95 | \details{
96 | \code{interplot.plot} is a S3 method from the \code{interplot}. It generates plots of conditional coefficients.
97 |
98 | Because the output function is based on \code{\link[ggplot2]{ggplot}}, any additional arguments and layers supported by \code{ggplot2} can be added with the \code{+}.
99 |
100 | \code{interplot} visualizes the conditional effect based on simulated marginal effects. The simulation provides a probabilistic distribution of moderation effect of the conditioning variable (\code{var2}) at every preset values (including the minimum and maximum values) of the conditioned variable (\code{var1}), denoted as Emin and Emax. This output allows the function to further examine the conditional effect statistically in two ways. One is to examine if the distribution of \eqn{Emax - Emin} covers zero. The other is to directly compare Emin and Emax through statistical tools for distributional comparisons. Users can choose either method by setting the argument \code{stats_cp} to "ci" or "ks".
101 | \itemize{
102 | \item "ci" provides the confidence interval of the difference of \eqn{Emax - Emin}. An interval including 0 suggests no statistical difference before and after the conditional effect is applied, and vise versa.
103 | \item "ks" presents the result of a two-sample Kolmogorov-Smirnov test of the simulated distributions of Emin and Emax. The output includes a D statistics and a p-value of the null hypothesis that the two distributions come from the same distribution at the 0.05 level.
104 | }
105 |
106 | See an illustration in the package vignette.
107 | }
108 |
--------------------------------------------------------------------------------
/docs/CONDUCT.html:
--------------------------------------------------------------------------------
1 |
2 | Contributor Code of Conduct • interplot
6 |
7 |
8 |
9 |
50 |
51 |
52 |
53 |
54 |
55 |
56 | Contributor Code of Conduct
57 |
58 |
59 |
60 |
61 | As contributors and maintainers of this project, we pledge to respect all people who contribute through reporting issues, posting feature requests, updating documentation, submitting pull requests or patches, and other activities.
62 | We are committed to making participation in this project a harassment-free experience for everyone, regardless of level of experience, gender, gender identity and expression, sexual orientation, disability, personal appearance, body size, race, ethnicity, age, or religion.
63 | Examples of unacceptable behavior by participants include the use of sexual language or imagery, derogatory comments or personal attacks, trolling, public or private harassment, insults, or other unprofessional conduct.
64 | Project maintainers have the right and responsibility to remove, edit, or reject comments, commits, code, wiki edits, issues, and other contributions that are not aligned to this Code of Conduct. Project maintainers who do not follow the Code of Conduct may be removed from the project team.
65 | Instances of abusive, harassing, or otherwise unacceptable behavior may be reported by opening an issue or contacting one or more of the project maintainers.
66 | This Code of Conduct is adapted from the Contributor Covenant (http:contributor-covenant.org), version 1.0.0, available at http://contributor-covenant.org/version/1/0/0/
67 |
68 |
69 |
70 |
71 |
74 |
75 |
76 |
77 |
78 |
79 |
88 |
89 |
90 |
91 |
92 |
93 |
94 |
95 |
96 |
--------------------------------------------------------------------------------
/man/interplot.default.Rd:
--------------------------------------------------------------------------------
1 | % Generated by roxygen2: do not edit by hand
2 | % Please edit documentation in R/Interplot_default.R
3 | \name{interplot.default}
4 | \alias{interplot.default}
5 | \title{Plot Conditional Coefficients in (Generalized) Linear Models with Interaction Terms}
6 | \source{
7 | Benjamini, Yoav, and Yosef Hochberg. 1995. "Controlling the False
8 | Discovery Rate: A Practical and Powerful Approach to Multiple Testing".
9 | Journal of the Royal Statistical Society, Series B 57(1): 289--300.
10 |
11 | Brambor, Thomas, William Roberts Clark, and Matt Golder.
12 | "Understanding interaction models: Improving empirical analyses". Political
13 | Analysis 14.1 (2006): 63-82.
14 |
15 | Esarey, Justin, and Jane Lawrence Sumner. 2015. "Marginal Effects in
16 | Interaction Models: Determining and Controlling the False Positive Rate".
17 | URL: \url{http://jee3.web.rice.edu/interaction-overconfidence.pdf}.
18 | }
19 | \usage{
20 | \method{interplot}{default}(
21 | m,
22 | var1,
23 | var2,
24 | plot = TRUE,
25 | steps = NULL,
26 | ci = 0.95,
27 | adjCI = FALSE,
28 | hist = FALSE,
29 | var2_dt = NA,
30 | predPro = FALSE,
31 | var2_vals = NULL,
32 | point = FALSE,
33 | sims = 1000,
34 | xmin = NA,
35 | xmax = NA,
36 | ercolor = NA,
37 | esize = 0.5,
38 | ralpha = 0.5,
39 | rfill = "grey70",
40 | stats_cp = "none",
41 | txt_caption = NULL,
42 | facet_labs = NULL,
43 | ...
44 | )
45 | }
46 | \arguments{
47 | \item{m}{A model object including an interaction term, or, alternately, a data frame recording conditional coefficients.}
48 |
49 | \item{var1}{The name (as a string) of the variable of interest in the interaction term; its conditional coefficient estimates will be plotted.}
50 |
51 | \item{var2}{The name (as a string) of the other variable in the interaction term.}
52 |
53 | \item{plot}{A logical value indicating whether the output is a plot or a dataframe including the conditional coefficient estimates of var1, their upper and lower bounds, and the corresponding values of var2.}
54 |
55 | \item{steps}{Desired length of the sequence. A non-negative number, which for seq and seq.int will be rounded up if fractional. The default is 100 or the unique categories in the \code{var2} (when it is less than 100. Also see \code{\link{unique}}).}
56 |
57 | \item{ci}{A numeric value defining the confidence intervals. The default value is 95\% (0.95).}
58 |
59 | \item{adjCI}{A logical value indication if applying the adjustment of confidence intervals to control the false discovery rate following the Esarey and Sumner (2017) procedure. (See also Benjamini and Hochberg 1995.) The default is FALSE; the plot presents the confidence intervals suggested by Brambor, Clark, and Golder (2006).}
60 |
61 | \item{hist}{A logical value indicating if there is a histogram of `var2` added at the bottom of the conditional effect plot.}
62 |
63 | \item{var2_dt}{A numerical value indicating the frequency distribution of `var2`. It is only used when `hist == TRUE`. When the object is a model, the default is the distribution of `var2` of the model.}
64 |
65 | \item{predPro}{A logical value with default of `FALSE`. When the `m` is an object of class `glm` and the argument is set to `TRUE`, the function will plot predicted probabilities at the values given by `var2_vals`.}
66 |
67 | \item{var2_vals}{A numerical value indicating the values the predicted probabilities are estimated, when `predPro` is `TRUE`.}
68 |
69 | \item{point}{A logical value determining the format of plot. By default, the function produces a line plot when var2 takes on ten or more distinct values and a point (dot-and-whisker) plot otherwise; option TRUE forces a point plot.}
70 |
71 | \item{sims}{Number of independent simulation draws used to calculate upper and lower bounds of coefficient estimates: lower values run faster; higher values produce smoother curves.}
72 |
73 | \item{xmin}{A numerical value indicating the minimum value shown of x shown in the graph. Rarely used.}
74 |
75 | \item{xmax}{A numerical value indicating the maximum value shown of x shown in the graph. Rarely used.}
76 |
77 | \item{ercolor}{A character value indicating the outline color of the whisker or ribbon.}
78 |
79 | \item{esize}{A numerical value indicating the size of the whisker or ribbon.}
80 |
81 | \item{ralpha}{A numerical value indicating the transparency of the ribbon.}
82 |
83 | \item{rfill}{A character value indicating the filling color of the ribbon.}
84 |
85 | \item{stats_cp}{A character value indicating what statistics to present as the plot note. Three options are available: "none", "ci", and "ks". The default is "none". See the Details for more information.}
86 |
87 | \item{txt_caption}{A character string to add a note for the plot, a value will sending to \code{ggplot2::labs(caption = txt_caption))}.}
88 |
89 | \item{facet_labs}{An optional character vector of facet labels to be used when plotting an interaction with a factor variable.}
90 |
91 | \item{...}{Other ggplot aesthetics arguments for points in the dot-whisker plot or lines in the line-ribbon plots. Not currently used.}
92 | }
93 | \value{
94 | The function returns a \code{ggplot} object.
95 | }
96 | \description{
97 | \code{interplot.default} is a method to calculate conditional coefficient estimates from the results of (generalized) linear regression models with interaction terms.
98 | }
99 | \details{
100 | \code{interplot.default} is a S3 method from the \code{interplot}. It works on two classes of objects:
101 | \itemize{
102 | \item Ordinary linear models (object class: \code{lm});
103 | \item Generalized linear models (object class: \code{glm}).
104 | }
105 |
106 | Because the output function is based on \code{\link[ggplot2]{ggplot}}, any additional arguments and layers supported by \code{ggplot2} can be added with the \code{+}.
107 |
108 | \code{interplot} visualizes the conditional effect based on simulated marginal effects. The simulation provides a probabilistic distribution of moderation effect of the conditioning variable (\code{var2}) at every preset values (including the minimum and maximum values) of the conditioned variable (\code{var1}), denoted as Emin and Emax. This output allows the function to further examine the conditional effect statistically in two ways. One is to examine if the distribution of \eqn{Emax - Emin} covers zero. The other is to directly compare Emin and Emax through statistical tools for distributional comparisons. Users can choose either method by setting the argument \code{stats_cp} to "ci" or "ks".
109 | \itemize{
110 | \item "ci" provides the confidence interval of the difference of \eqn{Emax - Emin}. An interval including 0 suggests no statistical difference before and after the conditional effect is applied, and vise versa.
111 | \item "ks" presents the result of a two-sample Kolmogorov-Smirnov test of the simulated distributions of Emin and Emax. The output includes a D statistics and a p-value of the null hypothesis that the two distributions come from the same distribution at the 0.05 level.
112 | }
113 |
114 | See an illustration in the package vignette.
115 | }
116 |
--------------------------------------------------------------------------------
/docs/reference/index.html:
--------------------------------------------------------------------------------
1 |
2 | Function reference • interplot
6 |
7 |
8 |
9 |
50 |
51 |
52 |
53 |
54 |
55 |
56 | Reference
57 |
58 |
59 |
60 | All functions
61 |
62 |
63 |
64 |
65 |
66 | Plot Conditional Coefficients in (Generalized) Linear Models with Interaction Terms
68 |
69 |
70 | Plot Conditional Coefficients in Mixed-Effects Models with Interaction Terms
72 |
73 |
74 | Plot Conditional Coefficients in (Generalized) Linear Models with Imputed Data and Interaction Terms
76 |
77 |
78 | Plot Conditional Coefficients in Mixed-Effects Models with Imputed Data and Interaction Terms
80 |
81 |
82 | Plot Conditional Coefficients in Models with Interaction Terms
84 |
85 |
86 | Plot Conditional Coefficients of a Variable in an Interaction Term
88 |
89 |
92 |
93 |
94 |
95 |
104 |
105 |
106 |
107 |
108 |
109 |
110 |
111 |
112 |
--------------------------------------------------------------------------------
/man/interplot.lmmi.Rd:
--------------------------------------------------------------------------------
1 | % Generated by roxygen2: do not edit by hand
2 | % Please edit documentation in R/Interplot_mi.R
3 | \name{interplot.lmmi}
4 | \alias{interplot.lmmi}
5 | \title{Plot Conditional Coefficients in (Generalized) Linear Models with Imputed Data and Interaction Terms}
6 | \source{
7 | Benjamini, Yoav, and Yosef Hochberg. 1995. "Controlling the False
8 | Discovery Rate: A Practical and Powerful Approach to Multiple Testing".
9 | Journal of the Royal Statistical Society, Series B 57(1): 289--300.
10 |
11 | Brambor, Thomas, William Roberts Clark, and Matt Golder.
12 | "Understanding interaction models: Improving empirical analyses". Political
13 | Analysis 14.1 (2006): 63-82.
14 |
15 | Esarey, Justin, and Jane Lawrence Sumner. 2015. "Marginal Effects in
16 | Interaction Models: Determining and Controlling the False Positive Rate".
17 | URL: \url{http://jee3.web.rice.edu/interaction-overconfidence.pdf}.
18 | }
19 | \usage{
20 | \method{interplot}{lmmi}(
21 | m,
22 | var1,
23 | var2,
24 | plot = TRUE,
25 | steps = NULL,
26 | ci = 0.95,
27 | adjCI = FALSE,
28 | hist = FALSE,
29 | var2_dt = NA,
30 | predPro = FALSE,
31 | var2_vals = NULL,
32 | point = FALSE,
33 | sims = 5000,
34 | xmin = NA,
35 | xmax = NA,
36 | ercolor = NA,
37 | esize = 0.5,
38 | ralpha = 0.5,
39 | rfill = "grey70",
40 | stats_cp = "none",
41 | txt_caption = NULL,
42 | facet_labs = NULL,
43 | ...
44 | )
45 | }
46 | \arguments{
47 | \item{m}{A model object including an interaction term, or, alternately, a data frame recording conditional coefficients.}
48 |
49 | \item{var1}{The name (as a string) of the variable of interest in the interaction term; its conditional coefficient estimates will be plotted.}
50 |
51 | \item{var2}{The name (as a string) of the other variable in the interaction term.}
52 |
53 | \item{plot}{A logical value indicating whether the output is a plot or a dataframe including the conditional coefficient estimates of var1, their upper and lower bounds, and the corresponding values of var2.}
54 |
55 | \item{steps}{Desired length of the sequence. A non-negative number, which for seq and seq.int will be rounded up if fractional. The default is 100 or the unique categories in the \code{var2} (when it is less than 100. Also see \code{\link{unique}}).}
56 |
57 | \item{ci}{A numeric value defining the confidence intervals. The default value is 95\% (0.95).}
58 |
59 | \item{adjCI}{A logical value indication if applying the adjustment of confidence intervals to control the false discovery rate following the Esarey and Sumner (2017) procedure. (See also Benjamini and Hochberg 1995.) The default is FALSE; the plot presents the confidence intervals suggested by Brambor, Clark, and Golder (2006).}
60 |
61 | \item{hist}{A logical value indicating if there is a histogram of `var2` added at the bottom of the conditional effect plot.}
62 |
63 | \item{var2_dt}{A numerical value indicating the frequency distribution of `var2`. It is only used when `hist == TRUE`. When the object is a model, the default is the distribution of `var2` of the model.}
64 |
65 | \item{predPro}{A logical value with default of `FALSE`. When the `m` is an object of class `glm` and the argument is set to `TRUE`, the function will plot predicted probabilities at the values given by `var2_vals`.}
66 |
67 | \item{var2_vals}{A numerical value indicating the values the predicted probabilities are estimated, when `predPro` is `TRUE`.}
68 |
69 | \item{point}{A logical value determining the format of plot. By default, the function produces a line plot when var2 takes on ten or more distinct values and a point (dot-and-whisker) plot otherwise; option TRUE forces a point plot.}
70 |
71 | \item{sims}{Number of independent simulation draws used to calculate upper and lower bounds of coefficient estimates: lower values run faster; higher values produce smoother curves.}
72 |
73 | \item{xmin}{A numerical value indicating the minimum value shown of x shown in the graph. Rarely used.}
74 |
75 | \item{xmax}{A numerical value indicating the maximum value shown of x shown in the graph. Rarely used.}
76 |
77 | \item{ercolor}{A character value indicating the outline color of the whisker or ribbon.}
78 |
79 | \item{esize}{A numerical value indicating the size of the whisker or ribbon.}
80 |
81 | \item{ralpha}{A numerical value indicating the transparency of the ribbon.}
82 |
83 | \item{rfill}{A character value indicating the filling color of the ribbon.}
84 |
85 | \item{stats_cp}{A character value indicating what statistics to present as the plot note. Three options are available: "none", "ci", and "ks". The default is "none". See the Details for more information.}
86 |
87 | \item{txt_caption}{A character string to add a note for the plot, a value will sending to \code{ggplot2::labs(caption = txt_caption))}.}
88 |
89 | \item{facet_labs}{An optional character vector of facet labels to be used when plotting an interaction with a factor variable.}
90 |
91 | \item{...}{Other ggplot aesthetics arguments for points in the dot-whisker plot or lines in the line-ribbon plots. Not currently used.}
92 | }
93 | \value{
94 | The function returns a \code{ggplot} object.
95 | }
96 | \description{
97 | \code{interplot.mi} is a method to calculate conditional coefficient estimates from the results of (generalized) linear regression models with interaction terms and multiply imputed data.
98 | }
99 | \details{
100 | \code{interplot.lmmi} and \code{interplot.glmmi} are S3 methods from the \code{interplot}. This function can work on interactions from results in the class of \code{list} generated by \code{mitools}.
101 |
102 | Because the output function is based on \code{\link[ggplot2]{ggplot}}, any additional arguments and layers supported by \code{ggplot2} can be added with the \code{+}.
103 |
104 | \code{interplot} visualizes the conditional effect based on simulated marginal effects. The simulation provides a probabilistic distribution of moderation effect of the conditioning variable (\code{var2}) at every preset values (including the minimum and maximum values) of the conditioned variable (\code{var1}), denoted as Emin and Emax. This output allows the function to further examine the conditional effect statistically in two ways. One is to examine if the distribution of \eqn{Emax - Emin} covers zero. The other is to directly compare Emin and Emax through statistical tools for distributional comparisons. Users can choose either method by setting the argument \code{stats_cp} to "ci" or "ks".
105 | \itemize{
106 | \item "ci" provides the confidence interval of the difference of \eqn{Emax - Emin}. An interval including 0 suggests no statistical difference before and after the conditional effect is applied, and vise versa.
107 | \item "ks" presents the result of a two-sample Kolmogorov-Smirnov test of the simulated distributions of Emin and Emax. The output includes a D statistics and a p-value of the null hypothesis that the two distributions come from the same distribution at the 0.05 level.
108 | }
109 |
110 |
111 | See an illustration in the package vignette.
112 | }
113 | \examples{
114 |
115 | library(interplot)
116 | library(mitools)
117 |
118 | data(smi)
119 | model1 <- with(smi, glm(drinkreg ~ wave * sex, family = binomial()))
120 |
121 | interplot(model1, var1 = "sex", var2 = "wave")
122 |
123 | }
124 |
--------------------------------------------------------------------------------
/vignettes/vignette.bib:
--------------------------------------------------------------------------------
1 | % Encoding: UTF-8
2 |
3 | @Article{Brambor2006,
4 | author = {Brambor, Thomas and Clark, William Roberts and Golder, Matt},
5 | title = {Understanding Interaction Models: Improving Empirical Analyses},
6 | journal = {Political Analysis},
7 | year = {2006},
8 | volume = {14},
9 | number = {1},
10 | pages = {63--82},
11 | __markedentry = {[Sammo:6]},
12 | publisher = {SPM-PMSAPSA}
13 | }
14 |
15 | @Book{Aiken1991,
16 | title = {Multiple regression: Testing and interpreting interactionsMultiple Regression: Testing and Interpreting Interactions},
17 | publisher = {Sage},
18 | year = {1991},
19 | author = {Aiken, Leona S and West, Stephen G and Reno, Raymond R}
20 | }
21 |
22 | @Article{WrightJr1976,
23 | author = {Wright Jr, Gerald C},
24 | title = {Linear Models for Evaluating Conditional Relationships},
25 | journal = {American Journal of Political Science},
26 | year = {1976},
27 | pages = {349--373},
28 | __markedentry = {[Sammo:]},
29 | publisher = {JSTOR}
30 | }
31 |
32 | @Article{Berry2012,
33 | author = {Berry, William D and Golder, Matt and Milton, Daniel},
34 | title = {Improving Tests of Theories Positing Interaction},
35 | journal = {Journal of Politics},
36 | year = {2012},
37 | volume = {74},
38 | number = {3},
39 | pages = {653--671},
40 | publisher = {Cambridge Univ Press}
41 | }
42 |
43 | @Unpublished{Hainmueller2016,
44 | author = {Hainmueller, Jens and Mummolo, Jonathan and Xu, Yiqing},
45 | title = {How Much Should We Trust Estimates from Multiplicative Interaction Models? Simple Tools to Improve Empirical Practice},
46 | year = {2016},
47 | timestamp = {2017-07-20},
48 | url = {https://www.ssrn.com/abstract=2739221},
49 | }
50 |
51 | @Article{Clark2006,
52 | author = {Clark, William Roberts and Gilligan, Michael J and Golder, Matt},
53 | title = {A Simple Multivariate Test for Asymmetric Hypotheses},
54 | journal = {Political Analysis},
55 | year = {2006},
56 | volume = {14},
57 | number = {3},
58 | pages = {311--331},
59 | publisher = {SPM-PMSAPSA}
60 | }
61 |
62 | @Article{Berry2016,
63 | author = {Berry, William D and DeMeritt, Jacqueline HR and Esarey, Justin},
64 | title = {Bias and Overconfidence in Parametric Models of Interactive Processes},
65 | journal = {American Journal of Political Science},
66 | year = {2016},
67 | volume = {60},
68 | number = {2},
69 | pages = {521--539},
70 | publisher = {Wiley Online Library},
71 | }
72 |
73 | @Unpublished{Esarey2017,
74 | author = {Esarey, Justin and Sumner, Jane Lawrence},
75 | title = {Marginal Effects in Interaction Models: Determining and Controlling the False Positive Rate},
76 | year = {2017},
77 | timestamp = {2017-07-20},
78 | url = {http://jee3.web.rice.edu/interaction-overconfidence.pdf},
79 | }
80 |
81 | @Article{Braumoeller2004,
82 | author = {Braumoeller, Bear F},
83 | title = {Hypothesis Testing and Multiplicative Interaction Terms},
84 | journal = {International Organization},
85 | year = {2004},
86 | volume = {58},
87 | number = {4},
88 | pages = {807--820},
89 | publisher = {Cambridge University Press},
90 | }
91 |
92 | @Article{Benjamini1995,
93 | author = {Benjamini, Yoav and Hochberg, Yosef},
94 | title = {Controlling the False Discovery Rate: A Practical and Powerful Approach to Multiple Testing},
95 | journal = {Journal of the Royal Statistical Society. Series B (Methodological)},
96 | year = {1995},
97 | pages = {289--300},
98 | publisher = {JSTOR},
99 | }
100 |
101 | @Article{Hanmer2013,
102 | author = {Hanmer, Michael J. and Ozan Kalkan, Kerem},
103 | title = {Behind the Curve: Clarifying the Best Approach to Calculating Predicted Probabilities and Marginal Effects from Limited Dependent Variable Models},
104 | journal = {American Journal of Political Science},
105 | year = {2013},
106 | volume = {57},
107 | number = {1},
108 | pages = {263--277},
109 | issn = {1540-5907},
110 | abstract = {Models designed for limited dependent variables are increasingly common in political science. Researchers estimating such models often give little attention to the coefficient estimates and instead focus on marginal effects, predicted probabilities, predicted counts, etc. Since the models are nonlinear, the estimated effects are sensitive to how one generates the predictions. The most common approach involves estimating the effect for the “average case.” But this approach creates a weaker connection between the results and the larger goals of the research enterprise and is thus less preferable than the observed-value approach. That is, rather than seeking to understand the effect for the average case, the goal is to obtain an estimate of the average effect in the population. In addition to the theoretical argument in favor of the observed-value approach, we illustrate via an empirical example and Monte Carlo simulations that the two approaches can produce substantively different results.},
111 | doi = {10.1111/j.1540-5907.2012.00602.x},
112 | publisher = {Blackwell Publishing Inc},
113 | timestamp = {2017-08-06},
114 | url = {http://dx.doi.org/10.1111/j.1540-5907.2012.00602.x},
115 | }
116 |
117 | @Article{Solt2016,
118 | author = {Solt, Frederick and Hu, Yue and Hudson, Kevan and Song, Jungmin and Yu, Dong Erico},
119 | title = {Economic Inequality and Belief in Meritocracy in the United States},
120 | journal = {Research \& Politics},
121 | year = {2016},
122 | volume = {3},
123 | number = {4},
124 | pages = {1--7},
125 | abstract = {How does the context of income inequality in which people live affect their belief in meritocracy, the ability to get ahead through hard work? A prominent recent study by Newman, Johnston, and Lown argues that, consistent with the conflict theory, exposure to higher levels of local income inequality leads lower-income people to become more likely to reject—and higher-income people to become more likely to accept—the dominant United States ideology of meritocracy. Here, we show that this conclusion is not supported by the studys own reported results and that even these results depend on pooling three distinctly different measures of meritocracy into a single analysis. We then demonstrate that analysis of a larger and more representative survey employing a single consistent measure of the dependent variable yields the opposite conclusion. Consistent with the relative power theory, among those with lower incomes, local contexts of greater inequality are associated with more widespread belief that people can get ahead if they are willing to work hard.},
126 | date = {2017-06-14},
127 | doi = {10.1177/2053168016672101},
128 | eprint = {http://rap.sagepub.com/content/3/4/2053168016672101.full.pdf},
129 | publisher = {SAGE Publications},
130 | timestamp = {2017-06-14},
131 | url = {http://rap.sagepub.com/content/3/4/2053168016672101},
132 | urldate = {2017-06-14},
133 | }
134 |
135 | @Comment{jabref-meta: databaseType:bibtex;}
136 |
--------------------------------------------------------------------------------
/dev/jss/jss_interplot.bib:
--------------------------------------------------------------------------------
1 | % Encoding: UTF-8
2 |
3 | @Article{Brambor2006,
4 | author = {Brambor, Thomas and Clark, William Roberts and Golder, Matt},
5 | title = {Understanding Interaction Models: Improving Empirical Analyses},
6 | journal = {Political Analysis},
7 | year = {2006},
8 | volume = {14},
9 | number = {1},
10 | pages = {63--82},
11 | __markedentry = {[Sammo:6]},
12 | publisher = {SPM-PMSAPSA}
13 | }
14 |
15 | @Book{Aiken1991,
16 | title = {Multiple regression: Testing and interpreting interactionsMultiple Regression: Testing and Interpreting Interactions},
17 | publisher = {Sage},
18 | year = {1991},
19 | author = {Aiken, Leona S and West, Stephen G and Reno, Raymond R}
20 | }
21 |
22 | @Article{WrightJr1976,
23 | author = {Wright Jr, Gerald C},
24 | title = {Linear Models for Evaluating Conditional Relationships},
25 | journal = {American Journal of Political Science},
26 | year = {1976},
27 | pages = {349--373},
28 | __markedentry = {[Sammo:]},
29 | publisher = {JSTOR}
30 | }
31 |
32 | @Article{Berry2012,
33 | author = {Berry, William D and Golder, Matt and Milton, Daniel},
34 | title = {Improving Tests of Theories Positing Interaction},
35 | journal = {Journal of Politics},
36 | year = {2012},
37 | volume = {74},
38 | number = {3},
39 | pages = {653--671},
40 | publisher = {Cambridge University Press}
41 | }
42 |
43 | @Article{Hainmueller2019,
44 | author = {Hainmueller, Jens and Mummolo, Jonathan and Xu, Yiqing},
45 | title = {How Much Should We Trust Estimates from Multiplicative Interaction Models? Simple Tools to Improve Empirical Practice},
46 | journal = {Political Analysis},
47 | year = {2012},
48 | volume = {27},
49 | number = {2},
50 | pages = {163-192},
51 | publisher = {SAGE}
52 | }
53 |
54 | @Article{Clark2006,
55 | author = {Clark, William Roberts and Gilligan, Michael J and Golder, Matt},
56 | title = {A Simple Multivariate Test for Asymmetric Hypotheses},
57 | journal = {Political Analysis},
58 | year = {2006},
59 | volume = {14},
60 | number = {3},
61 | pages = {311--331},
62 | publisher = {SPM-PMSAPSA}
63 | }
64 |
65 | @Article{Berry2016,
66 | author = {Berry, William D and DeMeritt, Jacqueline HR and Esarey, Justin},
67 | title = {Bias and Overconfidence in Parametric Models of Interactive Processes},
68 | journal = {American Journal of Political Science},
69 | year = {2016},
70 | volume = {60},
71 | number = {2},
72 | pages = {521--539},
73 | publisher = {Wiley Online Library},
74 | }
75 |
76 | @Article{Esarey2018,
77 | author = {Esarey, Justin and Sumner, Jane Lawrence},
78 | title = {Marginal Effects in Interaction Models: Determining and Controlling the False Positive Rate},
79 | journal = {Comparative Political Studies},
80 | year = {2016},
81 | volume = {51},
82 | number = {9},
83 | pages = {1144–1176},
84 | publisher = {JSTOR},
85 | }
86 |
87 | @Article{Braumoeller2004,
88 | author = {Braumoeller, Bear F},
89 | title = {Hypothesis Testing and Multiplicative Interaction Terms},
90 | journal = {International Organization},
91 | year = {2004},
92 | volume = {58},
93 | number = {4},
94 | pages = {807--820},
95 | publisher = {Cambridge University Press},
96 | }
97 |
98 | @Article{Benjamini1995,
99 | author = {Benjamini, Yoav and Hochberg, Yosef},
100 | title = {Controlling the False Discovery Rate: A Practical and Powerful Approach to Multiple Testing},
101 | journal = {Journal of the Royal Statistical Society. Series B (Methodological)},
102 | year = {1995},
103 | pages = {289--300},
104 | publisher = {JSTOR},
105 | }
106 |
107 | @Article{Hanmer2013,
108 | author = {Hanmer, Michael J. and Ozan Kalkan, Kerem},
109 | title = {Behind the Curve: Clarifying the Best Approach to Calculating Predicted Probabilities and Marginal Effects from Limited Dependent Variable Models},
110 | journal = {American Journal of Political Science},
111 | year = {2013},
112 | volume = {57},
113 | number = {1},
114 | pages = {263--277},
115 | issn = {1540-5907},
116 | abstract = {Models designed for limited dependent variables are increasingly common in political science. Researchers estimating such models often give little attention to the coefficient estimates and instead focus on marginal effects, predicted probabilities, predicted counts, etc. Since the models are nonlinear, the estimated effects are sensitive to how one generates the predictions. The most common approach involves estimating the effect for the “average case.” But this approach creates a weaker connection between the results and the larger goals of the research enterprise and is thus less preferable than the observed-value approach. That is, rather than seeking to understand the effect for the average case, the goal is to obtain an estimate of the average effect in the population. In addition to the theoretical argument in favor of the observed-value approach, we illustrate via an empirical example and Monte Carlo simulations that the two approaches can produce substantively different results.},
117 | doi = {10.1111/j.1540-5907.2012.00602.x},
118 | publisher = {Blackwell Publishing Inc},
119 | timestamp = {2017-08-06},
120 | url = {http://dx.doi.org/10.1111/j.1540-5907.2012.00602.x},
121 | }
122 |
123 | @Article{Solt2016,
124 | author = {Solt, Frederick and Hu, Yue and Hudson, Kevan and Song, Jungmin and Yu, Dong Erico},
125 | title = {Economic Inequality and Belief in Meritocracy in the United States},
126 | journal = {Research \& Politics},
127 | year = {2016},
128 | volume = {3},
129 | number = {4},
130 | pages = {1--7},
131 | abstract = {How does the context of income inequality in which people live affect their belief in meritocracy, the ability to get ahead through hard work? A prominent recent study by Newman, Johnston, and Lown argues that, consistent with the conflict theory, exposure to higher levels of local income inequality leads lower-income people to become more likely to reject—and higher-income people to become more likely to accept—the dominant United States ideology of meritocracy. Here, we show that this conclusion is not supported by the studys own reported results and that even these results depend on pooling three distinctly different measures of meritocracy into a single analysis. We then demonstrate that analysis of a larger and more representative survey employing a single consistent measure of the dependent variable yields the opposite conclusion. Consistent with the relative power theory, among those with lower incomes, local contexts of greater inequality are associated with more widespread belief that people can get ahead if they are willing to work hard.},
132 | date = {2017-06-14},
133 | doi = {10.1177/2053168016672101},
134 | eprint = {http://rap.sagepub.com/content/3/4/2053168016672101.full.pdf},
135 | publisher = {SAGE Publications},
136 | timestamp = {2017-06-14},
137 | url = {http://rap.sagepub.com/content/3/4/2053168016672101},
138 | urldate = {2017-06-14},
139 | }
140 |
141 | @Comment{jabref-meta: databaseType:bibtex;}
142 |
143 | @Article{Wright1976,
144 | title = {Linear models for evaluating conditional relationships},
145 | author = {Wright, Gerald C.},
146 | year = {1976},
147 | date = {1976},
148 | journal = {American Journal of Political Science},
149 | pages = {349},
150 | volume = {20},
151 | number = {2},
152 | doi = {10.2307/2110649},
153 | url = {https://www.jstor.org/stable/2110649?origin=crossref},
154 | langid = {en}
155 | }
156 |
157 |
158 |
--------------------------------------------------------------------------------
/docs/index.html:
--------------------------------------------------------------------------------
1 |
2 |
3 |
4 |
5 |
6 |
7 |
8 | Plot the Effects of Variables in Interaction Terms • interplot
9 |
10 |
11 |
12 |
13 |
14 |
15 |
17 |
21 |
22 |
23 |
24 |
25 |
26 |
75 |
76 |
77 |
78 |
79 |
80 |
81 |
82 | pkgdown
83 | 
84 |
85 |
86 |
88 | interplot is a tool for plotting the conditional coefficients (“marginal effects”) of variables included in multiplicative interaction terms. The function plots the changes in the coefficient of one variable in a two-way interaction term conditional on the value of the other included variable. The plot also includes simulated 95% confidential intervals of these coefficients.
89 | To install:
90 |
91 | - the latest released version:
install.packages("interplot").
92 | - the latest developing version:
remotes::install_github("sammo3182/interplot").
93 |
94 | More details are available at:
95 | http://cran.r-project.org/web/packages/interplot/vignettes/interplot-vignette.html
96 | Please note that this project is released with a Contributor Code of Conduct. By participating in this project you agree to abide by its terms.
97 |
98 |
99 |
100 |
101 |
140 |
141 |
142 |
143 |
154 |
155 |
156 |
157 |
158 |
159 |
160 |
161 |
162 |
163 |
--------------------------------------------------------------------------------
/man/interplot.Rd:
--------------------------------------------------------------------------------
1 | % Generated by roxygen2: do not edit by hand
2 | % Please edit documentation in R/Interplot.R
3 | \name{interplot}
4 | \alias{interplot}
5 | \title{Plot Conditional Coefficients of a Variable in an Interaction Term}
6 | \source{
7 | Benjamini, Yoav, and Yosef Hochberg. 1995. "Controlling the False
8 | Discovery Rate: A Practical and Powerful Approach to Multiple Testing".
9 | Journal of the Royal Statistical Society, Series B 57(1): 289--300.
10 |
11 | Brambor, Thomas, William Roberts Clark, and Matt Golder.
12 | "Understanding interaction models: Improving empirical analyses". Political
13 | Analysis 14.1 (2006): 63-82.
14 |
15 | Esarey, Justin, and Jane Lawrence Sumner. 2015. "Marginal Effects in
16 | Interaction Models: Determining and Controlling the False Positive Rate".
17 | URL: \url{https://jee3.web.rice.edu/interaction-overconfidence.pdf}.
18 | }
19 | \usage{
20 | interplot(
21 | m,
22 | var1,
23 | var2,
24 | plot = TRUE,
25 | steps = NULL,
26 | ci = 0.95,
27 | adjCI = FALSE,
28 | hist = FALSE,
29 | var2_dt = NA,
30 | predPro = FALSE,
31 | var2_vals = NULL,
32 | point = FALSE,
33 | sims = 1000,
34 | xmin = NA,
35 | xmax = NA,
36 | ercolor = NA,
37 | esize = 0.5,
38 | ralpha = 0.5,
39 | rfill = "grey70",
40 | stats_cp = "none",
41 | txt_caption = NULL,
42 | ...
43 | )
44 | }
45 | \arguments{
46 | \item{m}{A model object including an interaction term, or, alternately, a data frame generated by an earlier call to interplot using the argument plot = FALSE.}
47 |
48 | \item{var1}{The name (as a string) of the variable of interest in the interaction term; its conditional coefficient estimates will be plotted.}
49 |
50 | \item{var2}{The name (as a string) of the other variable in the interaction term}
51 |
52 | \item{plot}{A logical value indicating whether the output is a plot or a dataframe including the conditional coefficient estimates of var1, their upper and lower bounds, and the corresponding values of var2.}
53 |
54 | \item{steps}{Desired length of the sequence. A non-negative number, which for seq and seq.int will be rounded up if fractional. The default is 100 or the unique categories in the \code{var2} (when it is less than 100. Also see \code{\link{unique}}).}
55 |
56 | \item{ci}{A numeric value defining the confidence intervals. The default value is 95\% (0.95).}
57 |
58 | \item{adjCI}{A logical value indication if applying the adjustment of confidence intervals to control the false discovery rate following the Esarey and Sumner (2017) procedure. (See also Benjamini and Hochberg 1995.) The default is FALSE; the plot presents the confidence intervals suggested by Brambor, Clark, and Golder (2006). The functions dealing with multilevel model outputs in this package do not equip with this argument, because there is the controversy on the accurate degrees of freedom for multilevel models, esp. when random effect is engaged and the degrees of freedom is a necessary information to conduct the CI adjustment. See e.g., https://stat.ethz.ch/pipermail/r-help/2006-May/094765.html and https://stat.ethz.ch/pipermail/r-sig-mixed-models/2008q1/000517.html.}
59 |
60 | \item{hist}{A logical value indicating if there is a histogram of `var2` added at the bottom of the conditional effect plot.}
61 |
62 | \item{var2_dt}{A numerical value indicating the frequency distribution of `var2`. It is only used when `hist == TRUE`. When the object is a model, the default is the distribution of `var2` of the model.}
63 |
64 | \item{predPro}{A logical value with default of `FALSE`. When the `m` is an object of class `glm` or `glmerMod` and the argument is set to `TRUE`, the function will plot predicted probabilities at the values given by `var2_vals`.}
65 |
66 | \item{var2_vals}{A numerical value indicating the values the predicted probabilities are estimated, when `predPro` is `TRUE`.}
67 |
68 | \item{point}{A logical value determining the format of plot. By default, the function produces a line plot when var2 takes on ten or more distinct values and a point (dot-and-whisker) plot otherwise; option TRUE forces a point plot.}
69 |
70 | \item{sims}{Number of independent simulation draws used to calculate upper and lower bounds of coefficient estimates: lower values run faster; higher values produce smoother curves.}
71 |
72 | \item{xmin}{A numerical value indicating the minimum value shown of x shown in the graph. Rarely used.}
73 |
74 | \item{xmax}{A numerical value indicating the maximum value shown of x shown in the graph. Rarely used.}
75 |
76 | \item{ercolor}{A character value indicating the outline color of the whisker or ribbon.}
77 |
78 | \item{esize}{A numerical value indicating the size of the whisker or ribbon.}
79 |
80 | \item{ralpha}{A numerical value indicating the transparency of the ribbon.}
81 |
82 | \item{rfill}{A character value indicating the filling color of the ribbon.}
83 |
84 | \item{stats_cp}{A character value indicating what statistics to present as the plot note. Three options are available: "none", "ci", and "ks". The default is "none". See the Details for more information.}
85 |
86 | \item{txt_caption}{A character string to add a note for the plot, a value will sending to \code{ggplot2::labs(caption = txt_caption))}.}
87 |
88 | \item{...}{Other ggplot aesthetics arguments for points in the dot-whisker plot or lines in the line-ribbon plots. Not currently used.}
89 | }
90 | \value{
91 | The function returns a \code{ggplot} object.
92 | }
93 | \description{
94 | \code{interplot} is a generic function to produce a plot of the coefficient estimates of one variable in a two-way interaction conditional on the values of the other variable in the interaction term. The function invokes particular \code{methods} which depend on the \code{\link{class}} of the first argument.
95 | }
96 | \details{
97 | \code{interplot} visualizes the changes in the coefficient of one term in a two-way interaction conditioned by the other term. In the current version, the function works with interactions in the following classes of models:
98 | \itemize{
99 | \item Ordinary linear models (object class: \code{lm});
100 | \item Generalized linear models (object class: \code{glm});
101 | \item Linear mixed-effects models (object class: \code{lmerMod});
102 | \item Generalized linear mixed-effects models (object class: \code{glmerMod});
103 | \item Ordinary linear models with imputed data (object class: \code{list});
104 | \item Generalized linear models with imputed data (object class: \code{list})
105 | \item Linear mixed-effects models with imputed data (object class: \code{list});
106 | \item Generalized linear mixed-effects models with imputed data (object class: \code{list}).
107 | }
108 |
109 | The examples below illustrate how methods invoked by this generic deal with different type of objects.
110 |
111 | Because the output function is based on \code{\link[ggplot2]{ggplot}}, any additional arguments and layers supported by \code{ggplot2} can be added with the \code{+}.
112 |
113 | \code{interplot} visualizes the conditional effect based on simulated marginal effects. The simulation provides a probabilistic distribution of moderation effect of the conditioning variable (\code{var2}) at every preset values (including the minimum and maximum values) of the conditioned variable (\code{var1}), denoted as Emin and Emax. This output allows the function to further examine the conditional effect statistically in two ways. One is to examine if the distribution of \eqn{Emax - Emin} covers zero. The other is to directly compare Emin and Emax through statistical tools for distributional comparisons. Users can choose either method by setting the argument \code{stats_cp} to "ci" or "ks".
114 | \itemize{
115 | \item "ci" provides the confidence interval of the difference of \eqn{Emax - Emin}. An interval including 0 suggests no statistical difference before and after the conditional effect is applied, and vise versa.
116 | \item "ks" presents the result of a two-sample Kolmogorov-Smirnov test of the simulated distributions of Emin and Emax. The output includes a D statistics and a p-value of the null hypothesis that the two distributions come from the same distribution at the 0.05 level.
117 | }
118 |
119 | See an illustration in the package vignette.
120 | }
121 | \examples{
122 | data(mtcars)
123 | m_cyl <- lm(mpg ~ wt * cyl, data = mtcars)
124 | library(interplot)
125 |
126 | # Plot interactions with a continous conditioning variable
127 | interplot(m = m_cyl, var1 = 'cyl', var2 = 'wt') +
128 | xlab('Automobile Weight (thousands lbs)') +
129 | ylab('Estimated Coefficient for Number of Cylinders') +
130 | ggtitle('Estimated Coefficient of Engine Cylinders\non Mileage by Automobile Weight') +
131 | theme(plot.title = element_text(face='bold'))
132 |
133 |
134 | # Plot interactions with a categorical conditioning variable
135 | interplot(m = m_cyl, var1 = 'wt', var2 = 'cyl') +
136 | xlab('Number of Cylinders') +
137 | ylab('Estimated Coefficient for Automobile Weight (thousands lbs)') +
138 | ggtitle('Estimated Coefficient of Automobile Weight \non Mileage by Engine Cylinders') +
139 | theme(plot.title = element_text(face='bold'))
140 |
141 | }
142 |
--------------------------------------------------------------------------------
/docs/pkgdown.css:
--------------------------------------------------------------------------------
1 | /* Sticky footer */
2 |
3 | /**
4 | * Basic idea: https://philipwalton.github.io/solved-by-flexbox/demos/sticky-footer/
5 | * Details: https://github.com/philipwalton/solved-by-flexbox/blob/master/assets/css/components/site.css
6 | *
7 | * .Site -> body > .container
8 | * .Site-content -> body > .container .row
9 | * .footer -> footer
10 | *
11 | * Key idea seems to be to ensure that .container and __all its parents__
12 | * have height set to 100%
13 | *
14 | */
15 |
16 | html, body {
17 | height: 100%;
18 | }
19 |
20 | body {
21 | position: relative;
22 | }
23 |
24 | body > .container {
25 | display: flex;
26 | height: 100%;
27 | flex-direction: column;
28 | }
29 |
30 | body > .container .row {
31 | flex: 1 0 auto;
32 | }
33 |
34 | footer {
35 | margin-top: 45px;
36 | padding: 35px 0 36px;
37 | border-top: 1px solid #e5e5e5;
38 | color: #666;
39 | display: flex;
40 | flex-shrink: 0;
41 | }
42 | footer p {
43 | margin-bottom: 0;
44 | }
45 | footer div {
46 | flex: 1;
47 | }
48 | footer .pkgdown {
49 | text-align: right;
50 | }
51 | footer p {
52 | margin-bottom: 0;
53 | }
54 |
55 | img.icon {
56 | float: right;
57 | }
58 |
59 | /* Ensure in-page images don't run outside their container */
60 | .contents img {
61 | max-width: 100%;
62 | height: auto;
63 | }
64 |
65 | /* Fix bug in bootstrap (only seen in firefox) */
66 | summary {
67 | display: list-item;
68 | }
69 |
70 | /* Typographic tweaking ---------------------------------*/
71 |
72 | .contents .page-header {
73 | margin-top: calc(-60px + 1em);
74 | }
75 |
76 | dd {
77 | margin-left: 3em;
78 | }
79 |
80 | /* Section anchors ---------------------------------*/
81 |
82 | a.anchor {
83 | display: none;
84 | margin-left: 5px;
85 | width: 20px;
86 | height: 20px;
87 |
88 | background-image: url(./link.svg);
89 | background-repeat: no-repeat;
90 | background-size: 20px 20px;
91 | background-position: center center;
92 | }
93 |
94 | h1:hover .anchor,
95 | h2:hover .anchor,
96 | h3:hover .anchor,
97 | h4:hover .anchor,
98 | h5:hover .anchor,
99 | h6:hover .anchor {
100 | display: inline-block;
101 | }
102 |
103 | /* Fixes for fixed navbar --------------------------*/
104 |
105 | .contents h1, .contents h2, .contents h3, .contents h4 {
106 | padding-top: 60px;
107 | margin-top: -40px;
108 | }
109 |
110 | /* Navbar submenu --------------------------*/
111 |
112 | .dropdown-submenu {
113 | position: relative;
114 | }
115 |
116 | .dropdown-submenu>.dropdown-menu {
117 | top: 0;
118 | left: 100%;
119 | margin-top: -6px;
120 | margin-left: -1px;
121 | border-radius: 0 6px 6px 6px;
122 | }
123 |
124 | .dropdown-submenu:hover>.dropdown-menu {
125 | display: block;
126 | }
127 |
128 | .dropdown-submenu>a:after {
129 | display: block;
130 | content: " ";
131 | float: right;
132 | width: 0;
133 | height: 0;
134 | border-color: transparent;
135 | border-style: solid;
136 | border-width: 5px 0 5px 5px;
137 | border-left-color: #cccccc;
138 | margin-top: 5px;
139 | margin-right: -10px;
140 | }
141 |
142 | .dropdown-submenu:hover>a:after {
143 | border-left-color: #ffffff;
144 | }
145 |
146 | .dropdown-submenu.pull-left {
147 | float: none;
148 | }
149 |
150 | .dropdown-submenu.pull-left>.dropdown-menu {
151 | left: -100%;
152 | margin-left: 10px;
153 | border-radius: 6px 0 6px 6px;
154 | }
155 |
156 | /* Sidebar --------------------------*/
157 |
158 | #pkgdown-sidebar {
159 | margin-top: 30px;
160 | position: -webkit-sticky;
161 | position: sticky;
162 | top: 70px;
163 | }
164 |
165 | #pkgdown-sidebar h2 {
166 | font-size: 1.5em;
167 | margin-top: 1em;
168 | }
169 |
170 | #pkgdown-sidebar h2:first-child {
171 | margin-top: 0;
172 | }
173 |
174 | #pkgdown-sidebar .list-unstyled li {
175 | margin-bottom: 0.5em;
176 | }
177 |
178 | /* bootstrap-toc tweaks ------------------------------------------------------*/
179 |
180 | /* All levels of nav */
181 |
182 | nav[data-toggle='toc'] .nav > li > a {
183 | padding: 4px 20px 4px 6px;
184 | font-size: 1.5rem;
185 | font-weight: 400;
186 | color: inherit;
187 | }
188 |
189 | nav[data-toggle='toc'] .nav > li > a:hover,
190 | nav[data-toggle='toc'] .nav > li > a:focus {
191 | padding-left: 5px;
192 | color: inherit;
193 | border-left: 1px solid #878787;
194 | }
195 |
196 | nav[data-toggle='toc'] .nav > .active > a,
197 | nav[data-toggle='toc'] .nav > .active:hover > a,
198 | nav[data-toggle='toc'] .nav > .active:focus > a {
199 | padding-left: 5px;
200 | font-size: 1.5rem;
201 | font-weight: 400;
202 | color: inherit;
203 | border-left: 2px solid #878787;
204 | }
205 |
206 | /* Nav: second level (shown on .active) */
207 |
208 | nav[data-toggle='toc'] .nav .nav {
209 | display: none; /* Hide by default, but at >768px, show it */
210 | padding-bottom: 10px;
211 | }
212 |
213 | nav[data-toggle='toc'] .nav .nav > li > a {
214 | padding-left: 16px;
215 | font-size: 1.35rem;
216 | }
217 |
218 | nav[data-toggle='toc'] .nav .nav > li > a:hover,
219 | nav[data-toggle='toc'] .nav .nav > li > a:focus {
220 | padding-left: 15px;
221 | }
222 |
223 | nav[data-toggle='toc'] .nav .nav > .active > a,
224 | nav[data-toggle='toc'] .nav .nav > .active:hover > a,
225 | nav[data-toggle='toc'] .nav .nav > .active:focus > a {
226 | padding-left: 15px;
227 | font-weight: 500;
228 | font-size: 1.35rem;
229 | }
230 |
231 | /* orcid ------------------------------------------------------------------- */
232 |
233 | .orcid {
234 | font-size: 16px;
235 | color: #A6CE39;
236 | /* margins are required by official ORCID trademark and display guidelines */
237 | margin-left:4px;
238 | margin-right:4px;
239 | vertical-align: middle;
240 | }
241 |
242 | /* Reference index & topics ----------------------------------------------- */
243 |
244 | .ref-index th {font-weight: normal;}
245 |
246 | .ref-index td {vertical-align: top; min-width: 100px}
247 | .ref-index .icon {width: 40px;}
248 | .ref-index .alias {width: 40%;}
249 | .ref-index-icons .alias {width: calc(40% - 40px);}
250 | .ref-index .title {width: 60%;}
251 |
252 | .ref-arguments th {text-align: right; padding-right: 10px;}
253 | .ref-arguments th, .ref-arguments td {vertical-align: top; min-width: 100px}
254 | .ref-arguments .name {width: 20%;}
255 | .ref-arguments .desc {width: 80%;}
256 |
257 | /* Nice scrolling for wide elements --------------------------------------- */
258 |
259 | table {
260 | display: block;
261 | overflow: auto;
262 | }
263 |
264 | /* Syntax highlighting ---------------------------------------------------- */
265 |
266 | pre, code, pre code {
267 | background-color: #f8f8f8;
268 | color: #333;
269 | }
270 | pre, pre code {
271 | white-space: pre-wrap;
272 | word-break: break-all;
273 | overflow-wrap: break-word;
274 | }
275 |
276 | pre {
277 | border: 1px solid #eee;
278 | }
279 |
280 | pre .img, pre .r-plt {
281 | margin: 5px 0;
282 | }
283 |
284 | pre .img img, pre .r-plt img {
285 | background-color: #fff;
286 | }
287 |
288 | code a, pre a {
289 | color: #375f84;
290 | }
291 |
292 | a.sourceLine:hover {
293 | text-decoration: none;
294 | }
295 |
296 | .fl {color: #1514b5;}
297 | .fu {color: #000000;} /* function */
298 | .ch,.st {color: #036a07;} /* string */
299 | .kw {color: #264D66;} /* keyword */
300 | .co {color: #888888;} /* comment */
301 |
302 | .error {font-weight: bolder;}
303 | .warning {font-weight: bolder;}
304 |
305 | /* Clipboard --------------------------*/
306 |
307 | .hasCopyButton {
308 | position: relative;
309 | }
310 |
311 | .btn-copy-ex {
312 | position: absolute;
313 | right: 0;
314 | top: 0;
315 | visibility: hidden;
316 | }
317 |
318 | .hasCopyButton:hover button.btn-copy-ex {
319 | visibility: visible;
320 | }
321 |
322 | /* headroom.js ------------------------ */
323 |
324 | .headroom {
325 | will-change: transform;
326 | transition: transform 200ms linear;
327 | }
328 | .headroom--pinned {
329 | transform: translateY(0%);
330 | }
331 | .headroom--unpinned {
332 | transform: translateY(-100%);
333 | }
334 |
335 | /* mark.js ----------------------------*/
336 |
337 | mark {
338 | background-color: rgba(255, 255, 51, 0.5);
339 | border-bottom: 2px solid rgba(255, 153, 51, 0.3);
340 | padding: 1px;
341 | }
342 |
343 | /* vertical spacing after htmlwidgets */
344 | .html-widget {
345 | margin-bottom: 10px;
346 | }
347 |
348 | /* fontawesome ------------------------ */
349 |
350 | .fab {
351 | font-family: "Font Awesome 5 Brands" !important;
352 | }
353 |
354 | /* don't display links in code chunks when printing */
355 | /* source: https://stackoverflow.com/a/10781533 */
356 | @media print {
357 | code a:link:after, code a:visited:after {
358 | content: "";
359 | }
360 | }
361 |
362 | /* Section anchors ---------------------------------
363 | Added in pandoc 2.11: https://github.com/jgm/pandoc-templates/commit/9904bf71
364 | */
365 |
366 | div.csl-bib-body { }
367 | div.csl-entry {
368 | clear: both;
369 | }
370 | .hanging-indent div.csl-entry {
371 | margin-left:2em;
372 | text-indent:-2em;
373 | }
374 | div.csl-left-margin {
375 | min-width:2em;
376 | float:left;
377 | }
378 | div.csl-right-inline {
379 | margin-left:2em;
380 | padding-left:1em;
381 | }
382 | div.csl-indent {
383 | margin-left: 2em;
384 | }
385 |
--------------------------------------------------------------------------------
/R/Interplot.R:
--------------------------------------------------------------------------------
1 | #' Plot Conditional Coefficients of a Variable in an Interaction Term
2 | #'
3 | #' \code{interplot} is a generic function to produce a plot of the coefficient estimates of one variable in a two-way interaction conditional on the values of the other variable in the interaction term. The function invokes particular \code{methods} which depend on the \code{\link{class}} of the first argument.
4 | #'
5 | #' @param m A model object including an interaction term, or, alternately, a data frame generated by an earlier call to interplot using the argument plot = FALSE.
6 | #' @param var1 The name (as a string) of the variable of interest in the interaction term; its conditional coefficient estimates will be plotted.
7 | #' @param var2 The name (as a string) of the other variable in the interaction term
8 | #' @param plot A logical value indicating whether the output is a plot or a dataframe including the conditional coefficient estimates of var1, their upper and lower bounds, and the corresponding values of var2.
9 | #' @param steps Desired length of the sequence. A non-negative number, which for seq and seq.int will be rounded up if fractional. The default is 100 or the unique categories in the \code{var2} (when it is less than 100. Also see \code{\link{unique}}).
10 | #' @param ci A numeric value defining the confidence intervals. The default value is 95\% (0.95).
11 | #' @param adjCI A logical value indication if applying the adjustment of confidence intervals to control the false discovery rate following the Esarey and Sumner (2017) procedure. (See also Benjamini and Hochberg 1995.) The default is FALSE; the plot presents the confidence intervals suggested by Brambor, Clark, and Golder (2006). The functions dealing with multilevel model outputs in this package do not equip with this argument, because there is the controversy on the accurate degrees of freedom for multilevel models, esp. when random effect is engaged and the degrees of freedom is a necessary information to conduct the CI adjustment. See e.g., https://stat.ethz.ch/pipermail/r-help/2006-May/094765.html and https://stat.ethz.ch/pipermail/r-sig-mixed-models/2008q1/000517.html.
12 | #' @param hist A logical value indicating if there is a histogram of `var2` added at the bottom of the conditional effect plot.
13 | #' @param var2_dt A numerical value indicating the frequency distribution of `var2`. It is only used when `hist == TRUE`. When the object is a model, the default is the distribution of `var2` of the model.
14 | #' @param predPro A logical value with default of `FALSE`. When the `m` is an object of class `glm` or `glmerMod` and the argument is set to `TRUE`, the function will plot predicted probabilities at the values given by `var2_vals`.
15 | #' @param var2_vals A numerical value indicating the values the predicted probabilities are estimated, when `predPro` is `TRUE`.
16 | #' @param point A logical value determining the format of plot. By default, the function produces a line plot when var2 takes on ten or more distinct values and a point (dot-and-whisker) plot otherwise; option TRUE forces a point plot.
17 | #' @param sims Number of independent simulation draws used to calculate upper and lower bounds of coefficient estimates: lower values run faster; higher values produce smoother curves.
18 | #' @param xmin A numerical value indicating the minimum value shown of x shown in the graph. Rarely used.
19 | #' @param xmax A numerical value indicating the maximum value shown of x shown in the graph. Rarely used.
20 | #' @param ercolor A character value indicating the outline color of the whisker or ribbon.
21 | #' @param esize A numerical value indicating the size of the whisker or ribbon.
22 | #' @param ralpha A numerical value indicating the transparency of the ribbon.
23 | #' @param rfill A character value indicating the filling color of the ribbon.
24 | #' @param stats_cp A character value indicating what statistics to present as the plot note. Three options are available: "none", "ci", and "ks". The default is "none". See the Details for more information.
25 | #' @param txt_caption A character string to add a note for the plot, a value will sending to \code{ggplot2::labs(caption = txt_caption))}.
26 | #' @param ... Other ggplot aesthetics arguments for points in the dot-whisker plot or lines in the line-ribbon plots. Not currently used.
27 | #'
28 | #' @details \code{interplot} visualizes the changes in the coefficient of one term in a two-way interaction conditioned by the other term. In the current version, the function works with interactions in the following classes of models:
29 | #' \itemize{
30 | #' \item Ordinary linear models (object class: \code{lm});
31 | #' \item Generalized linear models (object class: \code{glm});
32 | #' \item Linear mixed-effects models (object class: \code{lmerMod});
33 | #' \item Generalized linear mixed-effects models (object class: \code{glmerMod});
34 | #' \item Ordinary linear models with imputed data (object class: \code{list});
35 | #' \item Generalized linear models with imputed data (object class: \code{list})
36 | #' \item Linear mixed-effects models with imputed data (object class: \code{list});
37 | #' \item Generalized linear mixed-effects models with imputed data (object class: \code{list}).
38 | #' }
39 | #'
40 | #' The examples below illustrate how methods invoked by this generic deal with different type of objects.
41 | #'
42 | #' Because the output function is based on \code{\link[ggplot2]{ggplot}}, any additional arguments and layers supported by \code{ggplot2} can be added with the \code{+}.
43 | #'
44 | #' \code{interplot} visualizes the conditional effect based on simulated marginal effects. The simulation provides a probabilistic distribution of moderation effect of the conditioning variable (\code{var2}) at every preset values (including the minimum and maximum values) of the conditioned variable (\code{var1}), denoted as Emin and Emax. This output allows the function to further examine the conditional effect statistically in two ways. One is to examine if the distribution of \eqn{Emax - Emin} covers zero. The other is to directly compare Emin and Emax through statistical tools for distributional comparisons. Users can choose either method by setting the argument \code{stats_cp} to "ci" or "ks".
45 | #' \itemize{
46 | #' \item "ci" provides the confidence interval of the difference of \eqn{Emax - Emin}. An interval including 0 suggests no statistical difference before and after the conditional effect is applied, and vise versa.
47 | #' \item "ks" presents the result of a two-sample Kolmogorov-Smirnov test of the simulated distributions of Emin and Emax. The output includes a D statistics and a p-value of the null hypothesis that the two distributions come from the same distribution at the 0.05 level.
48 | #' }
49 | #'
50 | #' See an illustration in the package vignette.
51 | #'
52 | #' @return The function returns a \code{ggplot} object.
53 | #'
54 | #'
55 | #' @examples
56 | #' data(mtcars)
57 | #' m_cyl <- lm(mpg ~ wt * cyl, data = mtcars)
58 | #' library(interplot)
59 | #'
60 | #' # Plot interactions with a continous conditioning variable
61 | #' interplot(m = m_cyl, var1 = 'cyl', var2 = 'wt') +
62 | #' xlab('Automobile Weight (thousands lbs)') +
63 | #' ylab('Estimated Coefficient for Number of Cylinders') +
64 | #' ggtitle('Estimated Coefficient of Engine Cylinders\non Mileage by Automobile Weight') +
65 | #' theme(plot.title = element_text(face='bold'))
66 | #'
67 | #'
68 | #' # Plot interactions with a categorical conditioning variable
69 | #' interplot(m = m_cyl, var1 = 'wt', var2 = 'cyl') +
70 | #' xlab('Number of Cylinders') +
71 | #' ylab('Estimated Coefficient for Automobile Weight (thousands lbs)') +
72 | #' ggtitle('Estimated Coefficient of Automobile Weight \non Mileage by Engine Cylinders') +
73 | #' theme(plot.title = element_text(face='bold'))
74 | #'
75 | #' @source Benjamini, Yoav, and Yosef Hochberg. 1995. "Controlling the False
76 | #' Discovery Rate: A Practical and Powerful Approach to Multiple Testing".
77 | #' Journal of the Royal Statistical Society, Series B 57(1): 289--300.
78 | #'
79 | #' Brambor, Thomas, William Roberts Clark, and Matt Golder.
80 | #' "Understanding interaction models: Improving empirical analyses". Political
81 | #' Analysis 14.1 (2006): 63-82.
82 | #'
83 | #' Esarey, Justin, and Jane Lawrence Sumner. 2015. "Marginal Effects in
84 | #' Interaction Models: Determining and Controlling the False Positive Rate".
85 | #' URL: \url{https://jee3.web.rice.edu/interaction-overconfidence.pdf}.
86 | #'
87 | #' @export
88 |
89 |
90 | interplot <- function(m, var1, var2, plot = TRUE, steps = NULL, ci = .95, adjCI = FALSE, hist = FALSE, var2_dt = NA, predPro = FALSE, var2_vals = NULL, point = FALSE, sims = 1000, xmin = NA, xmax = NA, ercolor = NA, esize = 0.5, ralpha = 0.5, rfill = "grey70", stats_cp = "none", txt_caption = NULL, ...) {
91 |
92 |
93 | if (class(m)[1] == "list") {
94 | if (class(m[[1]])[1] == "lmerMod") {
95 | class(m) <- "mlmmi"
96 | }
97 | if (class(m[[1]])[1] == "glmerMod") {
98 | class(m) <- "gmlmmi"
99 | }
100 | if (class(m[[1]])[1] == "lm") {
101 | class(m) <- "lmmi"
102 | }
103 | if (class(m[[1]])[1] == "glm") {
104 | class(m) <- "glmmi"
105 | }
106 | }
107 |
108 | if (class(m)[1] == "data.frame")
109 | class(m) <- "plot"
110 |
111 |
112 | UseMethod("interplot", m)
113 | }
114 |
--------------------------------------------------------------------------------
/docs/docsearch.css:
--------------------------------------------------------------------------------
1 | /* Docsearch -------------------------------------------------------------- */
2 | /*
3 | Source: https://github.com/algolia/docsearch/
4 | License: MIT
5 | */
6 |
7 | .algolia-autocomplete {
8 | display: block;
9 | -webkit-box-flex: 1;
10 | -ms-flex: 1;
11 | flex: 1
12 | }
13 |
14 | .algolia-autocomplete .ds-dropdown-menu {
15 | width: 100%;
16 | min-width: none;
17 | max-width: none;
18 | padding: .75rem 0;
19 | background-color: #fff;
20 | background-clip: padding-box;
21 | border: 1px solid rgba(0, 0, 0, .1);
22 | box-shadow: 0 .5rem 1rem rgba(0, 0, 0, .175);
23 | }
24 |
25 | @media (min-width:768px) {
26 | .algolia-autocomplete .ds-dropdown-menu {
27 | width: 175%
28 | }
29 | }
30 |
31 | .algolia-autocomplete .ds-dropdown-menu::before {
32 | display: none
33 | }
34 |
35 | .algolia-autocomplete .ds-dropdown-menu [class^=ds-dataset-] {
36 | padding: 0;
37 | background-color: rgb(255,255,255);
38 | border: 0;
39 | max-height: 80vh;
40 | }
41 |
42 | .algolia-autocomplete .ds-dropdown-menu .ds-suggestions {
43 | margin-top: 0
44 | }
45 |
46 | .algolia-autocomplete .algolia-docsearch-suggestion {
47 | padding: 0;
48 | overflow: visible
49 | }
50 |
51 | .algolia-autocomplete .algolia-docsearch-suggestion--category-header {
52 | padding: .125rem 1rem;
53 | margin-top: 0;
54 | font-size: 1.3em;
55 | font-weight: 500;
56 | color: #00008B;
57 | border-bottom: 0
58 | }
59 |
60 | .algolia-autocomplete .algolia-docsearch-suggestion--wrapper {
61 | float: none;
62 | padding-top: 0
63 | }
64 |
65 | .algolia-autocomplete .algolia-docsearch-suggestion--subcategory-column {
66 | float: none;
67 | width: auto;
68 | padding: 0;
69 | text-align: left
70 | }
71 |
72 | .algolia-autocomplete .algolia-docsearch-suggestion--content {
73 | float: none;
74 | width: auto;
75 | padding: 0
76 | }
77 |
78 | .algolia-autocomplete .algolia-docsearch-suggestion--content::before {
79 | display: none
80 | }
81 |
82 | .algolia-autocomplete .ds-suggestion:not(:first-child) .algolia-docsearch-suggestion--category-header {
83 | padding-top: .75rem;
84 | margin-top: .75rem;
85 | border-top: 1px solid rgba(0, 0, 0, .1)
86 | }
87 |
88 | .algolia-autocomplete .ds-suggestion .algolia-docsearch-suggestion--subcategory-column {
89 | display: block;
90 | padding: .1rem 1rem;
91 | margin-bottom: 0.1;
92 | font-size: 1.0em;
93 | font-weight: 400
94 | /* display: none */
95 | }
96 |
97 | .algolia-autocomplete .algolia-docsearch-suggestion--title {
98 | display: block;
99 | padding: .25rem 1rem;
100 | margin-bottom: 0;
101 | font-size: 0.9em;
102 | font-weight: 400
103 | }
104 |
105 | .algolia-autocomplete .algolia-docsearch-suggestion--text {
106 | padding: 0 1rem .5rem;
107 | margin-top: -.25rem;
108 | font-size: 0.8em;
109 | font-weight: 400;
110 | line-height: 1.25
111 | }
112 |
113 | .algolia-autocomplete .algolia-docsearch-footer {
114 | width: 110px;
115 | height: 20px;
116 | z-index: 3;
117 | margin-top: 10.66667px;
118 | float: right;
119 | font-size: 0;
120 | line-height: 0;
121 | }
122 |
123 | .algolia-autocomplete .algolia-docsearch-footer--logo {
124 | background-image: url("data:image/svg+xml;utf8,");
125 | background-repeat: no-repeat;
126 | background-position: 50%;
127 | background-size: 100%;
128 | overflow: hidden;
129 | text-indent: -9000px;
130 | width: 100%;
131 | height: 100%;
132 | display: block;
133 | transform: translate(-8px);
134 | }
135 |
136 | .algolia-autocomplete .algolia-docsearch-suggestion--highlight {
137 | color: #FF8C00;
138 | background: rgba(232, 189, 54, 0.1)
139 | }
140 |
141 |
142 | .algolia-autocomplete .algolia-docsearch-suggestion--text .algolia-docsearch-suggestion--highlight {
143 | box-shadow: inset 0 -2px 0 0 rgba(105, 105, 105, .5)
144 | }
145 |
146 | .algolia-autocomplete .ds-suggestion.ds-cursor .algolia-docsearch-suggestion--content {
147 | background-color: rgba(192, 192, 192, .15)
148 | }
149 |
--------------------------------------------------------------------------------
/docs/reference/interplot.lmerMod.html:
--------------------------------------------------------------------------------
1 |
2 | Plot Conditional Coefficients in Mixed-Effects Models with Interaction Terms — interplot.lmerMod • interplot
6 |
7 |
8 |
9 |
50 |
51 |
52 |
53 |
54 |
55 |
56 | Plot Conditional Coefficients in Mixed-Effects Models with Interaction Terms
57 | Source: R/Interplot_mlm.R
58 | interplot.lmerMod.Rd
59 |
60 |
61 |
62 | interplot.mlm is a method to calculate conditional coefficient estimates from the results of multilevel (mixed-effects) regression models with interaction terms.
63 |
64 |
65 |
66 | # S3 method for lmerMod
67 | interplot(
68 | m,
69 | var1,
70 | var2,
71 | plot = TRUE,
72 | steps = NULL,
73 | ci = 0.95,
74 | adjCI = FALSE,
75 | hist = FALSE,
76 | var2_dt = NA,
77 | predPro = FALSE,
78 | var2_vals = NULL,
79 | point = FALSE,
80 | sims = 5000,
81 | xmin = NA,
82 | xmax = NA,
83 | ercolor = NA,
84 | esize = 0.5,
85 | ralpha = 0.5,
86 | rfill = "grey70",
87 | stats_cp = "none",
88 | txt_caption = NULL,
89 | facet_labs = NULL,
90 | ...
91 | )
92 |
93 |
94 |
95 | Arguments
96 | - m
97 | A model object including an interaction term, or, alternately, a data frame recording conditional coefficients.
- var1
101 | The name (as a string) of the variable of interest in the interaction term; its conditional coefficient estimates will be plotted.
- var2
105 | The name (as a string) of the other variable in the interaction term.
- plot
109 | A logical value indicating whether the output is a plot or a dataframe including the conditional coefficient estimates of var1, their upper and lower bounds, and the corresponding values of var2.
- steps
113 | Desired length of the sequence. A non-negative number, which for seq and seq.int will be rounded up if fractional. The default is 100 or the unique categories in the var2 (when it is less than 100. Also see unique).
- ci
117 | A numeric value defining the confidence intervals. The default value is 95% (0.95).
- adjCI
121 | Not working for `lmer` outputs yet.
- hist
125 | A logical value indicating if there is a histogram of `var2` added at the bottom of the conditional effect plot.
- var2_dt
129 | A numerical value indicating the frequency distribution of `var2`. It is only used when `hist == TRUE`. When the object is a model, the default is the distribution of `var2` of the model.
- predPro
133 | A logical value with default of `FALSE`. When the `m` is an object of class `glmerMod` and the argument is set to `TRUE`, the function will plot predicted probabilities at the values given by `var2_vals`.
- var2_vals
137 | A numerical value indicating the values the predicted probabilities are estimated, when `predPro` is `TRUE`.
- point
141 | A logical value determining the format of plot. By default, the function produces a line plot when var2 takes on ten or more distinct values and a point (dot-and-whisker) plot otherwise; option TRUE forces a point plot.
- sims
145 | Number of independent simulation draws used to calculate upper and lower bounds of coefficient estimates: lower values run faster; higher values produce smoother curves.
- xmin
149 | A numerical value indicating the minimum value shown of x shown in the graph. Rarely used.
- xmax
153 | A numerical value indicating the maximum value shown of x shown in the graph. Rarely used.
- ercolor
157 | A character value indicating the outline color of the whisker or ribbon.
- esize
161 | A numerical value indicating the size of the whisker or ribbon.
- ralpha
165 | A numerical value indicating the transparency of the ribbon.
- rfill
169 | A character value indicating the filling color of the ribbon.
- stats_cp
173 | A character value indicating what statistics to present as the plot note. Three options are available: "none", "ci", and "ks". The default is "none". See the Details for more information.
- txt_caption
177 | A character string to add a note for the plot, a value will sending to ggplot2::labs(caption = txt_caption)).
- facet_labs
181 | An optional character vector of facet labels to be used when plotting an interaction with a factor variable.
- ...
185 | Other ggplot aesthetics arguments for points in the dot-whisker plot or lines in the line-ribbon plots. Not currently used.
188 |
189 | Value
190 |
191 |
192 | The function returns a ggplot object.
193 |
194 |
195 | Details
196 | interplot.mlm is a S3 method from the interplot. It works on mixed-effects objects with class lmerMod and glmerMod.
197 | Because the output function is based on ggplot, any additional arguments and layers supported by ggplot2 can be added with the +.
198 | interplot visualizes the conditional effect based on simulated marginal effects. The simulation provides a probabilistic distribution of moderation effect of the conditioning variable (var2) at every preset values (including the minimum and maximum values) of the conditioned variable (var1), denoted as Emin and Emax. This output allows the function to further examine the conditional effect statistically in two ways. One is to examine if the distribution of \(Emax - Emin\) covers zero. The other is to directly compare Emin and Emax through statistical tools for distributional comparisons. Users can choose either method by setting the argument stats_cp to "ci" or "ks".
"ci" provides the confidence interval of the difference of \(Emax - Emin\). An interval including 0 suggests no statistical difference before and after the conditional effect is applied, and vise versa.
"ks" presents the result of a two-sample Kolmogorov-Smirnov test of the simulated distributions of Emin and Emax. The output includes a D statistics and a p-value of the null hypothesis that the two distributions come from the same distribution at the 0.05 level.
See an illustration in the package vignette.
201 |
202 |
203 |
204 |
207 |
208 |
209 |
210 |
219 |
220 |
221 |
222 |
223 |
224 |
225 |
226 |
227 |
--------------------------------------------------------------------------------
`).
76 | getTopLevel: function($scope) {
77 | for (var i = 1; i <= 6; i++) {
78 | var $headings = this.findOrFilter($scope, 'h' + i);
79 | if ($headings.length > 1) {
80 | return i;
81 | }
82 | }
83 |
84 | return 1;
85 | },
86 |
87 | // returns the elements for the top level, and the next below it
88 | getHeadings: function($scope, topLevel) {
89 | var topSelector = 'h' + topLevel;
90 |
91 | var secondaryLevel = topLevel + 1;
92 | var secondarySelector = 'h' + secondaryLevel;
93 |
94 | return this.findOrFilter($scope, topSelector + ',' + secondarySelector);
95 | },
96 |
97 | getNavLevel: function(el) {
98 | return parseInt(el.tagName.charAt(1), 10);
99 | },
100 |
101 | populateNav: function($topContext, topLevel, $headings) {
102 | var $context = $topContext;
103 | var $prevNav;
104 |
105 | var helpers = this;
106 | $headings.each(function(i, el) {
107 | var $newNav = helpers.generateNavItem(el);
108 | var navLevel = helpers.getNavLevel(el);
109 |
110 | // determine the proper $context
111 | if (navLevel === topLevel) {
112 | // use top level
113 | $context = $topContext;
114 | } else if ($prevNav && $context === $topContext) {
115 | // create a new level of the tree and switch to it
116 | $context = helpers.createChildNavList($prevNav);
117 | } // else use the current $context
118 |
119 | $context.append($newNav);
120 |
121 | $prevNav = $newNav;
122 | });
123 | },
124 |
125 | parseOps: function(arg) {
126 | var opts;
127 | if (arg.jquery) {
128 | opts = {
129 | $nav: arg
130 | };
131 | } else {
132 | opts = arg;
133 | }
134 | opts.$scope = opts.$scope || $(document.body);
135 | return opts;
136 | }
137 | },
138 |
139 | // accepts a jQuery object, or an options object
140 | init: function(opts) {
141 | opts = this.helpers.parseOps(opts);
142 |
143 | // ensure that the data attribute is in place for styling
144 | opts.$nav.attr('data-toggle', 'toc');
145 |
146 | var $topContext = this.helpers.createChildNavList(opts.$nav);
147 | var topLevel = this.helpers.getTopLevel(opts.$scope);
148 | var $headings = this.helpers.getHeadings(opts.$scope, topLevel);
149 | this.helpers.populateNav($topContext, topLevel, $headings);
150 | }
151 | };
152 |
153 | $(function() {
154 | $('nav[data-toggle="toc"]').each(function(i, el) {
155 | var $nav = $(el);
156 | Toc.init($nav);
157 | });
158 | });
159 | })();
160 |
--------------------------------------------------------------------------------
/vignettes/vignette.bib.bak:
--------------------------------------------------------------------------------
1 | % Encoding: UTF-8
2 |
3 | @Article{Brambor2006,
4 | author = {Brambor, Thomas and Clark, William Roberts and Golder, Matt},
5 | title = {Understanding Interaction Models: Improving Empirical Analyses},
6 | journal = {Political Analysis},
7 | year = {2006},
8 | volume = {14},
9 | number = {1},
10 | pages = {63--82},
11 | __markedentry = {[Sammo:6]},
12 | publisher = {SPM-PMSAPSA}
13 | }
14 |
15 | @Book{Aiken1991,
16 | title = {Multiple regression: Testing and interpreting interactionsMultiple Regression: Testing and Interpreting Interactions},
17 | publisher = {Sage},
18 | year = {1991},
19 | author = {Aiken, Leona S and West, Stephen G and Reno, Raymond R}
20 | }
21 |
22 | @Article{WrightJr1976,
23 | author = {Wright Jr, Gerald C},
24 | title = {Linear Models for Evaluating Conditional Relationships},
25 | journal = {American Journal of Political Science},
26 | year = {1976},
27 | pages = {349--373},
28 | __markedentry = {[Sammo:]},
29 | publisher = {JSTOR}
30 | }
31 |
32 | @Article{Berry2012,
33 | author = {Berry, William D and Golder, Matt and Milton, Daniel},
34 | title = {Improving Tests of Theories Positing Interaction},
35 | journal = {Journal of Politics},
36 | year = {2012},
37 | volume = {74},
38 | number = {3},
39 | pages = {653--671},
40 | publisher = {Cambridge Univ Press}
41 | }
42 |
43 | @Unpublished{Hainmueller2016,
44 | author = {Hainmueller, Jens and Mummolo, Jonathan and Xu, Yiqing},
45 | title = {How Much Should We Trust Estimates from Multiplicative Interaction Models? Simple Tools to Improve Empirical Practice},
46 | year = {2016},
47 | timestamp = {2017-07-20},
48 | url = {https://ssrn.com/abstract=2739221},
49 | }
50 |
51 | @Article{Clark2006,
52 | author = {Clark, William Roberts and Gilligan, Michael J and Golder, Matt},
53 | title = {A Simple Multivariate Test for Asymmetric Hypotheses},
54 | journal = {Political Analysis},
55 | year = {2006},
56 | volume = {14},
57 | number = {3},
58 | pages = {311--331},
59 | publisher = {SPM-PMSAPSA}
60 | }
61 |
62 | @Article{Berry2016,
63 | author = {Berry, William D and DeMeritt, Jacqueline HR and Esarey, Justin},
64 | title = {Bias and Overconfidence in Parametric Models of Interactive Processes},
65 | journal = {American Journal of Political Science},
66 | year = {2016},
67 | volume = {60},
68 | number = {2},
69 | pages = {521--539},
70 | publisher = {Wiley Online Library},
71 | }
72 |
73 | @Unpublished{Esarey2017,
74 | author = {Esarey, Justin and Sumner, Jane Lawrence},
75 | title = {Marginal Effects in Interaction Models: Determining and Controlling the False Positive Rate},
76 | year = {2017},
77 | timestamp = {2017-07-20},
78 | url = {http://jee3.web.rice.edu/interaction-overconfidence.pdf},
79 | }
80 |
81 | @Article{Braumoeller2004,
82 | author = {Braumoeller, Bear F},
83 | title = {Hypothesis Testing and Multiplicative Interaction Terms},
84 | journal = {International Organization},
85 | year = {2004},
86 | volume = {58},
87 | number = {4},
88 | pages = {807--820},
89 | publisher = {Cambridge University Press},
90 | }
91 |
92 | @Article{Benjamini1995,
93 | author = {Benjamini, Yoav and Hochberg, Yosef},
94 | title = {Controlling the False Discovery Rate: A Practical and Powerful Approach to Multiple Testing},
95 | journal = {Journal of the Royal Statistical Society. Series B (Methodological)},
96 | year = {1995},
97 | pages = {289--300},
98 | publisher = {JSTOR},
99 | }
100 |
101 | @Article{Hanmer2013,
102 | author = {Hanmer, Michael J. and Ozan Kalkan, Kerem},
103 | title = {Behind the Curve: Clarifying the Best Approach to Calculating Predicted Probabilities and Marginal Effects from Limited Dependent Variable Models},
104 | journal = {American Journal of Political Science},
105 | year = {2013},
106 | volume = {57},
107 | number = {1},
108 | pages = {263--277},
109 | issn = {1540-5907},
110 | abstract = {Models designed for limited dependent variables are increasingly common in political science. Researchers estimating such models often give little attention to the coefficient estimates and instead focus on marginal effects, predicted probabilities, predicted counts, etc. Since the models are nonlinear, the estimated effects are sensitive to how one generates the predictions. The most common approach involves estimating the effect for the “average case.” But this approach creates a weaker connection between the results and the larger goals of the research enterprise and is thus less preferable than the observed-value approach. That is, rather than seeking to understand the effect for the average case, the goal is to obtain an estimate of the average effect in the population. In addition to the theoretical argument in favor of the observed-value approach, we illustrate via an empirical example and Monte Carlo simulations that the two approaches can produce substantively different results.},
111 | doi = {10.1111/j.1540-5907.2012.00602.x},
112 | publisher = {Blackwell Publishing Inc},
113 | timestamp = {2017-08-06},
114 | url = {http://dx.doi.org/10.1111/j.1540-5907.2012.00602.x},
115 | }
116 |
117 | @Comment{jabref-meta: databaseType:bibtex;}
118 |
--------------------------------------------------------------------------------
/docs/404.html:
--------------------------------------------------------------------------------
1 |
2 |
3 |
4 |
5 |
6 |
7 |
8 | Page not found (404) • interplot
9 |
10 |
11 |
12 |
13 |
14 |
15 |
19 |
20 |
21 |
22 |
23 |
24 |
73 |
74 |
75 |
76 |
77 |
78 |
79 |
80 | Page not found (404)
81 |
82 |
83 | Content not found. Please use links in the navbar.
84 |
85 |
86 |
87 |
91 |
92 |
93 |
94 |
95 |
96 |
107 |
108 |
109 |
110 |
111 |
112 |
113 |
114 |
115 |
116 |
--------------------------------------------------------------------------------
/man/interplot.lmerMod.Rd:
--------------------------------------------------------------------------------
1 | % Generated by roxygen2: do not edit by hand
2 | % Please edit documentation in R/Interplot_mlm.R
3 | \name{interplot.lmerMod}
4 | \alias{interplot.lmerMod}
5 | \title{Plot Conditional Coefficients in Mixed-Effects Models with Interaction Terms}
6 | \usage{
7 | \method{interplot}{lmerMod}(
8 | m,
9 | var1,
10 | var2,
11 | plot = TRUE,
12 | steps = NULL,
13 | ci = 0.95,
14 | adjCI = FALSE,
15 | hist = FALSE,
16 | var2_dt = NA,
17 | predPro = FALSE,
18 | var2_vals = NULL,
19 | point = FALSE,
20 | sims = 5000,
21 | xmin = NA,
22 | xmax = NA,
23 | ercolor = NA,
24 | esize = 0.5,
25 | ralpha = 0.5,
26 | rfill = "grey70",
27 | stats_cp = "none",
28 | txt_caption = NULL,
29 | facet_labs = NULL,
30 | ...
31 | )
32 | }
33 | \arguments{
34 | \item{m}{A model object including an interaction term, or, alternately, a data frame recording conditional coefficients.}
35 |
36 | \item{var1}{The name (as a string) of the variable of interest in the interaction term; its conditional coefficient estimates will be plotted.}
37 |
38 | \item{var2}{The name (as a string) of the other variable in the interaction term.}
39 |
40 | \item{plot}{A logical value indicating whether the output is a plot or a dataframe including the conditional coefficient estimates of var1, their upper and lower bounds, and the corresponding values of var2.}
41 |
42 | \item{steps}{Desired length of the sequence. A non-negative number, which for seq and seq.int will be rounded up if fractional. The default is 100 or the unique categories in the \code{var2} (when it is less than 100. Also see \code{\link{unique}}).}
43 |
44 | \item{ci}{A numeric value defining the confidence intervals. The default value is 95\% (0.95).}
45 |
46 | \item{adjCI}{Not working for `lmer` outputs yet.}
47 |
48 | \item{hist}{A logical value indicating if there is a histogram of `var2` added at the bottom of the conditional effect plot.}
49 |
50 | \item{var2_dt}{A numerical value indicating the frequency distribution of `var2`. It is only used when `hist == TRUE`. When the object is a model, the default is the distribution of `var2` of the model.}
51 |
52 | \item{predPro}{A logical value with default of `FALSE`. When the `m` is an object of class `glmerMod` and the argument is set to `TRUE`, the function will plot predicted probabilities at the values given by `var2_vals`.}
53 |
54 | \item{var2_vals}{A numerical value indicating the values the predicted probabilities are estimated, when `predPro` is `TRUE`.}
55 |
56 | \item{point}{A logical value determining the format of plot. By default, the function produces a line plot when var2 takes on ten or more distinct values and a point (dot-and-whisker) plot otherwise; option TRUE forces a point plot.}
57 |
58 | \item{sims}{Number of independent simulation draws used to calculate upper and lower bounds of coefficient estimates: lower values run faster; higher values produce smoother curves.}
59 |
60 | \item{xmin}{A numerical value indicating the minimum value shown of x shown in the graph. Rarely used.}
61 |
62 | \item{xmax}{A numerical value indicating the maximum value shown of x shown in the graph. Rarely used.}
63 |
64 | \item{ercolor}{A character value indicating the outline color of the whisker or ribbon.}
65 |
66 | \item{esize}{A numerical value indicating the size of the whisker or ribbon.}
67 |
68 | \item{ralpha}{A numerical value indicating the transparency of the ribbon.}
69 |
70 | \item{rfill}{A character value indicating the filling color of the ribbon.}
71 |
72 | \item{stats_cp}{A character value indicating what statistics to present as the plot note. Three options are available: "none", "ci", and "ks". The default is "none". See the Details for more information.}
73 |
74 | \item{txt_caption}{A character string to add a note for the plot, a value will sending to \code{ggplot2::labs(caption = txt_caption))}.}
75 |
76 | \item{facet_labs}{An optional character vector of facet labels to be used when plotting an interaction with a factor variable.}
77 |
78 | \item{...}{Other ggplot aesthetics arguments for points in the dot-whisker plot or lines in the line-ribbon plots. Not currently used.}
79 | }
80 | \value{
81 | The function returns a \code{ggplot} object.
82 | }
83 | \description{
84 | \code{interplot.mlm} is a method to calculate conditional coefficient estimates from the results of multilevel (mixed-effects) regression models with interaction terms.
85 | }
86 | \details{
87 | \code{interplot.mlm} is a S3 method from the \code{interplot}. It works on mixed-effects objects with class \code{lmerMod} and \code{glmerMod}.
88 |
89 | Because the output function is based on \code{\link[ggplot2]{ggplot}}, any additional arguments and layers supported by \code{ggplot2} can be added with the \code{+}.
90 |
91 | \code{interplot} visualizes the conditional effect based on simulated marginal effects. The simulation provides a probabilistic distribution of moderation effect of the conditioning variable (\code{var2}) at every preset values (including the minimum and maximum values) of the conditioned variable (\code{var1}), denoted as Emin and Emax. This output allows the function to further examine the conditional effect statistically in two ways. One is to examine if the distribution of \eqn{Emax - Emin} covers zero. The other is to directly compare Emin and Emax through statistical tools for distributional comparisons. Users can choose either method by setting the argument \code{stats_cp} to "ci" or "ks".
92 | \itemize{
93 | \item "ci" provides the confidence interval of the difference of \eqn{Emax - Emin}. An interval including 0 suggests no statistical difference before and after the conditional effect is applied, and vise versa.
94 | \item "ks" presents the result of a two-sample Kolmogorov-Smirnov test of the simulated distributions of Emin and Emax. The output includes a D statistics and a p-value of the null hypothesis that the two distributions come from the same distribution at the 0.05 level.
95 | }
96 |
97 | See an illustration in the package vignette.
98 | }
99 |
--------------------------------------------------------------------------------
/vignettes/interplot-vignette.R:
--------------------------------------------------------------------------------
1 | ## ----setup, include=FALSE-----------------------------------------------------
2 | knitr::opts_chunk$set(
3 | message = FALSE,
4 | warning = FALSE,
5 | dpi = 300
6 | )
7 |
8 | ## -----------------------------------------------------------------------------
9 | data(mtcars) #load the data
10 |
11 | ## -----------------------------------------------------------------------------
12 | m_cyl <- lm(mpg ~ wt * cyl, data = mtcars)
13 | summary(m_cyl)
14 |
15 | ## ----out.width="80%"----------------------------------------------------------
16 | library(interplot)
17 |
18 | interplot(m = m_cyl, var1 = "cyl", var2 = "wt")
19 |
20 | ## ----out.width="80%"----------------------------------------------------------
21 | interplot(m = m_cyl, var1 = "wt", var2 = "cyl")
22 |
23 | ## ----out.width="80%"----------------------------------------------------------
24 | interplot(m = m_cyl, var1 = "wt", var2 = "cyl", ci = .9, point = T)
25 |
26 | ## ----out.width="80%"----------------------------------------------------------
27 | interplot(m = m_cyl, var1 = "cyl", var2 = "wt", point = T) +
28 | # changing the angle of x labels for a clearer vision
29 | theme(axis.text.x = element_text(angle=90))
30 |
31 | ## ----out.width="80%"----------------------------------------------------------
32 | interplot(m = m_cyl, var1 = "cyl", var2 = "wt") +
33 | # Add labels for X and Y axes
34 | xlab("Automobile Weight (thousands lbs)") +
35 | ylab("Estimated Coefficient for\nNumber of Cylinders") +
36 | # Change the background
37 | theme_bw() +
38 | # Add the title
39 | ggtitle("Estimated Coefficient of Engine Cylinders \non Mileage by Automobile Weight") +
40 | theme(plot.title = element_text(face="bold")) +
41 | # Add a horizontal line at y = 0
42 | geom_hline(yintercept = 0, linetype = "dashed")
43 |
44 | ## ----out.width="80%"----------------------------------------------------------
45 | interplot(m = m_cyl, var1 = "wt", var2 = "cyl", ercolor = "blue", esize = 1.5) +
46 | geom_point(size = 2, color = "red")
47 |
48 | ## ----out.width="80%"----------------------------------------------------------
49 | m_wt <- lm(mpg ~ wt + I(wt^2), data = mtcars)
50 |
51 | interplot(m = m_wt, var1 = "wt", var2 = "wt")
52 |
53 | ## ----out.width="80%"----------------------------------------------------------
54 | mtcars$gear <- factor(mtcars$gear)
55 | m_gear <- lm(mpg ~ gear * wt, data = mtcars)
56 |
57 | interplot(m = m_gear, var1 = "wt", var2 = "gear")
58 |
59 | ## ----out.width="80%"----------------------------------------------------------
60 | interplot(m = m_gear, var1 = "wt", var2 = "gear", facet_labs = c("4-speed", "5-speed"))
61 |
62 | ## ----out.width="80%"----------------------------------------------------------
63 | interplot(m = m_cyl, var1 = "cyl", var2 = "wt", hist = TRUE) +
64 | geom_hline(yintercept = 0, linetype = "dashed")
65 |
66 | ## ----out.width="80%"----------------------------------------------------------
67 | interplot(m = m_cyl, var1 = "cyl", var2 = "wt", hist = TRUE) +
68 | aes(color = "pink") + theme(legend.position="none") + # geom_line(color = "pink") +
69 | geom_hline(yintercept = 0, linetype = "dashed")
70 |
71 | ## -----------------------------------------------------------------------------
72 | set.seed(313)
73 |
74 | interplot(m = m_cyl, var1 = "cyl", var2 = "wt", stats_cp = "ci")
75 |
76 | ## -----------------------------------------------------------------------------
77 | set.seed(313)
78 |
79 | interplot(m = m_cyl, var1 = "cyl", var2 = "wt", stats_cp = "ci",
80 | txt_caption = "\n Source: Motor Trend Car Road Tests 1973.")
81 |
82 | ## ----out.width="80%"----------------------------------------------------------
83 | stdCI_plot <- interplot(m = m_wt, var1 = "wt", var2 = "wt", adjCI = FALSE) +
84 | ggtitle("Marginal Effects with Standard CIs")
85 | adjCI_plot <- interplot(m = m_wt, var1 = "wt", var2 = "wt", adjCI = TRUE) +
86 | ggtitle("Marginal Effects with Adjusted CIs")
87 |
88 | library(gridExtra)
89 | grid.arrange(stdCI_plot, adjCI_plot, ncol = 2)
90 |
91 | ## ----out.width="80%"----------------------------------------------------------
92 | pew1.w <- read.csv("pew1_w.csv")
93 |
94 | m <- glm(formula=meritocracy~ginicnty+income_i+ginicnty:income_i+income_cnty+black_cnty+
95 | perc_bush04+pop_cnty+educ_i+age_i+gender_i+unemp_i+union_i+partyid_i+
96 | ideo_i+attend_i+survid2006+survid2007+survid2009,
97 | data=pew1.w,family=binomial(link="logit"))
98 |
99 | plot_avg <- interplot(m, var1 = "ginicnty",var2 = "income_i", predPro = FALSE) +
100 | ggtitle("Average Conditional Effects")
101 |
102 | plot_3val <- interplot(m, var1 = "ginicnty",var2 = "income_i", predPro = TRUE, var2_vals = c(min(pew1.w$income_i), max(pew1.w$income_i))) +
103 | ggtitle("Conditional Predicted Probabilities for \nCitizens with Low and High Incomes") +
104 | scale_colour_discrete(guide = guide_legend(title = "Income"), labels = c("Low", "High")) +
105 | scale_fill_discrete(guide = guide_legend(title = "Income"), labels = c("Low", "High")) +
106 | theme(legend.position = c(0, .8), legend.justification = c(0, .5))
107 |
108 | grid.arrange(plot_avg, plot_3val, ncol = 2)
109 |
110 | ## ----out.width="80%"----------------------------------------------------------
111 | # Create a fake dataset of conditional effects
112 | fake <- rnorm(100, 0, 1)
113 | coef1 <- fake * sample(.5:2.5, 100, replace = T)
114 | lb <- coef1 - .5
115 | ub <- coef1 + .5
116 |
117 | df_fake <- data.frame(cbind(fake, coef1, lb, ub))
118 |
119 | # Use interplot directly with the dataset
120 | interplot(df_fake)
121 |
122 | ## ----out.width="80%"----------------------------------------------------------
123 | var2_fake <- fake
124 | # Set `hist` to TRUE is required to superimpose a histogram.
125 | interplot(df_fake, hist = TRUE, var2_dt = var2_fake)
126 | ## The ribbon and histogram do not fit. This is just an illustration
127 |
128 |
--------------------------------------------------------------------------------
/man/interplot.mlmmi.Rd:
--------------------------------------------------------------------------------
1 | % Generated by roxygen2: do not edit by hand
2 | % Please edit documentation in R/Interplot_mlmmi.R
3 | \name{interplot.mlmmi}
4 | \alias{interplot.mlmmi}
5 | \title{Plot Conditional Coefficients in Mixed-Effects Models with Imputed Data and Interaction Terms}
6 | \usage{
7 | \method{interplot}{mlmmi}(
8 | m,
9 | var1,
10 | var2,
11 | plot = TRUE,
12 | steps = NULL,
13 | ci = 0.95,
14 | adjCI = FALSE,
15 | hist = FALSE,
16 | var2_dt = NA,
17 | predPro = FALSE,
18 | var2_vals = NULL,
19 | point = FALSE,
20 | sims = 5000,
21 | xmin = NA,
22 | xmax = NA,
23 | ercolor = NA,
24 | esize = 0.5,
25 | ralpha = 0.5,
26 | rfill = "grey70",
27 | stats_cp = "none",
28 | txt_caption = NULL,
29 | facet_labs = NULL,
30 | ...
31 | )
32 | }
33 | \arguments{
34 | \item{m}{A model object including an interaction term, or, alternately, a data frame recording conditional coefficients.}
35 |
36 | \item{var1}{The name (as a string) of the variable of interest in the interaction term; its conditional coefficient estimates will be plotted.}
37 |
38 | \item{var2}{The name (as a string) of the other variable in the interaction term.}
39 |
40 | \item{plot}{A logical value indicating whether the output is a plot or a dataframe including the conditional coefficient estimates of var1, their upper and lower bounds, and the corresponding values of var2.}
41 |
42 | \item{steps}{Desired length of the sequence. A non-negative number, which for seq and seq.int will be rounded up if fractional. The default is 100 or the unique categories in the \code{var2} (when it is less than 100. Also see \code{\link{unique}}).}
43 |
44 | \item{ci}{A numeric value defining the confidence intervals. The default value is 95\% (0.95).}
45 |
46 | \item{adjCI}{Not working for `lmer` outputs yet.}
47 |
48 | \item{hist}{A logical value indicating if there is a histogram of `var2` added at the bottom of the conditional effect plot.}
49 |
50 | \item{var2_dt}{A numerical value indicating the frequency distribution of `var2`. It is only used when `hist == TRUE`. When the object is a model, the default is the distribution of `var2` of the model.}
51 |
52 | \item{predPro}{A logical value with default of `FALSE`. When the `m` is an object of class `glmerMod` and the argument is set to `TRUE`, the function will plot predicted probabilities at the values given by `var2_vals`.}
53 |
54 | \item{var2_vals}{A numerical value indicating the values the predicted probabilities are estimated, when `predPro` is `TRUE`.}
55 |
56 | \item{point}{A logical value determining the format of plot. By default, the function produces a line plot when var2 takes on ten or more distinct values and a point (dot-and-whisker) plot otherwise; option TRUE forces a point plot.}
57 |
58 | \item{sims}{Number of independent simulation draws used to calculate upper and lower bounds of coefficient estimates: lower values run faster; higher values produce smoother curves.}
59 |
60 | \item{xmin}{A numerical value indicating the minimum value shown of x shown in the graph. Rarely used.}
61 |
62 | \item{xmax}{A numerical value indicating the maximum value shown of x shown in the graph. Rarely used.}
63 |
64 | \item{ercolor}{A character value indicating the outline color of the whisker or ribbon.}
65 |
66 | \item{esize}{A numerical value indicating the size of the whisker or ribbon.}
67 |
68 | \item{ralpha}{A numerical value indicating the transparency of the ribbon.}
69 |
70 | \item{rfill}{A character value indicating the filling color of the ribbon.}
71 |
72 | \item{stats_cp}{A character value indicating what statistics to present as the plot note. Three options are available: "none", "ci", and "ks". The default is "none". See the Details for more information.}
73 |
74 | \item{txt_caption}{A character string to add a note for the plot, a value will sending to \code{ggplot2::labs(caption = txt_caption))}.}
75 |
76 | \item{facet_labs}{An optional character vector of facet labels to be used when plotting an interaction with a factor variable.}
77 |
78 | \item{...}{Other ggplot aesthetics arguments for points in the dot-whisker plot or lines in the line-ribbon plots. Not currently used.}
79 | }
80 | \value{
81 | The function returns a \code{ggplot} object.
82 | }
83 | \description{
84 | \code{interplot.mlmmi} is a method to calculate conditional coefficient estimates from the results of multilevel (mixed-effects) regression models with interaction terms and multiply imputed data.
85 | }
86 | \details{
87 | \code{interplot.mlmmi} and \code{interplot.gmlmmi} are S3 methods from the \code{interplot}. It works on lists of mixed-effects objects with class \code{lmerMod} and \code{glmerMod} generated by \code{mitools} and \code{lme4}.
88 |
89 | Because the output function is based on \code{\link[ggplot2]{ggplot}}, any additional arguments and layers supported by \code{ggplot2} can be added with the \code{+}.
90 |
91 | \code{interplot} visualizes the conditional effect based on simulated marginal effects. The simulation provides a probabilistic distribution of moderation effect of the conditioning variable (\code{var2}) at every preset values (including the minimum and maximum values) of the conditioned variable (\code{var1}), denoted as Emin and Emax. This output allows the function to further examine the conditional effect statistically in two ways. One is to examine if the distribution of \eqn{Emax - Emin} covers zero. The other is to directly compare Emin and Emax through statistical tools for distributional comparisons. Users can choose either method by setting the argument \code{stats_cp} to "ci" or "ks".
92 | \itemize{
93 | \item "ci" provides the confidence interval of the difference of \eqn{Emax - Emin}. An interval including 0 suggests no statistical difference before and after the conditional effect is applied, and vise versa.
94 | \item "ks" presents the result of a two-sample Kolmogorov-Smirnov test of the simulated distributions of Emin and Emax. The output includes a D statistics and a p-value of the null hypothesis that the two distributions come from the same distribution at the 0.05 level.
95 | }
96 |
97 | See an illustration in the package vignette.
98 | }
99 |
--------------------------------------------------------------------------------
/docs/authors.html:
--------------------------------------------------------------------------------
1 |
2 | Authors and Citation • interplot
6 |
7 |
8 |
108 |
109 |
110 |
111 |
112 |
113 |
114 |
115 |
116 |
--------------------------------------------------------------------------------
/man/interplot.plot.Rd:
--------------------------------------------------------------------------------
1 | % Generated by roxygen2: do not edit by hand
2 | % Please edit documentation in R/Interplot_plot.R
3 | \name{interplot.plot}
4 | \alias{interplot.plot}
5 | \title{Plot Conditional Coefficients in Models with Interaction Terms}
6 | \usage{
7 | \method{interplot}{plot}(
8 | m,
9 | var1 = NULL,
10 | var2 = NULL,
11 | plot = TRUE,
12 | steps = NULL,
13 | ci = 0.95,
14 | adjCI = FALSE,
15 | hist = FALSE,
16 | var2_dt = NULL,
17 | predPro = FALSE,
18 | var2_vals = NULL,
19 | point = FALSE,
20 | sims = 5000,
21 | xmin = NA,
22 | xmax = NA,
23 | ercolor = NA,
24 | esize = 0.5,
25 | ralpha = 0.5,
26 | rfill = "grey70",
27 | stats_cp = "none",
28 | txt_caption = NULL,
29 | ci_diff = NULL,
30 | ks_diff = NULL,
31 | ...
32 | )
33 | }
34 | \arguments{
35 | \item{m}{A model object including an interaction term, or, alternately, a data frame recording conditional coefficients. This data frame should includes four columns:
36 | \itemize{
37 | \item fake: The sequence of \code{var1} (the item whose effect will be conditioned on in the interaction);
38 | \item coef1: The point estimates of the coefficient of \code{var1} at each break point.
39 | \item ub: The upper bound of the simulated 95\% CI.
40 | \item lb: The lower bound of the simulated 95\% CI.
41 | }}
42 |
43 | \item{var1}{The name (as a string) of the variable of interest in the interaction term; its conditional coefficient estimates will be plotted.}
44 |
45 | \item{var2}{The name (as a string) of the other variable in the interaction term.}
46 |
47 | \item{plot}{A logical value indicating whether the output is a plot or a dataframe including the conditional coefficient estimates of var1, their upper and lower bounds, and the corresponding values of var2.}
48 |
49 | \item{steps}{Desired length of the sequence. A non-negative number, which for seq and seq.int will be rounded up if fractional. The default is 100 or the unique categories in the \code{var2} (when it is less than 100. Also see \code{\link{unique}}).}
50 |
51 | \item{ci}{is a numeric value inherited from the data wrangling functions in this package. Adding it here is just for the method consistency.}
52 |
53 | \item{adjCI}{Succeeded from the data management functions in `interplot` package.}
54 |
55 | \item{hist}{A logical value indicating if there is a histogram of `var2` added at the bottom of the conditional effect plot.}
56 |
57 | \item{var2_dt}{A numerical value indicating the frequency distribution of `var2`. It is only used when `hist == TRUE`. When the object is a model, the default is the distribution of `var2` of the model.}
58 |
59 | \item{predPro}{A logical value with default of `FALSE`. When the `m` is an output of a general linear model (class `glm` or `glmerMod`) and the argument is set to `TRUE`, the function will plot predicted probabilities at the values given by `var2_vals`.}
60 |
61 | \item{var2_vals}{A numerical value indicating the values the predicted probabilities are estimated, when `predPro` is `TRUE`.}
62 |
63 | \item{point}{A logical value determining the format of plot. By default, the function produces a line plot when var2 takes on ten or more distinct values and a point (dot-and-whisker) plot otherwise; option TRUE forces a point plot.}
64 |
65 | \item{sims}{Number of independent simulation draws used to calculate upper and lower bounds of coefficient estimates: lower values run faster; higher values produce smoother curves.}
66 |
67 | \item{xmin}{A numerical value indicating the minimum value shown of x shown in the graph. Rarely used.}
68 |
69 | \item{xmax}{A numerical value indicating the maximum value shown of x shown in the graph. Rarely used.}
70 |
71 | \item{ercolor}{A character value indicating the outline color of the whisker or ribbon.}
72 |
73 | \item{esize}{A numerical value indicating the size of the whisker or ribbon.}
74 |
75 | \item{ralpha}{A numerical value indicating the transparency of the ribbon.}
76 |
77 | \item{rfill}{A character value indicating the filling color of the ribbon.}
78 |
79 | \item{stats_cp}{A character value indicating what statistics to present as the plot note. Three options are available: "none", "ci", and "ks". The default is "none". See the Details for more information.}
80 |
81 | \item{txt_caption}{A character string to add a note for the plot, a value will sending to \code{ggplot2::labs(caption = txt_caption))}.}
82 |
83 | \item{ci_diff}{A numerical vector with a pair of values indicating the confidence intervals of the difference between \code{var1} and \code{var2}.}
84 |
85 | \item{ks_diff}{A \code{ks.test} object of the effect of \code{var1} conditioned on \code{var2}.}
86 |
87 | \item{...}{Other ggplot aesthetics arguments for points in the dot-whisker plot or lines in the line-ribbon plots. Not currently used.}
88 | }
89 | \value{
90 | The function returns a \code{ggplot} object.
91 | }
92 | \description{
93 | Graph based on the data frame of statistics about the conditional effect of an interaction.
94 | }
95 | \details{
96 | \code{interplot.plot} is a S3 method from the \code{interplot}. It generates plots of conditional coefficients.
97 |
98 | Because the output function is based on \code{\link[ggplot2]{ggplot}}, any additional arguments and layers supported by \code{ggplot2} can be added with the \code{+}.
99 |
100 | \code{interplot} visualizes the conditional effect based on simulated marginal effects. The simulation provides a probabilistic distribution of moderation effect of the conditioning variable (\code{var2}) at every preset values (including the minimum and maximum values) of the conditioned variable (\code{var1}), denoted as Emin and Emax. This output allows the function to further examine the conditional effect statistically in two ways. One is to examine if the distribution of \eqn{Emax - Emin} covers zero. The other is to directly compare Emin and Emax through statistical tools for distributional comparisons. Users can choose either method by setting the argument \code{stats_cp} to "ci" or "ks".
101 | \itemize{
102 | \item "ci" provides the confidence interval of the difference of \eqn{Emax - Emin}. An interval including 0 suggests no statistical difference before and after the conditional effect is applied, and vise versa.
103 | \item "ks" presents the result of a two-sample Kolmogorov-Smirnov test of the simulated distributions of Emin and Emax. The output includes a D statistics and a p-value of the null hypothesis that the two distributions come from the same distribution at the 0.05 level.
104 | }
105 |
106 | See an illustration in the package vignette.
107 | }
108 |
--------------------------------------------------------------------------------
/docs/CONDUCT.html:
--------------------------------------------------------------------------------
1 |
2 | Contributor Code of Conduct • interplot
6 |
7 |
8 |
9 |
50 |
51 |
52 |
53 |
54 |
55 |
56 | Contributor Code of Conduct
57 |
58 |
59 |
60 |
61 | As contributors and maintainers of this project, we pledge to respect all people who contribute through reporting issues, posting feature requests, updating documentation, submitting pull requests or patches, and other activities.
62 | We are committed to making participation in this project a harassment-free experience for everyone, regardless of level of experience, gender, gender identity and expression, sexual orientation, disability, personal appearance, body size, race, ethnicity, age, or religion.
63 | Examples of unacceptable behavior by participants include the use of sexual language or imagery, derogatory comments or personal attacks, trolling, public or private harassment, insults, or other unprofessional conduct.
64 | Project maintainers have the right and responsibility to remove, edit, or reject comments, commits, code, wiki edits, issues, and other contributions that are not aligned to this Code of Conduct. Project maintainers who do not follow the Code of Conduct may be removed from the project team.
65 | Instances of abusive, harassing, or otherwise unacceptable behavior may be reported by opening an issue or contacting one or more of the project maintainers.
66 | This Code of Conduct is adapted from the Contributor Covenant (http:contributor-covenant.org), version 1.0.0, available at http://contributor-covenant.org/version/1/0/0/
67 |
68 |
69 |
70 |
71 |
74 |
75 |
76 |
77 |
78 |
79 |
88 |
89 |
90 |
91 |
92 |
93 |
94 |
95 |
96 |
--------------------------------------------------------------------------------
/man/interplot.default.Rd:
--------------------------------------------------------------------------------
1 | % Generated by roxygen2: do not edit by hand
2 | % Please edit documentation in R/Interplot_default.R
3 | \name{interplot.default}
4 | \alias{interplot.default}
5 | \title{Plot Conditional Coefficients in (Generalized) Linear Models with Interaction Terms}
6 | \source{
7 | Benjamini, Yoav, and Yosef Hochberg. 1995. "Controlling the False
8 | Discovery Rate: A Practical and Powerful Approach to Multiple Testing".
9 | Journal of the Royal Statistical Society, Series B 57(1): 289--300.
10 |
11 | Brambor, Thomas, William Roberts Clark, and Matt Golder.
12 | "Understanding interaction models: Improving empirical analyses". Political
13 | Analysis 14.1 (2006): 63-82.
14 |
15 | Esarey, Justin, and Jane Lawrence Sumner. 2015. "Marginal Effects in
16 | Interaction Models: Determining and Controlling the False Positive Rate".
17 | URL: \url{http://jee3.web.rice.edu/interaction-overconfidence.pdf}.
18 | }
19 | \usage{
20 | \method{interplot}{default}(
21 | m,
22 | var1,
23 | var2,
24 | plot = TRUE,
25 | steps = NULL,
26 | ci = 0.95,
27 | adjCI = FALSE,
28 | hist = FALSE,
29 | var2_dt = NA,
30 | predPro = FALSE,
31 | var2_vals = NULL,
32 | point = FALSE,
33 | sims = 1000,
34 | xmin = NA,
35 | xmax = NA,
36 | ercolor = NA,
37 | esize = 0.5,
38 | ralpha = 0.5,
39 | rfill = "grey70",
40 | stats_cp = "none",
41 | txt_caption = NULL,
42 | facet_labs = NULL,
43 | ...
44 | )
45 | }
46 | \arguments{
47 | \item{m}{A model object including an interaction term, or, alternately, a data frame recording conditional coefficients.}
48 |
49 | \item{var1}{The name (as a string) of the variable of interest in the interaction term; its conditional coefficient estimates will be plotted.}
50 |
51 | \item{var2}{The name (as a string) of the other variable in the interaction term.}
52 |
53 | \item{plot}{A logical value indicating whether the output is a plot or a dataframe including the conditional coefficient estimates of var1, their upper and lower bounds, and the corresponding values of var2.}
54 |
55 | \item{steps}{Desired length of the sequence. A non-negative number, which for seq and seq.int will be rounded up if fractional. The default is 100 or the unique categories in the \code{var2} (when it is less than 100. Also see \code{\link{unique}}).}
56 |
57 | \item{ci}{A numeric value defining the confidence intervals. The default value is 95\% (0.95).}
58 |
59 | \item{adjCI}{A logical value indication if applying the adjustment of confidence intervals to control the false discovery rate following the Esarey and Sumner (2017) procedure. (See also Benjamini and Hochberg 1995.) The default is FALSE; the plot presents the confidence intervals suggested by Brambor, Clark, and Golder (2006).}
60 |
61 | \item{hist}{A logical value indicating if there is a histogram of `var2` added at the bottom of the conditional effect plot.}
62 |
63 | \item{var2_dt}{A numerical value indicating the frequency distribution of `var2`. It is only used when `hist == TRUE`. When the object is a model, the default is the distribution of `var2` of the model.}
64 |
65 | \item{predPro}{A logical value with default of `FALSE`. When the `m` is an object of class `glm` and the argument is set to `TRUE`, the function will plot predicted probabilities at the values given by `var2_vals`.}
66 |
67 | \item{var2_vals}{A numerical value indicating the values the predicted probabilities are estimated, when `predPro` is `TRUE`.}
68 |
69 | \item{point}{A logical value determining the format of plot. By default, the function produces a line plot when var2 takes on ten or more distinct values and a point (dot-and-whisker) plot otherwise; option TRUE forces a point plot.}
70 |
71 | \item{sims}{Number of independent simulation draws used to calculate upper and lower bounds of coefficient estimates: lower values run faster; higher values produce smoother curves.}
72 |
73 | \item{xmin}{A numerical value indicating the minimum value shown of x shown in the graph. Rarely used.}
74 |
75 | \item{xmax}{A numerical value indicating the maximum value shown of x shown in the graph. Rarely used.}
76 |
77 | \item{ercolor}{A character value indicating the outline color of the whisker or ribbon.}
78 |
79 | \item{esize}{A numerical value indicating the size of the whisker or ribbon.}
80 |
81 | \item{ralpha}{A numerical value indicating the transparency of the ribbon.}
82 |
83 | \item{rfill}{A character value indicating the filling color of the ribbon.}
84 |
85 | \item{stats_cp}{A character value indicating what statistics to present as the plot note. Three options are available: "none", "ci", and "ks". The default is "none". See the Details for more information.}
86 |
87 | \item{txt_caption}{A character string to add a note for the plot, a value will sending to \code{ggplot2::labs(caption = txt_caption))}.}
88 |
89 | \item{facet_labs}{An optional character vector of facet labels to be used when plotting an interaction with a factor variable.}
90 |
91 | \item{...}{Other ggplot aesthetics arguments for points in the dot-whisker plot or lines in the line-ribbon plots. Not currently used.}
92 | }
93 | \value{
94 | The function returns a \code{ggplot} object.
95 | }
96 | \description{
97 | \code{interplot.default} is a method to calculate conditional coefficient estimates from the results of (generalized) linear regression models with interaction terms.
98 | }
99 | \details{
100 | \code{interplot.default} is a S3 method from the \code{interplot}. It works on two classes of objects:
101 | \itemize{
102 | \item Ordinary linear models (object class: \code{lm});
103 | \item Generalized linear models (object class: \code{glm}).
104 | }
105 |
106 | Because the output function is based on \code{\link[ggplot2]{ggplot}}, any additional arguments and layers supported by \code{ggplot2} can be added with the \code{+}.
107 |
108 | \code{interplot} visualizes the conditional effect based on simulated marginal effects. The simulation provides a probabilistic distribution of moderation effect of the conditioning variable (\code{var2}) at every preset values (including the minimum and maximum values) of the conditioned variable (\code{var1}), denoted as Emin and Emax. This output allows the function to further examine the conditional effect statistically in two ways. One is to examine if the distribution of \eqn{Emax - Emin} covers zero. The other is to directly compare Emin and Emax through statistical tools for distributional comparisons. Users can choose either method by setting the argument \code{stats_cp} to "ci" or "ks".
109 | \itemize{
110 | \item "ci" provides the confidence interval of the difference of \eqn{Emax - Emin}. An interval including 0 suggests no statistical difference before and after the conditional effect is applied, and vise versa.
111 | \item "ks" presents the result of a two-sample Kolmogorov-Smirnov test of the simulated distributions of Emin and Emax. The output includes a D statistics and a p-value of the null hypothesis that the two distributions come from the same distribution at the 0.05 level.
112 | }
113 |
114 | See an illustration in the package vignette.
115 | }
116 |
--------------------------------------------------------------------------------
/docs/reference/index.html:
--------------------------------------------------------------------------------
1 |
2 | Function reference • interplot
6 |
7 |
8 |
9 |
50 |
51 |
52 |
53 |
54 |
55 |
56 | Reference
57 |
58 |
59 |
60 | All functions
61 |
62 |
63 |
64 |
65 |
66 | Plot Conditional Coefficients in (Generalized) Linear Models with Interaction Terms
68 |
69 |
70 | Plot Conditional Coefficients in Mixed-Effects Models with Interaction Terms
72 |
73 |
74 | Plot Conditional Coefficients in (Generalized) Linear Models with Imputed Data and Interaction Terms
76 |
77 |
78 | Plot Conditional Coefficients in Mixed-Effects Models with Imputed Data and Interaction Terms
80 |
81 |
82 | Plot Conditional Coefficients in Models with Interaction Terms
84 |
85 |
86 | Plot Conditional Coefficients of a Variable in an Interaction Term
88 |
89 |
92 |
93 |
94 |
95 |
104 |
105 |
106 |
107 |
108 |
109 |
110 |
111 |
112 |
--------------------------------------------------------------------------------
/man/interplot.lmmi.Rd:
--------------------------------------------------------------------------------
1 | % Generated by roxygen2: do not edit by hand
2 | % Please edit documentation in R/Interplot_mi.R
3 | \name{interplot.lmmi}
4 | \alias{interplot.lmmi}
5 | \title{Plot Conditional Coefficients in (Generalized) Linear Models with Imputed Data and Interaction Terms}
6 | \source{
7 | Benjamini, Yoav, and Yosef Hochberg. 1995. "Controlling the False
8 | Discovery Rate: A Practical and Powerful Approach to Multiple Testing".
9 | Journal of the Royal Statistical Society, Series B 57(1): 289--300.
10 |
11 | Brambor, Thomas, William Roberts Clark, and Matt Golder.
12 | "Understanding interaction models: Improving empirical analyses". Political
13 | Analysis 14.1 (2006): 63-82.
14 |
15 | Esarey, Justin, and Jane Lawrence Sumner. 2015. "Marginal Effects in
16 | Interaction Models: Determining and Controlling the False Positive Rate".
17 | URL: \url{http://jee3.web.rice.edu/interaction-overconfidence.pdf}.
18 | }
19 | \usage{
20 | \method{interplot}{lmmi}(
21 | m,
22 | var1,
23 | var2,
24 | plot = TRUE,
25 | steps = NULL,
26 | ci = 0.95,
27 | adjCI = FALSE,
28 | hist = FALSE,
29 | var2_dt = NA,
30 | predPro = FALSE,
31 | var2_vals = NULL,
32 | point = FALSE,
33 | sims = 5000,
34 | xmin = NA,
35 | xmax = NA,
36 | ercolor = NA,
37 | esize = 0.5,
38 | ralpha = 0.5,
39 | rfill = "grey70",
40 | stats_cp = "none",
41 | txt_caption = NULL,
42 | facet_labs = NULL,
43 | ...
44 | )
45 | }
46 | \arguments{
47 | \item{m}{A model object including an interaction term, or, alternately, a data frame recording conditional coefficients.}
48 |
49 | \item{var1}{The name (as a string) of the variable of interest in the interaction term; its conditional coefficient estimates will be plotted.}
50 |
51 | \item{var2}{The name (as a string) of the other variable in the interaction term.}
52 |
53 | \item{plot}{A logical value indicating whether the output is a plot or a dataframe including the conditional coefficient estimates of var1, their upper and lower bounds, and the corresponding values of var2.}
54 |
55 | \item{steps}{Desired length of the sequence. A non-negative number, which for seq and seq.int will be rounded up if fractional. The default is 100 or the unique categories in the \code{var2} (when it is less than 100. Also see \code{\link{unique}}).}
56 |
57 | \item{ci}{A numeric value defining the confidence intervals. The default value is 95\% (0.95).}
58 |
59 | \item{adjCI}{A logical value indication if applying the adjustment of confidence intervals to control the false discovery rate following the Esarey and Sumner (2017) procedure. (See also Benjamini and Hochberg 1995.) The default is FALSE; the plot presents the confidence intervals suggested by Brambor, Clark, and Golder (2006).}
60 |
61 | \item{hist}{A logical value indicating if there is a histogram of `var2` added at the bottom of the conditional effect plot.}
62 |
63 | \item{var2_dt}{A numerical value indicating the frequency distribution of `var2`. It is only used when `hist == TRUE`. When the object is a model, the default is the distribution of `var2` of the model.}
64 |
65 | \item{predPro}{A logical value with default of `FALSE`. When the `m` is an object of class `glm` and the argument is set to `TRUE`, the function will plot predicted probabilities at the values given by `var2_vals`.}
66 |
67 | \item{var2_vals}{A numerical value indicating the values the predicted probabilities are estimated, when `predPro` is `TRUE`.}
68 |
69 | \item{point}{A logical value determining the format of plot. By default, the function produces a line plot when var2 takes on ten or more distinct values and a point (dot-and-whisker) plot otherwise; option TRUE forces a point plot.}
70 |
71 | \item{sims}{Number of independent simulation draws used to calculate upper and lower bounds of coefficient estimates: lower values run faster; higher values produce smoother curves.}
72 |
73 | \item{xmin}{A numerical value indicating the minimum value shown of x shown in the graph. Rarely used.}
74 |
75 | \item{xmax}{A numerical value indicating the maximum value shown of x shown in the graph. Rarely used.}
76 |
77 | \item{ercolor}{A character value indicating the outline color of the whisker or ribbon.}
78 |
79 | \item{esize}{A numerical value indicating the size of the whisker or ribbon.}
80 |
81 | \item{ralpha}{A numerical value indicating the transparency of the ribbon.}
82 |
83 | \item{rfill}{A character value indicating the filling color of the ribbon.}
84 |
85 | \item{stats_cp}{A character value indicating what statistics to present as the plot note. Three options are available: "none", "ci", and "ks". The default is "none". See the Details for more information.}
86 |
87 | \item{txt_caption}{A character string to add a note for the plot, a value will sending to \code{ggplot2::labs(caption = txt_caption))}.}
88 |
89 | \item{facet_labs}{An optional character vector of facet labels to be used when plotting an interaction with a factor variable.}
90 |
91 | \item{...}{Other ggplot aesthetics arguments for points in the dot-whisker plot or lines in the line-ribbon plots. Not currently used.}
92 | }
93 | \value{
94 | The function returns a \code{ggplot} object.
95 | }
96 | \description{
97 | \code{interplot.mi} is a method to calculate conditional coefficient estimates from the results of (generalized) linear regression models with interaction terms and multiply imputed data.
98 | }
99 | \details{
100 | \code{interplot.lmmi} and \code{interplot.glmmi} are S3 methods from the \code{interplot}. This function can work on interactions from results in the class of \code{list} generated by \code{mitools}.
101 |
102 | Because the output function is based on \code{\link[ggplot2]{ggplot}}, any additional arguments and layers supported by \code{ggplot2} can be added with the \code{+}.
103 |
104 | \code{interplot} visualizes the conditional effect based on simulated marginal effects. The simulation provides a probabilistic distribution of moderation effect of the conditioning variable (\code{var2}) at every preset values (including the minimum and maximum values) of the conditioned variable (\code{var1}), denoted as Emin and Emax. This output allows the function to further examine the conditional effect statistically in two ways. One is to examine if the distribution of \eqn{Emax - Emin} covers zero. The other is to directly compare Emin and Emax through statistical tools for distributional comparisons. Users can choose either method by setting the argument \code{stats_cp} to "ci" or "ks".
105 | \itemize{
106 | \item "ci" provides the confidence interval of the difference of \eqn{Emax - Emin}. An interval including 0 suggests no statistical difference before and after the conditional effect is applied, and vise versa.
107 | \item "ks" presents the result of a two-sample Kolmogorov-Smirnov test of the simulated distributions of Emin and Emax. The output includes a D statistics and a p-value of the null hypothesis that the two distributions come from the same distribution at the 0.05 level.
108 | }
109 |
110 |
111 | See an illustration in the package vignette.
112 | }
113 | \examples{
114 |
115 | library(interplot)
116 | library(mitools)
117 |
118 | data(smi)
119 | model1 <- with(smi, glm(drinkreg ~ wave * sex, family = binomial()))
120 |
121 | interplot(model1, var1 = "sex", var2 = "wave")
122 |
123 | }
124 |
--------------------------------------------------------------------------------
/vignettes/vignette.bib:
--------------------------------------------------------------------------------
1 | % Encoding: UTF-8
2 |
3 | @Article{Brambor2006,
4 | author = {Brambor, Thomas and Clark, William Roberts and Golder, Matt},
5 | title = {Understanding Interaction Models: Improving Empirical Analyses},
6 | journal = {Political Analysis},
7 | year = {2006},
8 | volume = {14},
9 | number = {1},
10 | pages = {63--82},
11 | __markedentry = {[Sammo:6]},
12 | publisher = {SPM-PMSAPSA}
13 | }
14 |
15 | @Book{Aiken1991,
16 | title = {Multiple regression: Testing and interpreting interactionsMultiple Regression: Testing and Interpreting Interactions},
17 | publisher = {Sage},
18 | year = {1991},
19 | author = {Aiken, Leona S and West, Stephen G and Reno, Raymond R}
20 | }
21 |
22 | @Article{WrightJr1976,
23 | author = {Wright Jr, Gerald C},
24 | title = {Linear Models for Evaluating Conditional Relationships},
25 | journal = {American Journal of Political Science},
26 | year = {1976},
27 | pages = {349--373},
28 | __markedentry = {[Sammo:]},
29 | publisher = {JSTOR}
30 | }
31 |
32 | @Article{Berry2012,
33 | author = {Berry, William D and Golder, Matt and Milton, Daniel},
34 | title = {Improving Tests of Theories Positing Interaction},
35 | journal = {Journal of Politics},
36 | year = {2012},
37 | volume = {74},
38 | number = {3},
39 | pages = {653--671},
40 | publisher = {Cambridge Univ Press}
41 | }
42 |
43 | @Unpublished{Hainmueller2016,
44 | author = {Hainmueller, Jens and Mummolo, Jonathan and Xu, Yiqing},
45 | title = {How Much Should We Trust Estimates from Multiplicative Interaction Models? Simple Tools to Improve Empirical Practice},
46 | year = {2016},
47 | timestamp = {2017-07-20},
48 | url = {https://www.ssrn.com/abstract=2739221},
49 | }
50 |
51 | @Article{Clark2006,
52 | author = {Clark, William Roberts and Gilligan, Michael J and Golder, Matt},
53 | title = {A Simple Multivariate Test for Asymmetric Hypotheses},
54 | journal = {Political Analysis},
55 | year = {2006},
56 | volume = {14},
57 | number = {3},
58 | pages = {311--331},
59 | publisher = {SPM-PMSAPSA}
60 | }
61 |
62 | @Article{Berry2016,
63 | author = {Berry, William D and DeMeritt, Jacqueline HR and Esarey, Justin},
64 | title = {Bias and Overconfidence in Parametric Models of Interactive Processes},
65 | journal = {American Journal of Political Science},
66 | year = {2016},
67 | volume = {60},
68 | number = {2},
69 | pages = {521--539},
70 | publisher = {Wiley Online Library},
71 | }
72 |
73 | @Unpublished{Esarey2017,
74 | author = {Esarey, Justin and Sumner, Jane Lawrence},
75 | title = {Marginal Effects in Interaction Models: Determining and Controlling the False Positive Rate},
76 | year = {2017},
77 | timestamp = {2017-07-20},
78 | url = {http://jee3.web.rice.edu/interaction-overconfidence.pdf},
79 | }
80 |
81 | @Article{Braumoeller2004,
82 | author = {Braumoeller, Bear F},
83 | title = {Hypothesis Testing and Multiplicative Interaction Terms},
84 | journal = {International Organization},
85 | year = {2004},
86 | volume = {58},
87 | number = {4},
88 | pages = {807--820},
89 | publisher = {Cambridge University Press},
90 | }
91 |
92 | @Article{Benjamini1995,
93 | author = {Benjamini, Yoav and Hochberg, Yosef},
94 | title = {Controlling the False Discovery Rate: A Practical and Powerful Approach to Multiple Testing},
95 | journal = {Journal of the Royal Statistical Society. Series B (Methodological)},
96 | year = {1995},
97 | pages = {289--300},
98 | publisher = {JSTOR},
99 | }
100 |
101 | @Article{Hanmer2013,
102 | author = {Hanmer, Michael J. and Ozan Kalkan, Kerem},
103 | title = {Behind the Curve: Clarifying the Best Approach to Calculating Predicted Probabilities and Marginal Effects from Limited Dependent Variable Models},
104 | journal = {American Journal of Political Science},
105 | year = {2013},
106 | volume = {57},
107 | number = {1},
108 | pages = {263--277},
109 | issn = {1540-5907},
110 | abstract = {Models designed for limited dependent variables are increasingly common in political science. Researchers estimating such models often give little attention to the coefficient estimates and instead focus on marginal effects, predicted probabilities, predicted counts, etc. Since the models are nonlinear, the estimated effects are sensitive to how one generates the predictions. The most common approach involves estimating the effect for the “average case.” But this approach creates a weaker connection between the results and the larger goals of the research enterprise and is thus less preferable than the observed-value approach. That is, rather than seeking to understand the effect for the average case, the goal is to obtain an estimate of the average effect in the population. In addition to the theoretical argument in favor of the observed-value approach, we illustrate via an empirical example and Monte Carlo simulations that the two approaches can produce substantively different results.},
111 | doi = {10.1111/j.1540-5907.2012.00602.x},
112 | publisher = {Blackwell Publishing Inc},
113 | timestamp = {2017-08-06},
114 | url = {http://dx.doi.org/10.1111/j.1540-5907.2012.00602.x},
115 | }
116 |
117 | @Article{Solt2016,
118 | author = {Solt, Frederick and Hu, Yue and Hudson, Kevan and Song, Jungmin and Yu, Dong Erico},
119 | title = {Economic Inequality and Belief in Meritocracy in the United States},
120 | journal = {Research \& Politics},
121 | year = {2016},
122 | volume = {3},
123 | number = {4},
124 | pages = {1--7},
125 | abstract = {How does the context of income inequality in which people live affect their belief in meritocracy, the ability to get ahead through hard work? A prominent recent study by Newman, Johnston, and Lown argues that, consistent with the conflict theory, exposure to higher levels of local income inequality leads lower-income people to become more likely to reject—and higher-income people to become more likely to accept—the dominant United States ideology of meritocracy. Here, we show that this conclusion is not supported by the studys own reported results and that even these results depend on pooling three distinctly different measures of meritocracy into a single analysis. We then demonstrate that analysis of a larger and more representative survey employing a single consistent measure of the dependent variable yields the opposite conclusion. Consistent with the relative power theory, among those with lower incomes, local contexts of greater inequality are associated with more widespread belief that people can get ahead if they are willing to work hard.},
126 | date = {2017-06-14},
127 | doi = {10.1177/2053168016672101},
128 | eprint = {http://rap.sagepub.com/content/3/4/2053168016672101.full.pdf},
129 | publisher = {SAGE Publications},
130 | timestamp = {2017-06-14},
131 | url = {http://rap.sagepub.com/content/3/4/2053168016672101},
132 | urldate = {2017-06-14},
133 | }
134 |
135 | @Comment{jabref-meta: databaseType:bibtex;}
136 |
--------------------------------------------------------------------------------
/dev/jss/jss_interplot.bib:
--------------------------------------------------------------------------------
1 | % Encoding: UTF-8
2 |
3 | @Article{Brambor2006,
4 | author = {Brambor, Thomas and Clark, William Roberts and Golder, Matt},
5 | title = {Understanding Interaction Models: Improving Empirical Analyses},
6 | journal = {Political Analysis},
7 | year = {2006},
8 | volume = {14},
9 | number = {1},
10 | pages = {63--82},
11 | __markedentry = {[Sammo:6]},
12 | publisher = {SPM-PMSAPSA}
13 | }
14 |
15 | @Book{Aiken1991,
16 | title = {Multiple regression: Testing and interpreting interactionsMultiple Regression: Testing and Interpreting Interactions},
17 | publisher = {Sage},
18 | year = {1991},
19 | author = {Aiken, Leona S and West, Stephen G and Reno, Raymond R}
20 | }
21 |
22 | @Article{WrightJr1976,
23 | author = {Wright Jr, Gerald C},
24 | title = {Linear Models for Evaluating Conditional Relationships},
25 | journal = {American Journal of Political Science},
26 | year = {1976},
27 | pages = {349--373},
28 | __markedentry = {[Sammo:]},
29 | publisher = {JSTOR}
30 | }
31 |
32 | @Article{Berry2012,
33 | author = {Berry, William D and Golder, Matt and Milton, Daniel},
34 | title = {Improving Tests of Theories Positing Interaction},
35 | journal = {Journal of Politics},
36 | year = {2012},
37 | volume = {74},
38 | number = {3},
39 | pages = {653--671},
40 | publisher = {Cambridge University Press}
41 | }
42 |
43 | @Article{Hainmueller2019,
44 | author = {Hainmueller, Jens and Mummolo, Jonathan and Xu, Yiqing},
45 | title = {How Much Should We Trust Estimates from Multiplicative Interaction Models? Simple Tools to Improve Empirical Practice},
46 | journal = {Political Analysis},
47 | year = {2012},
48 | volume = {27},
49 | number = {2},
50 | pages = {163-192},
51 | publisher = {SAGE}
52 | }
53 |
54 | @Article{Clark2006,
55 | author = {Clark, William Roberts and Gilligan, Michael J and Golder, Matt},
56 | title = {A Simple Multivariate Test for Asymmetric Hypotheses},
57 | journal = {Political Analysis},
58 | year = {2006},
59 | volume = {14},
60 | number = {3},
61 | pages = {311--331},
62 | publisher = {SPM-PMSAPSA}
63 | }
64 |
65 | @Article{Berry2016,
66 | author = {Berry, William D and DeMeritt, Jacqueline HR and Esarey, Justin},
67 | title = {Bias and Overconfidence in Parametric Models of Interactive Processes},
68 | journal = {American Journal of Political Science},
69 | year = {2016},
70 | volume = {60},
71 | number = {2},
72 | pages = {521--539},
73 | publisher = {Wiley Online Library},
74 | }
75 |
76 | @Article{Esarey2018,
77 | author = {Esarey, Justin and Sumner, Jane Lawrence},
78 | title = {Marginal Effects in Interaction Models: Determining and Controlling the False Positive Rate},
79 | journal = {Comparative Political Studies},
80 | year = {2016},
81 | volume = {51},
82 | number = {9},
83 | pages = {1144–1176},
84 | publisher = {JSTOR},
85 | }
86 |
87 | @Article{Braumoeller2004,
88 | author = {Braumoeller, Bear F},
89 | title = {Hypothesis Testing and Multiplicative Interaction Terms},
90 | journal = {International Organization},
91 | year = {2004},
92 | volume = {58},
93 | number = {4},
94 | pages = {807--820},
95 | publisher = {Cambridge University Press},
96 | }
97 |
98 | @Article{Benjamini1995,
99 | author = {Benjamini, Yoav and Hochberg, Yosef},
100 | title = {Controlling the False Discovery Rate: A Practical and Powerful Approach to Multiple Testing},
101 | journal = {Journal of the Royal Statistical Society. Series B (Methodological)},
102 | year = {1995},
103 | pages = {289--300},
104 | publisher = {JSTOR},
105 | }
106 |
107 | @Article{Hanmer2013,
108 | author = {Hanmer, Michael J. and Ozan Kalkan, Kerem},
109 | title = {Behind the Curve: Clarifying the Best Approach to Calculating Predicted Probabilities and Marginal Effects from Limited Dependent Variable Models},
110 | journal = {American Journal of Political Science},
111 | year = {2013},
112 | volume = {57},
113 | number = {1},
114 | pages = {263--277},
115 | issn = {1540-5907},
116 | abstract = {Models designed for limited dependent variables are increasingly common in political science. Researchers estimating such models often give little attention to the coefficient estimates and instead focus on marginal effects, predicted probabilities, predicted counts, etc. Since the models are nonlinear, the estimated effects are sensitive to how one generates the predictions. The most common approach involves estimating the effect for the “average case.” But this approach creates a weaker connection between the results and the larger goals of the research enterprise and is thus less preferable than the observed-value approach. That is, rather than seeking to understand the effect for the average case, the goal is to obtain an estimate of the average effect in the population. In addition to the theoretical argument in favor of the observed-value approach, we illustrate via an empirical example and Monte Carlo simulations that the two approaches can produce substantively different results.},
117 | doi = {10.1111/j.1540-5907.2012.00602.x},
118 | publisher = {Blackwell Publishing Inc},
119 | timestamp = {2017-08-06},
120 | url = {http://dx.doi.org/10.1111/j.1540-5907.2012.00602.x},
121 | }
122 |
123 | @Article{Solt2016,
124 | author = {Solt, Frederick and Hu, Yue and Hudson, Kevan and Song, Jungmin and Yu, Dong Erico},
125 | title = {Economic Inequality and Belief in Meritocracy in the United States},
126 | journal = {Research \& Politics},
127 | year = {2016},
128 | volume = {3},
129 | number = {4},
130 | pages = {1--7},
131 | abstract = {How does the context of income inequality in which people live affect their belief in meritocracy, the ability to get ahead through hard work? A prominent recent study by Newman, Johnston, and Lown argues that, consistent with the conflict theory, exposure to higher levels of local income inequality leads lower-income people to become more likely to reject—and higher-income people to become more likely to accept—the dominant United States ideology of meritocracy. Here, we show that this conclusion is not supported by the studys own reported results and that even these results depend on pooling three distinctly different measures of meritocracy into a single analysis. We then demonstrate that analysis of a larger and more representative survey employing a single consistent measure of the dependent variable yields the opposite conclusion. Consistent with the relative power theory, among those with lower incomes, local contexts of greater inequality are associated with more widespread belief that people can get ahead if they are willing to work hard.},
132 | date = {2017-06-14},
133 | doi = {10.1177/2053168016672101},
134 | eprint = {http://rap.sagepub.com/content/3/4/2053168016672101.full.pdf},
135 | publisher = {SAGE Publications},
136 | timestamp = {2017-06-14},
137 | url = {http://rap.sagepub.com/content/3/4/2053168016672101},
138 | urldate = {2017-06-14},
139 | }
140 |
141 | @Comment{jabref-meta: databaseType:bibtex;}
142 |
143 | @Article{Wright1976,
144 | title = {Linear models for evaluating conditional relationships},
145 | author = {Wright, Gerald C.},
146 | year = {1976},
147 | date = {1976},
148 | journal = {American Journal of Political Science},
149 | pages = {349},
150 | volume = {20},
151 | number = {2},
152 | doi = {10.2307/2110649},
153 | url = {https://www.jstor.org/stable/2110649?origin=crossref},
154 | langid = {en}
155 | }
156 |
157 |
158 |
--------------------------------------------------------------------------------
/docs/index.html:
--------------------------------------------------------------------------------
1 |
2 |
3 |
4 |
5 |
6 |
7 |
8 | Plot the Effects of Variables in Interaction Terms • interplot
9 |
10 |
11 |
12 |
13 |
14 |
15 |
17 |
21 |
22 |
23 |
24 |
25 |
26 |
75 |
76 |
77 |
78 |
79 |
80 |
81 |
82 | pkgdown
83 |
84 |
85 |
86 |
88 | interplot is a tool for plotting the conditional coefficients (“marginal effects”) of variables included in multiplicative interaction terms. The function plots the changes in the coefficient of one variable in a two-way interaction term conditional on the value of the other included variable. The plot also includes simulated 95% confidential intervals of these coefficients.
89 | To install:
90 |
91 | - the latest released version:
install.packages("interplot").
92 | - the latest developing version:
remotes::install_github("sammo3182/interplot").
93 |
94 | More details are available at:
95 | http://cran.r-project.org/web/packages/interplot/vignettes/interplot-vignette.html
96 | Please note that this project is released with a Contributor Code of Conduct. By participating in this project you agree to abide by its terms.
97 |
98 |
99 |
100 |
101 |
140 |
141 |
142 |
143 |
154 |
155 |
156 |
157 |
158 |
159 |
160 |
161 |
162 |
163 |
--------------------------------------------------------------------------------
/man/interplot.Rd:
--------------------------------------------------------------------------------
1 | % Generated by roxygen2: do not edit by hand
2 | % Please edit documentation in R/Interplot.R
3 | \name{interplot}
4 | \alias{interplot}
5 | \title{Plot Conditional Coefficients of a Variable in an Interaction Term}
6 | \source{
7 | Benjamini, Yoav, and Yosef Hochberg. 1995. "Controlling the False
8 | Discovery Rate: A Practical and Powerful Approach to Multiple Testing".
9 | Journal of the Royal Statistical Society, Series B 57(1): 289--300.
10 |
11 | Brambor, Thomas, William Roberts Clark, and Matt Golder.
12 | "Understanding interaction models: Improving empirical analyses". Political
13 | Analysis 14.1 (2006): 63-82.
14 |
15 | Esarey, Justin, and Jane Lawrence Sumner. 2015. "Marginal Effects in
16 | Interaction Models: Determining and Controlling the False Positive Rate".
17 | URL: \url{https://jee3.web.rice.edu/interaction-overconfidence.pdf}.
18 | }
19 | \usage{
20 | interplot(
21 | m,
22 | var1,
23 | var2,
24 | plot = TRUE,
25 | steps = NULL,
26 | ci = 0.95,
27 | adjCI = FALSE,
28 | hist = FALSE,
29 | var2_dt = NA,
30 | predPro = FALSE,
31 | var2_vals = NULL,
32 | point = FALSE,
33 | sims = 1000,
34 | xmin = NA,
35 | xmax = NA,
36 | ercolor = NA,
37 | esize = 0.5,
38 | ralpha = 0.5,
39 | rfill = "grey70",
40 | stats_cp = "none",
41 | txt_caption = NULL,
42 | ...
43 | )
44 | }
45 | \arguments{
46 | \item{m}{A model object including an interaction term, or, alternately, a data frame generated by an earlier call to interplot using the argument plot = FALSE.}
47 |
48 | \item{var1}{The name (as a string) of the variable of interest in the interaction term; its conditional coefficient estimates will be plotted.}
49 |
50 | \item{var2}{The name (as a string) of the other variable in the interaction term}
51 |
52 | \item{plot}{A logical value indicating whether the output is a plot or a dataframe including the conditional coefficient estimates of var1, their upper and lower bounds, and the corresponding values of var2.}
53 |
54 | \item{steps}{Desired length of the sequence. A non-negative number, which for seq and seq.int will be rounded up if fractional. The default is 100 or the unique categories in the \code{var2} (when it is less than 100. Also see \code{\link{unique}}).}
55 |
56 | \item{ci}{A numeric value defining the confidence intervals. The default value is 95\% (0.95).}
57 |
58 | \item{adjCI}{A logical value indication if applying the adjustment of confidence intervals to control the false discovery rate following the Esarey and Sumner (2017) procedure. (See also Benjamini and Hochberg 1995.) The default is FALSE; the plot presents the confidence intervals suggested by Brambor, Clark, and Golder (2006). The functions dealing with multilevel model outputs in this package do not equip with this argument, because there is the controversy on the accurate degrees of freedom for multilevel models, esp. when random effect is engaged and the degrees of freedom is a necessary information to conduct the CI adjustment. See e.g., https://stat.ethz.ch/pipermail/r-help/2006-May/094765.html and https://stat.ethz.ch/pipermail/r-sig-mixed-models/2008q1/000517.html.}
59 |
60 | \item{hist}{A logical value indicating if there is a histogram of `var2` added at the bottom of the conditional effect plot.}
61 |
62 | \item{var2_dt}{A numerical value indicating the frequency distribution of `var2`. It is only used when `hist == TRUE`. When the object is a model, the default is the distribution of `var2` of the model.}
63 |
64 | \item{predPro}{A logical value with default of `FALSE`. When the `m` is an object of class `glm` or `glmerMod` and the argument is set to `TRUE`, the function will plot predicted probabilities at the values given by `var2_vals`.}
65 |
66 | \item{var2_vals}{A numerical value indicating the values the predicted probabilities are estimated, when `predPro` is `TRUE`.}
67 |
68 | \item{point}{A logical value determining the format of plot. By default, the function produces a line plot when var2 takes on ten or more distinct values and a point (dot-and-whisker) plot otherwise; option TRUE forces a point plot.}
69 |
70 | \item{sims}{Number of independent simulation draws used to calculate upper and lower bounds of coefficient estimates: lower values run faster; higher values produce smoother curves.}
71 |
72 | \item{xmin}{A numerical value indicating the minimum value shown of x shown in the graph. Rarely used.}
73 |
74 | \item{xmax}{A numerical value indicating the maximum value shown of x shown in the graph. Rarely used.}
75 |
76 | \item{ercolor}{A character value indicating the outline color of the whisker or ribbon.}
77 |
78 | \item{esize}{A numerical value indicating the size of the whisker or ribbon.}
79 |
80 | \item{ralpha}{A numerical value indicating the transparency of the ribbon.}
81 |
82 | \item{rfill}{A character value indicating the filling color of the ribbon.}
83 |
84 | \item{stats_cp}{A character value indicating what statistics to present as the plot note. Three options are available: "none", "ci", and "ks". The default is "none". See the Details for more information.}
85 |
86 | \item{txt_caption}{A character string to add a note for the plot, a value will sending to \code{ggplot2::labs(caption = txt_caption))}.}
87 |
88 | \item{...}{Other ggplot aesthetics arguments for points in the dot-whisker plot or lines in the line-ribbon plots. Not currently used.}
89 | }
90 | \value{
91 | The function returns a \code{ggplot} object.
92 | }
93 | \description{
94 | \code{interplot} is a generic function to produce a plot of the coefficient estimates of one variable in a two-way interaction conditional on the values of the other variable in the interaction term. The function invokes particular \code{methods} which depend on the \code{\link{class}} of the first argument.
95 | }
96 | \details{
97 | \code{interplot} visualizes the changes in the coefficient of one term in a two-way interaction conditioned by the other term. In the current version, the function works with interactions in the following classes of models:
98 | \itemize{
99 | \item Ordinary linear models (object class: \code{lm});
100 | \item Generalized linear models (object class: \code{glm});
101 | \item Linear mixed-effects models (object class: \code{lmerMod});
102 | \item Generalized linear mixed-effects models (object class: \code{glmerMod});
103 | \item Ordinary linear models with imputed data (object class: \code{list});
104 | \item Generalized linear models with imputed data (object class: \code{list})
105 | \item Linear mixed-effects models with imputed data (object class: \code{list});
106 | \item Generalized linear mixed-effects models with imputed data (object class: \code{list}).
107 | }
108 |
109 | The examples below illustrate how methods invoked by this generic deal with different type of objects.
110 |
111 | Because the output function is based on \code{\link[ggplot2]{ggplot}}, any additional arguments and layers supported by \code{ggplot2} can be added with the \code{+}.
112 |
113 | \code{interplot} visualizes the conditional effect based on simulated marginal effects. The simulation provides a probabilistic distribution of moderation effect of the conditioning variable (\code{var2}) at every preset values (including the minimum and maximum values) of the conditioned variable (\code{var1}), denoted as Emin and Emax. This output allows the function to further examine the conditional effect statistically in two ways. One is to examine if the distribution of \eqn{Emax - Emin} covers zero. The other is to directly compare Emin and Emax through statistical tools for distributional comparisons. Users can choose either method by setting the argument \code{stats_cp} to "ci" or "ks".
114 | \itemize{
115 | \item "ci" provides the confidence interval of the difference of \eqn{Emax - Emin}. An interval including 0 suggests no statistical difference before and after the conditional effect is applied, and vise versa.
116 | \item "ks" presents the result of a two-sample Kolmogorov-Smirnov test of the simulated distributions of Emin and Emax. The output includes a D statistics and a p-value of the null hypothesis that the two distributions come from the same distribution at the 0.05 level.
117 | }
118 |
119 | See an illustration in the package vignette.
120 | }
121 | \examples{
122 | data(mtcars)
123 | m_cyl <- lm(mpg ~ wt * cyl, data = mtcars)
124 | library(interplot)
125 |
126 | # Plot interactions with a continous conditioning variable
127 | interplot(m = m_cyl, var1 = 'cyl', var2 = 'wt') +
128 | xlab('Automobile Weight (thousands lbs)') +
129 | ylab('Estimated Coefficient for Number of Cylinders') +
130 | ggtitle('Estimated Coefficient of Engine Cylinders\non Mileage by Automobile Weight') +
131 | theme(plot.title = element_text(face='bold'))
132 |
133 |
134 | # Plot interactions with a categorical conditioning variable
135 | interplot(m = m_cyl, var1 = 'wt', var2 = 'cyl') +
136 | xlab('Number of Cylinders') +
137 | ylab('Estimated Coefficient for Automobile Weight (thousands lbs)') +
138 | ggtitle('Estimated Coefficient of Automobile Weight \non Mileage by Engine Cylinders') +
139 | theme(plot.title = element_text(face='bold'))
140 |
141 | }
142 |
--------------------------------------------------------------------------------
/docs/pkgdown.css:
--------------------------------------------------------------------------------
1 | /* Sticky footer */
2 |
3 | /**
4 | * Basic idea: https://philipwalton.github.io/solved-by-flexbox/demos/sticky-footer/
5 | * Details: https://github.com/philipwalton/solved-by-flexbox/blob/master/assets/css/components/site.css
6 | *
7 | * .Site -> body > .container
8 | * .Site-content -> body > .container .row
9 | * .footer -> footer
10 | *
11 | * Key idea seems to be to ensure that .container and __all its parents__
12 | * have height set to 100%
13 | *
14 | */
15 |
16 | html, body {
17 | height: 100%;
18 | }
19 |
20 | body {
21 | position: relative;
22 | }
23 |
24 | body > .container {
25 | display: flex;
26 | height: 100%;
27 | flex-direction: column;
28 | }
29 |
30 | body > .container .row {
31 | flex: 1 0 auto;
32 | }
33 |
34 | footer {
35 | margin-top: 45px;
36 | padding: 35px 0 36px;
37 | border-top: 1px solid #e5e5e5;
38 | color: #666;
39 | display: flex;
40 | flex-shrink: 0;
41 | }
42 | footer p {
43 | margin-bottom: 0;
44 | }
45 | footer div {
46 | flex: 1;
47 | }
48 | footer .pkgdown {
49 | text-align: right;
50 | }
51 | footer p {
52 | margin-bottom: 0;
53 | }
54 |
55 | img.icon {
56 | float: right;
57 | }
58 |
59 | /* Ensure in-page images don't run outside their container */
60 | .contents img {
61 | max-width: 100%;
62 | height: auto;
63 | }
64 |
65 | /* Fix bug in bootstrap (only seen in firefox) */
66 | summary {
67 | display: list-item;
68 | }
69 |
70 | /* Typographic tweaking ---------------------------------*/
71 |
72 | .contents .page-header {
73 | margin-top: calc(-60px + 1em);
74 | }
75 |
76 | dd {
77 | margin-left: 3em;
78 | }
79 |
80 | /* Section anchors ---------------------------------*/
81 |
82 | a.anchor {
83 | display: none;
84 | margin-left: 5px;
85 | width: 20px;
86 | height: 20px;
87 |
88 | background-image: url(./link.svg);
89 | background-repeat: no-repeat;
90 | background-size: 20px 20px;
91 | background-position: center center;
92 | }
93 |
94 | h1:hover .anchor,
95 | h2:hover .anchor,
96 | h3:hover .anchor,
97 | h4:hover .anchor,
98 | h5:hover .anchor,
99 | h6:hover .anchor {
100 | display: inline-block;
101 | }
102 |
103 | /* Fixes for fixed navbar --------------------------*/
104 |
105 | .contents h1, .contents h2, .contents h3, .contents h4 {
106 | padding-top: 60px;
107 | margin-top: -40px;
108 | }
109 |
110 | /* Navbar submenu --------------------------*/
111 |
112 | .dropdown-submenu {
113 | position: relative;
114 | }
115 |
116 | .dropdown-submenu>.dropdown-menu {
117 | top: 0;
118 | left: 100%;
119 | margin-top: -6px;
120 | margin-left: -1px;
121 | border-radius: 0 6px 6px 6px;
122 | }
123 |
124 | .dropdown-submenu:hover>.dropdown-menu {
125 | display: block;
126 | }
127 |
128 | .dropdown-submenu>a:after {
129 | display: block;
130 | content: " ";
131 | float: right;
132 | width: 0;
133 | height: 0;
134 | border-color: transparent;
135 | border-style: solid;
136 | border-width: 5px 0 5px 5px;
137 | border-left-color: #cccccc;
138 | margin-top: 5px;
139 | margin-right: -10px;
140 | }
141 |
142 | .dropdown-submenu:hover>a:after {
143 | border-left-color: #ffffff;
144 | }
145 |
146 | .dropdown-submenu.pull-left {
147 | float: none;
148 | }
149 |
150 | .dropdown-submenu.pull-left>.dropdown-menu {
151 | left: -100%;
152 | margin-left: 10px;
153 | border-radius: 6px 0 6px 6px;
154 | }
155 |
156 | /* Sidebar --------------------------*/
157 |
158 | #pkgdown-sidebar {
159 | margin-top: 30px;
160 | position: -webkit-sticky;
161 | position: sticky;
162 | top: 70px;
163 | }
164 |
165 | #pkgdown-sidebar h2 {
166 | font-size: 1.5em;
167 | margin-top: 1em;
168 | }
169 |
170 | #pkgdown-sidebar h2:first-child {
171 | margin-top: 0;
172 | }
173 |
174 | #pkgdown-sidebar .list-unstyled li {
175 | margin-bottom: 0.5em;
176 | }
177 |
178 | /* bootstrap-toc tweaks ------------------------------------------------------*/
179 |
180 | /* All levels of nav */
181 |
182 | nav[data-toggle='toc'] .nav > li > a {
183 | padding: 4px 20px 4px 6px;
184 | font-size: 1.5rem;
185 | font-weight: 400;
186 | color: inherit;
187 | }
188 |
189 | nav[data-toggle='toc'] .nav > li > a:hover,
190 | nav[data-toggle='toc'] .nav > li > a:focus {
191 | padding-left: 5px;
192 | color: inherit;
193 | border-left: 1px solid #878787;
194 | }
195 |
196 | nav[data-toggle='toc'] .nav > .active > a,
197 | nav[data-toggle='toc'] .nav > .active:hover > a,
198 | nav[data-toggle='toc'] .nav > .active:focus > a {
199 | padding-left: 5px;
200 | font-size: 1.5rem;
201 | font-weight: 400;
202 | color: inherit;
203 | border-left: 2px solid #878787;
204 | }
205 |
206 | /* Nav: second level (shown on .active) */
207 |
208 | nav[data-toggle='toc'] .nav .nav {
209 | display: none; /* Hide by default, but at >768px, show it */
210 | padding-bottom: 10px;
211 | }
212 |
213 | nav[data-toggle='toc'] .nav .nav > li > a {
214 | padding-left: 16px;
215 | font-size: 1.35rem;
216 | }
217 |
218 | nav[data-toggle='toc'] .nav .nav > li > a:hover,
219 | nav[data-toggle='toc'] .nav .nav > li > a:focus {
220 | padding-left: 15px;
221 | }
222 |
223 | nav[data-toggle='toc'] .nav .nav > .active > a,
224 | nav[data-toggle='toc'] .nav .nav > .active:hover > a,
225 | nav[data-toggle='toc'] .nav .nav > .active:focus > a {
226 | padding-left: 15px;
227 | font-weight: 500;
228 | font-size: 1.35rem;
229 | }
230 |
231 | /* orcid ------------------------------------------------------------------- */
232 |
233 | .orcid {
234 | font-size: 16px;
235 | color: #A6CE39;
236 | /* margins are required by official ORCID trademark and display guidelines */
237 | margin-left:4px;
238 | margin-right:4px;
239 | vertical-align: middle;
240 | }
241 |
242 | /* Reference index & topics ----------------------------------------------- */
243 |
244 | .ref-index th {font-weight: normal;}
245 |
246 | .ref-index td {vertical-align: top; min-width: 100px}
247 | .ref-index .icon {width: 40px;}
248 | .ref-index .alias {width: 40%;}
249 | .ref-index-icons .alias {width: calc(40% - 40px);}
250 | .ref-index .title {width: 60%;}
251 |
252 | .ref-arguments th {text-align: right; padding-right: 10px;}
253 | .ref-arguments th, .ref-arguments td {vertical-align: top; min-width: 100px}
254 | .ref-arguments .name {width: 20%;}
255 | .ref-arguments .desc {width: 80%;}
256 |
257 | /* Nice scrolling for wide elements --------------------------------------- */
258 |
259 | table {
260 | display: block;
261 | overflow: auto;
262 | }
263 |
264 | /* Syntax highlighting ---------------------------------------------------- */
265 |
266 | pre, code, pre code {
267 | background-color: #f8f8f8;
268 | color: #333;
269 | }
270 | pre, pre code {
271 | white-space: pre-wrap;
272 | word-break: break-all;
273 | overflow-wrap: break-word;
274 | }
275 |
276 | pre {
277 | border: 1px solid #eee;
278 | }
279 |
280 | pre .img, pre .r-plt {
281 | margin: 5px 0;
282 | }
283 |
284 | pre .img img, pre .r-plt img {
285 | background-color: #fff;
286 | }
287 |
288 | code a, pre a {
289 | color: #375f84;
290 | }
291 |
292 | a.sourceLine:hover {
293 | text-decoration: none;
294 | }
295 |
296 | .fl {color: #1514b5;}
297 | .fu {color: #000000;} /* function */
298 | .ch,.st {color: #036a07;} /* string */
299 | .kw {color: #264D66;} /* keyword */
300 | .co {color: #888888;} /* comment */
301 |
302 | .error {font-weight: bolder;}
303 | .warning {font-weight: bolder;}
304 |
305 | /* Clipboard --------------------------*/
306 |
307 | .hasCopyButton {
308 | position: relative;
309 | }
310 |
311 | .btn-copy-ex {
312 | position: absolute;
313 | right: 0;
314 | top: 0;
315 | visibility: hidden;
316 | }
317 |
318 | .hasCopyButton:hover button.btn-copy-ex {
319 | visibility: visible;
320 | }
321 |
322 | /* headroom.js ------------------------ */
323 |
324 | .headroom {
325 | will-change: transform;
326 | transition: transform 200ms linear;
327 | }
328 | .headroom--pinned {
329 | transform: translateY(0%);
330 | }
331 | .headroom--unpinned {
332 | transform: translateY(-100%);
333 | }
334 |
335 | /* mark.js ----------------------------*/
336 |
337 | mark {
338 | background-color: rgba(255, 255, 51, 0.5);
339 | border-bottom: 2px solid rgba(255, 153, 51, 0.3);
340 | padding: 1px;
341 | }
342 |
343 | /* vertical spacing after htmlwidgets */
344 | .html-widget {
345 | margin-bottom: 10px;
346 | }
347 |
348 | /* fontawesome ------------------------ */
349 |
350 | .fab {
351 | font-family: "Font Awesome 5 Brands" !important;
352 | }
353 |
354 | /* don't display links in code chunks when printing */
355 | /* source: https://stackoverflow.com/a/10781533 */
356 | @media print {
357 | code a:link:after, code a:visited:after {
358 | content: "";
359 | }
360 | }
361 |
362 | /* Section anchors ---------------------------------
363 | Added in pandoc 2.11: https://github.com/jgm/pandoc-templates/commit/9904bf71
364 | */
365 |
366 | div.csl-bib-body { }
367 | div.csl-entry {
368 | clear: both;
369 | }
370 | .hanging-indent div.csl-entry {
371 | margin-left:2em;
372 | text-indent:-2em;
373 | }
374 | div.csl-left-margin {
375 | min-width:2em;
376 | float:left;
377 | }
378 | div.csl-right-inline {
379 | margin-left:2em;
380 | padding-left:1em;
381 | }
382 | div.csl-indent {
383 | margin-left: 2em;
384 | }
385 |
--------------------------------------------------------------------------------
/R/Interplot.R:
--------------------------------------------------------------------------------
1 | #' Plot Conditional Coefficients of a Variable in an Interaction Term
2 | #'
3 | #' \code{interplot} is a generic function to produce a plot of the coefficient estimates of one variable in a two-way interaction conditional on the values of the other variable in the interaction term. The function invokes particular \code{methods} which depend on the \code{\link{class}} of the first argument.
4 | #'
5 | #' @param m A model object including an interaction term, or, alternately, a data frame generated by an earlier call to interplot using the argument plot = FALSE.
6 | #' @param var1 The name (as a string) of the variable of interest in the interaction term; its conditional coefficient estimates will be plotted.
7 | #' @param var2 The name (as a string) of the other variable in the interaction term
8 | #' @param plot A logical value indicating whether the output is a plot or a dataframe including the conditional coefficient estimates of var1, their upper and lower bounds, and the corresponding values of var2.
9 | #' @param steps Desired length of the sequence. A non-negative number, which for seq and seq.int will be rounded up if fractional. The default is 100 or the unique categories in the \code{var2} (when it is less than 100. Also see \code{\link{unique}}).
10 | #' @param ci A numeric value defining the confidence intervals. The default value is 95\% (0.95).
11 | #' @param adjCI A logical value indication if applying the adjustment of confidence intervals to control the false discovery rate following the Esarey and Sumner (2017) procedure. (See also Benjamini and Hochberg 1995.) The default is FALSE; the plot presents the confidence intervals suggested by Brambor, Clark, and Golder (2006). The functions dealing with multilevel model outputs in this package do not equip with this argument, because there is the controversy on the accurate degrees of freedom for multilevel models, esp. when random effect is engaged and the degrees of freedom is a necessary information to conduct the CI adjustment. See e.g., https://stat.ethz.ch/pipermail/r-help/2006-May/094765.html and https://stat.ethz.ch/pipermail/r-sig-mixed-models/2008q1/000517.html.
12 | #' @param hist A logical value indicating if there is a histogram of `var2` added at the bottom of the conditional effect plot.
13 | #' @param var2_dt A numerical value indicating the frequency distribution of `var2`. It is only used when `hist == TRUE`. When the object is a model, the default is the distribution of `var2` of the model.
14 | #' @param predPro A logical value with default of `FALSE`. When the `m` is an object of class `glm` or `glmerMod` and the argument is set to `TRUE`, the function will plot predicted probabilities at the values given by `var2_vals`.
15 | #' @param var2_vals A numerical value indicating the values the predicted probabilities are estimated, when `predPro` is `TRUE`.
16 | #' @param point A logical value determining the format of plot. By default, the function produces a line plot when var2 takes on ten or more distinct values and a point (dot-and-whisker) plot otherwise; option TRUE forces a point plot.
17 | #' @param sims Number of independent simulation draws used to calculate upper and lower bounds of coefficient estimates: lower values run faster; higher values produce smoother curves.
18 | #' @param xmin A numerical value indicating the minimum value shown of x shown in the graph. Rarely used.
19 | #' @param xmax A numerical value indicating the maximum value shown of x shown in the graph. Rarely used.
20 | #' @param ercolor A character value indicating the outline color of the whisker or ribbon.
21 | #' @param esize A numerical value indicating the size of the whisker or ribbon.
22 | #' @param ralpha A numerical value indicating the transparency of the ribbon.
23 | #' @param rfill A character value indicating the filling color of the ribbon.
24 | #' @param stats_cp A character value indicating what statistics to present as the plot note. Three options are available: "none", "ci", and "ks". The default is "none". See the Details for more information.
25 | #' @param txt_caption A character string to add a note for the plot, a value will sending to \code{ggplot2::labs(caption = txt_caption))}.
26 | #' @param ... Other ggplot aesthetics arguments for points in the dot-whisker plot or lines in the line-ribbon plots. Not currently used.
27 | #'
28 | #' @details \code{interplot} visualizes the changes in the coefficient of one term in a two-way interaction conditioned by the other term. In the current version, the function works with interactions in the following classes of models:
29 | #' \itemize{
30 | #' \item Ordinary linear models (object class: \code{lm});
31 | #' \item Generalized linear models (object class: \code{glm});
32 | #' \item Linear mixed-effects models (object class: \code{lmerMod});
33 | #' \item Generalized linear mixed-effects models (object class: \code{glmerMod});
34 | #' \item Ordinary linear models with imputed data (object class: \code{list});
35 | #' \item Generalized linear models with imputed data (object class: \code{list})
36 | #' \item Linear mixed-effects models with imputed data (object class: \code{list});
37 | #' \item Generalized linear mixed-effects models with imputed data (object class: \code{list}).
38 | #' }
39 | #'
40 | #' The examples below illustrate how methods invoked by this generic deal with different type of objects.
41 | #'
42 | #' Because the output function is based on \code{\link[ggplot2]{ggplot}}, any additional arguments and layers supported by \code{ggplot2} can be added with the \code{+}.
43 | #'
44 | #' \code{interplot} visualizes the conditional effect based on simulated marginal effects. The simulation provides a probabilistic distribution of moderation effect of the conditioning variable (\code{var2}) at every preset values (including the minimum and maximum values) of the conditioned variable (\code{var1}), denoted as Emin and Emax. This output allows the function to further examine the conditional effect statistically in two ways. One is to examine if the distribution of \eqn{Emax - Emin} covers zero. The other is to directly compare Emin and Emax through statistical tools for distributional comparisons. Users can choose either method by setting the argument \code{stats_cp} to "ci" or "ks".
45 | #' \itemize{
46 | #' \item "ci" provides the confidence interval of the difference of \eqn{Emax - Emin}. An interval including 0 suggests no statistical difference before and after the conditional effect is applied, and vise versa.
47 | #' \item "ks" presents the result of a two-sample Kolmogorov-Smirnov test of the simulated distributions of Emin and Emax. The output includes a D statistics and a p-value of the null hypothesis that the two distributions come from the same distribution at the 0.05 level.
48 | #' }
49 | #'
50 | #' See an illustration in the package vignette.
51 | #'
52 | #' @return The function returns a \code{ggplot} object.
53 | #'
54 | #'
55 | #' @examples
56 | #' data(mtcars)
57 | #' m_cyl <- lm(mpg ~ wt * cyl, data = mtcars)
58 | #' library(interplot)
59 | #'
60 | #' # Plot interactions with a continous conditioning variable
61 | #' interplot(m = m_cyl, var1 = 'cyl', var2 = 'wt') +
62 | #' xlab('Automobile Weight (thousands lbs)') +
63 | #' ylab('Estimated Coefficient for Number of Cylinders') +
64 | #' ggtitle('Estimated Coefficient of Engine Cylinders\non Mileage by Automobile Weight') +
65 | #' theme(plot.title = element_text(face='bold'))
66 | #'
67 | #'
68 | #' # Plot interactions with a categorical conditioning variable
69 | #' interplot(m = m_cyl, var1 = 'wt', var2 = 'cyl') +
70 | #' xlab('Number of Cylinders') +
71 | #' ylab('Estimated Coefficient for Automobile Weight (thousands lbs)') +
72 | #' ggtitle('Estimated Coefficient of Automobile Weight \non Mileage by Engine Cylinders') +
73 | #' theme(plot.title = element_text(face='bold'))
74 | #'
75 | #' @source Benjamini, Yoav, and Yosef Hochberg. 1995. "Controlling the False
76 | #' Discovery Rate: A Practical and Powerful Approach to Multiple Testing".
77 | #' Journal of the Royal Statistical Society, Series B 57(1): 289--300.
78 | #'
79 | #' Brambor, Thomas, William Roberts Clark, and Matt Golder.
80 | #' "Understanding interaction models: Improving empirical analyses". Political
81 | #' Analysis 14.1 (2006): 63-82.
82 | #'
83 | #' Esarey, Justin, and Jane Lawrence Sumner. 2015. "Marginal Effects in
84 | #' Interaction Models: Determining and Controlling the False Positive Rate".
85 | #' URL: \url{https://jee3.web.rice.edu/interaction-overconfidence.pdf}.
86 | #'
87 | #' @export
88 |
89 |
90 | interplot <- function(m, var1, var2, plot = TRUE, steps = NULL, ci = .95, adjCI = FALSE, hist = FALSE, var2_dt = NA, predPro = FALSE, var2_vals = NULL, point = FALSE, sims = 1000, xmin = NA, xmax = NA, ercolor = NA, esize = 0.5, ralpha = 0.5, rfill = "grey70", stats_cp = "none", txt_caption = NULL, ...) {
91 |
92 |
93 | if (class(m)[1] == "list") {
94 | if (class(m[[1]])[1] == "lmerMod") {
95 | class(m) <- "mlmmi"
96 | }
97 | if (class(m[[1]])[1] == "glmerMod") {
98 | class(m) <- "gmlmmi"
99 | }
100 | if (class(m[[1]])[1] == "lm") {
101 | class(m) <- "lmmi"
102 | }
103 | if (class(m[[1]])[1] == "glm") {
104 | class(m) <- "glmmi"
105 | }
106 | }
107 |
108 | if (class(m)[1] == "data.frame")
109 | class(m) <- "plot"
110 |
111 |
112 | UseMethod("interplot", m)
113 | }
114 |
--------------------------------------------------------------------------------
/docs/docsearch.css:
--------------------------------------------------------------------------------
1 | /* Docsearch -------------------------------------------------------------- */
2 | /*
3 | Source: https://github.com/algolia/docsearch/
4 | License: MIT
5 | */
6 |
7 | .algolia-autocomplete {
8 | display: block;
9 | -webkit-box-flex: 1;
10 | -ms-flex: 1;
11 | flex: 1
12 | }
13 |
14 | .algolia-autocomplete .ds-dropdown-menu {
15 | width: 100%;
16 | min-width: none;
17 | max-width: none;
18 | padding: .75rem 0;
19 | background-color: #fff;
20 | background-clip: padding-box;
21 | border: 1px solid rgba(0, 0, 0, .1);
22 | box-shadow: 0 .5rem 1rem rgba(0, 0, 0, .175);
23 | }
24 |
25 | @media (min-width:768px) {
26 | .algolia-autocomplete .ds-dropdown-menu {
27 | width: 175%
28 | }
29 | }
30 |
31 | .algolia-autocomplete .ds-dropdown-menu::before {
32 | display: none
33 | }
34 |
35 | .algolia-autocomplete .ds-dropdown-menu [class^=ds-dataset-] {
36 | padding: 0;
37 | background-color: rgb(255,255,255);
38 | border: 0;
39 | max-height: 80vh;
40 | }
41 |
42 | .algolia-autocomplete .ds-dropdown-menu .ds-suggestions {
43 | margin-top: 0
44 | }
45 |
46 | .algolia-autocomplete .algolia-docsearch-suggestion {
47 | padding: 0;
48 | overflow: visible
49 | }
50 |
51 | .algolia-autocomplete .algolia-docsearch-suggestion--category-header {
52 | padding: .125rem 1rem;
53 | margin-top: 0;
54 | font-size: 1.3em;
55 | font-weight: 500;
56 | color: #00008B;
57 | border-bottom: 0
58 | }
59 |
60 | .algolia-autocomplete .algolia-docsearch-suggestion--wrapper {
61 | float: none;
62 | padding-top: 0
63 | }
64 |
65 | .algolia-autocomplete .algolia-docsearch-suggestion--subcategory-column {
66 | float: none;
67 | width: auto;
68 | padding: 0;
69 | text-align: left
70 | }
71 |
72 | .algolia-autocomplete .algolia-docsearch-suggestion--content {
73 | float: none;
74 | width: auto;
75 | padding: 0
76 | }
77 |
78 | .algolia-autocomplete .algolia-docsearch-suggestion--content::before {
79 | display: none
80 | }
81 |
82 | .algolia-autocomplete .ds-suggestion:not(:first-child) .algolia-docsearch-suggestion--category-header {
83 | padding-top: .75rem;
84 | margin-top: .75rem;
85 | border-top: 1px solid rgba(0, 0, 0, .1)
86 | }
87 |
88 | .algolia-autocomplete .ds-suggestion .algolia-docsearch-suggestion--subcategory-column {
89 | display: block;
90 | padding: .1rem 1rem;
91 | margin-bottom: 0.1;
92 | font-size: 1.0em;
93 | font-weight: 400
94 | /* display: none */
95 | }
96 |
97 | .algolia-autocomplete .algolia-docsearch-suggestion--title {
98 | display: block;
99 | padding: .25rem 1rem;
100 | margin-bottom: 0;
101 | font-size: 0.9em;
102 | font-weight: 400
103 | }
104 |
105 | .algolia-autocomplete .algolia-docsearch-suggestion--text {
106 | padding: 0 1rem .5rem;
107 | margin-top: -.25rem;
108 | font-size: 0.8em;
109 | font-weight: 400;
110 | line-height: 1.25
111 | }
112 |
113 | .algolia-autocomplete .algolia-docsearch-footer {
114 | width: 110px;
115 | height: 20px;
116 | z-index: 3;
117 | margin-top: 10.66667px;
118 | float: right;
119 | font-size: 0;
120 | line-height: 0;
121 | }
122 |
123 | .algolia-autocomplete .algolia-docsearch-footer--logo {
124 | background-image: url("data:image/svg+xml;utf8,");
125 | background-repeat: no-repeat;
126 | background-position: 50%;
127 | background-size: 100%;
128 | overflow: hidden;
129 | text-indent: -9000px;
130 | width: 100%;
131 | height: 100%;
132 | display: block;
133 | transform: translate(-8px);
134 | }
135 |
136 | .algolia-autocomplete .algolia-docsearch-suggestion--highlight {
137 | color: #FF8C00;
138 | background: rgba(232, 189, 54, 0.1)
139 | }
140 |
141 |
142 | .algolia-autocomplete .algolia-docsearch-suggestion--text .algolia-docsearch-suggestion--highlight {
143 | box-shadow: inset 0 -2px 0 0 rgba(105, 105, 105, .5)
144 | }
145 |
146 | .algolia-autocomplete .ds-suggestion.ds-cursor .algolia-docsearch-suggestion--content {
147 | background-color: rgba(192, 192, 192, .15)
148 | }
149 |
--------------------------------------------------------------------------------
/docs/reference/interplot.lmerMod.html:
--------------------------------------------------------------------------------
1 |
2 | Plot Conditional Coefficients in Mixed-Effects Models with Interaction Terms — interplot.lmerMod • interplot
6 |
7 |
8 |
9 |
50 |
51 |
52 |
53 |
54 |
55 |
56 | Plot Conditional Coefficients in Mixed-Effects Models with Interaction Terms
57 | Source: R/Interplot_mlm.R
58 | interplot.lmerMod.Rd
59 |
60 |
61 |
62 | interplot.mlm is a method to calculate conditional coefficient estimates from the results of multilevel (mixed-effects) regression models with interaction terms.
63 |
64 |
65 |
66 | # S3 method for lmerMod
67 | interplot(
68 | m,
69 | var1,
70 | var2,
71 | plot = TRUE,
72 | steps = NULL,
73 | ci = 0.95,
74 | adjCI = FALSE,
75 | hist = FALSE,
76 | var2_dt = NA,
77 | predPro = FALSE,
78 | var2_vals = NULL,
79 | point = FALSE,
80 | sims = 5000,
81 | xmin = NA,
82 | xmax = NA,
83 | ercolor = NA,
84 | esize = 0.5,
85 | ralpha = 0.5,
86 | rfill = "grey70",
87 | stats_cp = "none",
88 | txt_caption = NULL,
89 | facet_labs = NULL,
90 | ...
91 | )
92 |
93 |
94 |
95 | Arguments
96 | - m
97 | A model object including an interaction term, or, alternately, a data frame recording conditional coefficients.
- var1
101 | The name (as a string) of the variable of interest in the interaction term; its conditional coefficient estimates will be plotted.
- var2
105 | The name (as a string) of the other variable in the interaction term.
- plot
109 | A logical value indicating whether the output is a plot or a dataframe including the conditional coefficient estimates of var1, their upper and lower bounds, and the corresponding values of var2.
- steps
113 | Desired length of the sequence. A non-negative number, which for seq and seq.int will be rounded up if fractional. The default is 100 or the unique categories in the var2 (when it is less than 100. Also see unique).
- ci
117 | A numeric value defining the confidence intervals. The default value is 95% (0.95).
- adjCI
121 | Not working for `lmer` outputs yet.
- hist
125 | A logical value indicating if there is a histogram of `var2` added at the bottom of the conditional effect plot.
- var2_dt
129 | A numerical value indicating the frequency distribution of `var2`. It is only used when `hist == TRUE`. When the object is a model, the default is the distribution of `var2` of the model.
- predPro
133 | A logical value with default of `FALSE`. When the `m` is an object of class `glmerMod` and the argument is set to `TRUE`, the function will plot predicted probabilities at the values given by `var2_vals`.
- var2_vals
137 | A numerical value indicating the values the predicted probabilities are estimated, when `predPro` is `TRUE`.
- point
141 | A logical value determining the format of plot. By default, the function produces a line plot when var2 takes on ten or more distinct values and a point (dot-and-whisker) plot otherwise; option TRUE forces a point plot.
- sims
145 | Number of independent simulation draws used to calculate upper and lower bounds of coefficient estimates: lower values run faster; higher values produce smoother curves.
- xmin
149 | A numerical value indicating the minimum value shown of x shown in the graph. Rarely used.
- xmax
153 | A numerical value indicating the maximum value shown of x shown in the graph. Rarely used.
- ercolor
157 | A character value indicating the outline color of the whisker or ribbon.
- esize
161 | A numerical value indicating the size of the whisker or ribbon.
- ralpha
165 | A numerical value indicating the transparency of the ribbon.
- rfill
169 | A character value indicating the filling color of the ribbon.
- stats_cp
173 | A character value indicating what statistics to present as the plot note. Three options are available: "none", "ci", and "ks". The default is "none". See the Details for more information.
- txt_caption
177 | A character string to add a note for the plot, a value will sending to ggplot2::labs(caption = txt_caption)).
- facet_labs
181 | An optional character vector of facet labels to be used when plotting an interaction with a factor variable.
- ...
185 | Other ggplot aesthetics arguments for points in the dot-whisker plot or lines in the line-ribbon plots. Not currently used.
188 |
189 | Value
190 |
191 |
192 | The function returns a ggplot object.
193 |
194 |
195 | Details
196 | interplot.mlm is a S3 method from the interplot. It works on mixed-effects objects with class lmerMod and glmerMod.
197 | Because the output function is based on ggplot, any additional arguments and layers supported by ggplot2 can be added with the +.
198 | interplot visualizes the conditional effect based on simulated marginal effects. The simulation provides a probabilistic distribution of moderation effect of the conditioning variable (var2) at every preset values (including the minimum and maximum values) of the conditioned variable (var1), denoted as Emin and Emax. This output allows the function to further examine the conditional effect statistically in two ways. One is to examine if the distribution of \(Emax - Emin\) covers zero. The other is to directly compare Emin and Emax through statistical tools for distributional comparisons. Users can choose either method by setting the argument stats_cp to "ci" or "ks".
"ci" provides the confidence interval of the difference of \(Emax - Emin\). An interval including 0 suggests no statistical difference before and after the conditional effect is applied, and vise versa.
"ks" presents the result of a two-sample Kolmogorov-Smirnov test of the simulated distributions of Emin and Emax. The output includes a D statistics and a p-value of the null hypothesis that the two distributions come from the same distribution at the 0.05 level.
See an illustration in the package vignette.
201 |
202 |
203 |
204 |
207 |
208 |
209 |
210 |
219 |
220 |
221 |
222 |
223 |
224 |
225 |
226 |
227 |
--------------------------------------------------------------------------------
78 |
93 |
94 |
95 |
96 |
107 |
79 |
86 |
87 |
91 |
92 |
80 |
82 |
83 | Content not found. Please use links in the navbar.
84 |
85 | Page not found (404)
81 |
54 |
76 |
77 |
78 |
79 |
55 |
70 |
71 |
74 |
75 |
56 |
58 |
59 | Contributor Code of Conduct
57 |
60 |
61 |
68 |
69 | As contributors and maintainers of this project, we pledge to respect all people who contribute through reporting issues, posting feature requests, updating documentation, submitting pull requests or patches, and other activities.
62 |We are committed to making participation in this project a harassment-free experience for everyone, regardless of level of experience, gender, gender identity and expression, sexual orientation, disability, personal appearance, body size, race, ethnicity, age, or religion.
63 |Examples of unacceptable behavior by participants include the use of sexual language or imagery, derogatory comments or personal attacks, trolling, public or private harassment, insults, or other unprofessional conduct.
64 |Project maintainers have the right and responsibility to remove, edit, or reject comments, commits, code, wiki edits, issues, and other contributions that are not aligned to this Code of Conduct. Project maintainers who do not follow the Code of Conduct may be removed from the project team.
65 |Instances of abusive, harassing, or otherwise unacceptable behavior may be reported by opening an issue or contacting one or more of the project maintainers.
66 |This Code of Conduct is adapted from the Contributor Covenant (http:contributor-covenant.org), version 1.0.0, available at http://contributor-covenant.org/version/1/0/0/
67 |
54 |
93 |
94 |
95 |
55 |
88 |
89 |
92 |
56 |
58 |
59 | Reference
57 |
60 | All functions61 | 62 | |
63 | |
|---|---|
| 64 | 65 | | 66 |Plot Conditional Coefficients in (Generalized) Linear Models with Interaction Terms |
67 |
| 68 | 69 | | 70 |Plot Conditional Coefficients in Mixed-Effects Models with Interaction Terms |
71 |
| 72 | 73 | | 74 |Plot Conditional Coefficients in (Generalized) Linear Models with Imputed Data and Interaction Terms |
75 |
| 76 | 77 | | 78 |Plot Conditional Coefficients in Mixed-Effects Models with Imputed Data and Interaction Terms |
79 |
| 80 | 81 | | 82 |Plot Conditional Coefficients in Models with Interaction Terms |
83 |
| 84 | 85 | | 86 |Plot Conditional Coefficients of a Variable in an Interaction Term |
87 |
80 |
141 |
142 |
143 |
154 |
81 |
82 |
85 |
100 |
101 |
140 | pkgdown
85 |
86 |
88 |
98 |
99 | interplot is a tool for plotting the conditional coefficients (“marginal effects”) of variables included in multiplicative interaction terms. The function plots the changes in the coefficient of one variable in a two-way interaction term conditional on the value of the other included variable. The plot also includes simulated 95% confidential intervals of these coefficients.
To install:
90 |-
91 |
- the latest released version:
install.packages("interplot").
92 | - the latest developing version:
remotes::install_github("sammo3182/interplot").
93 |
More details are available at:
95 |http://cran.r-project.org/web/packages/interplot/vignettes/interplot-vignette.html
96 |Please note that this project is released with a Contributor Code of Conduct. By participating in this project you agree to abide by its terms.
97 |
54 |
208 |
209 |
210 |
55 |
204 |
207 |
56 |
59 |
60 |
61 | Plot Conditional Coefficients in Mixed-Effects Models with Interaction Terms
57 | Source:R/Interplot_mlm.R
58 | interplot.lmerMod.Rd
62 |
64 |
65 | interplot.mlm is a method to calculate conditional coefficient estimates from the results of multilevel (mixed-effects) regression models with interaction terms.
66 |
92 |
93 |
94 | # S3 method for lmerMod
67 | interplot(
68 | m,
69 | var1,
70 | var2,
71 | plot = TRUE,
72 | steps = NULL,
73 | ci = 0.95,
74 | adjCI = FALSE,
75 | hist = FALSE,
76 | var2_dt = NA,
77 | predPro = FALSE,
78 | var2_vals = NULL,
79 | point = FALSE,
80 | sims = 5000,
81 | xmin = NA,
82 | xmax = NA,
83 | ercolor = NA,
84 | esize = 0.5,
85 | ralpha = 0.5,
86 | rfill = "grey70",
87 | stats_cp = "none",
88 | txt_caption = NULL,
89 | facet_labs = NULL,
90 | ...
91 | )
95 |
188 | Arguments
96 |- m 97 |
A model object including an interaction term, or, alternately, a data frame recording conditional coefficients.
98 |
99 |
100 | - var1 101 |
The name (as a string) of the variable of interest in the interaction term; its conditional coefficient estimates will be plotted.
102 |
103 |
104 | - var2 105 |
The name (as a string) of the other variable in the interaction term.
106 |
107 |
108 | - plot 109 |
A logical value indicating whether the output is a plot or a dataframe including the conditional coefficient estimates of var1, their upper and lower bounds, and the corresponding values of var2.
110 |
111 |
112 | - steps 113 |
Desired length of the sequence. A non-negative number, which for seq and seq.int will be rounded up if fractional. The default is 100 or the unique categories in the
var2(when it is less than 100. Also seeunique).
114 |
115 |
116 | - ci 117 |
A numeric value defining the confidence intervals. The default value is 95% (0.95).
118 |
119 |
120 | - adjCI 121 |
Not working for `lmer` outputs yet.
122 |
123 |
124 | - hist 125 |
A logical value indicating if there is a histogram of `var2` added at the bottom of the conditional effect plot.
126 |
127 |
128 | - var2_dt 129 |
A numerical value indicating the frequency distribution of `var2`. It is only used when `hist == TRUE`. When the object is a model, the default is the distribution of `var2` of the model.
130 |
131 |
132 | - predPro 133 |
A logical value with default of `FALSE`. When the `m` is an object of class `glmerMod` and the argument is set to `TRUE`, the function will plot predicted probabilities at the values given by `var2_vals`.
134 |
135 |
136 | - var2_vals 137 |
A numerical value indicating the values the predicted probabilities are estimated, when `predPro` is `TRUE`.
138 |
139 |
140 | - point 141 |
A logical value determining the format of plot. By default, the function produces a line plot when var2 takes on ten or more distinct values and a point (dot-and-whisker) plot otherwise; option TRUE forces a point plot.
142 |
143 |
144 | - sims 145 |
Number of independent simulation draws used to calculate upper and lower bounds of coefficient estimates: lower values run faster; higher values produce smoother curves.
146 |
147 |
148 | - xmin 149 |
A numerical value indicating the minimum value shown of x shown in the graph. Rarely used.
150 |
151 |
152 | - xmax 153 |
A numerical value indicating the maximum value shown of x shown in the graph. Rarely used.
154 |
155 |
156 | - ercolor 157 |
A character value indicating the outline color of the whisker or ribbon.
158 |
159 |
160 | - esize 161 |
A numerical value indicating the size of the whisker or ribbon.
162 |
163 |
164 | - ralpha 165 |
A numerical value indicating the transparency of the ribbon.
166 |
167 |
168 | - rfill 169 |
A character value indicating the filling color of the ribbon.
170 |
171 |
172 | - stats_cp 173 |
A character value indicating what statistics to present as the plot note. Three options are available: "none", "ci", and "ks". The default is "none". See the Details for more information.
174 |
175 |
176 | - txt_caption 177 |
A character string to add a note for the plot, a value will sending to
ggplot2::labs(caption = txt_caption)).
178 |
179 |
180 | - facet_labs 181 |
An optional character vector of facet labels to be used when plotting an interaction with a factor variable.
182 |
183 |
184 | - ... 185 |
Other ggplot aesthetics arguments for points in the dot-whisker plot or lines in the line-ribbon plots. Not currently used.
186 |
187 |
189 |
194 | Value
190 | 191 | 192 |The function returns a ggplot object.
195 |
202 |
203 | Details
196 |interplot.mlm is a S3 method from the interplot. It works on mixed-effects objects with class lmerMod and glmerMod.
Because the output function is based on ggplot, any additional arguments and layers supported by ggplot2 can be added with the +.
interplot visualizes the conditional effect based on simulated marginal effects. The simulation provides a probabilistic distribution of moderation effect of the conditioning variable (var2) at every preset values (including the minimum and maximum values) of the conditioned variable (var1), denoted as Emin and Emax. This output allows the function to further examine the conditional effect statistically in two ways. One is to examine if the distribution of \(Emax - Emin\) covers zero. The other is to directly compare Emin and Emax through statistical tools for distributional comparisons. Users can choose either method by setting the argument stats_cp to "ci" or "ks".
"ci" provides the confidence interval of the difference of \(Emax - Emin\). An interval including 0 suggests no statistical difference before and after the conditional effect is applied, and vise versa.
199 | "ks" presents the result of a two-sample Kolmogorov-Smirnov test of the simulated distributions of Emin and Emax. The output includes a D statistics and a p-value of the null hypothesis that the two distributions come from the same distribution at the 0.05 level.
200 |
See an illustration in the package vignette.
201 |