more updates

JohnnyDoorn · Oct 30, 2024 · 7d1a4a6 · 7d1a4a6
1 parent 2a4b740
commit 7d1a4a6
Show file tree

Hide file tree

Showing 27 changed files with 429 additions and 411 deletions.
diff --git a/SSR/.quarto/xref/INDEX b/SSR/.quarto/xref/INDEX
@@ -99,5 +99,8 @@
   "2024/21. multiple_regression/21. multiple_regresson.qmd": {
     "21. multiple_regresson.html": "58ddcc28",
     "21. multiple_regresson.slide.html": "0cf5212e"
+  },
+  "topics/block3_opening/block3_opening.qmd": {
+    "block3_opening.html": "74fdf750"
   }
 }
diff --git a/SSR/2024/20. correlation_and_simple_regression/20. correlation_simple_regression.html b/SSR/2024/20. correlation_and_simple_regression/20. correlation_simple_regression.html
diff --git a/SSR/2024/20. correlation_and_simple_regression/20. correlation_simple_regression.qmd b/SSR/2024/20. correlation_and_simple_regression/20. correlation_simple_regression.qmd
@@ -9,13 +9,20 @@ format:
     output-ext: slide.html
 ---
 
+## Fishing Expedition!
+<img src="qr_kkd7n.png">
+
+Thank you Julia, Kim, Barnaba, Lioba, Joaquin, Tadeja, and Ewa!
+
+<!-- Nikolas and Zeynep for error-spotting -->
+
 ```{r child="../../topics/block3_opening/block3_opening.qmd"}
 ```
 
-
 ```{r child="../../topics/correlation/correlation_albumsales.qmd"}
 ```
 
+
 ```{r child="../../topics/regression_one_predictor/regression_one_predictor.qmd", eval=TRUE}
 ```
 

diff --git a/SSR/2024/20. correlation_and_simple_regression/20. correlation_simple_regression.slide.html b/SSR/2024/20. correlation_and_simple_regression/20. correlation_simple_regression.slide.html
diff --git a/...sion/20.-correlation_simple_regression_files/figure-html/unnamed-chunk-55-1.png b/...sion/20.-correlation_simple_regression_files/figure-html/unnamed-chunk-55-1.png
diff --git a/...sion/20.-correlation_simple_regression_files/figure-html/unnamed-chunk-57-1.png b/...sion/20.-correlation_simple_regression_files/figure-html/unnamed-chunk-57-1.png
diff --git a/...sion/20.-correlation_simple_regression_files/figure-html/unnamed-chunk-62-1.png b/...sion/20.-correlation_simple_regression_files/figure-html/unnamed-chunk-62-1.png
diff --git a/...sion/20.-correlation_simple_regression_files/figure-html/unnamed-chunk-63-1.png b/...sion/20.-correlation_simple_regression_files/figure-html/unnamed-chunk-63-1.png
diff --git a/...sion/20.-correlation_simple_regression_files/figure-html/unnamed-chunk-76-1.png b/...sion/20.-correlation_simple_regression_files/figure-html/unnamed-chunk-76-1.png
diff --git a/...sion/20.-correlation_simple_regression_files/figure-html/unnamed-chunk-77-1.png b/...sion/20.-correlation_simple_regression_files/figure-html/unnamed-chunk-77-1.png
diff --git a/.../20.-correlation_simple_regression_files/figure-revealjs/unnamed-chunk-55-1.png b/.../20.-correlation_simple_regression_files/figure-revealjs/unnamed-chunk-55-1.png
diff --git a/.../20.-correlation_simple_regression_files/figure-revealjs/unnamed-chunk-57-1.png b/.../20.-correlation_simple_regression_files/figure-revealjs/unnamed-chunk-57-1.png
diff --git a/.../20.-correlation_simple_regression_files/figure-revealjs/unnamed-chunk-62-1.png b/.../20.-correlation_simple_regression_files/figure-revealjs/unnamed-chunk-62-1.png
diff --git a/.../20.-correlation_simple_regression_files/figure-revealjs/unnamed-chunk-63-1.png b/.../20.-correlation_simple_regression_files/figure-revealjs/unnamed-chunk-63-1.png
diff --git a/.../20.-correlation_simple_regression_files/figure-revealjs/unnamed-chunk-76-1.png b/.../20.-correlation_simple_regression_files/figure-revealjs/unnamed-chunk-76-1.png
diff --git a/.../20.-correlation_simple_regression_files/figure-revealjs/unnamed-chunk-77-1.png b/.../20.-correlation_simple_regression_files/figure-revealjs/unnamed-chunk-77-1.png
diff --git a/SSR/2024/20. correlation_and_simple_regression/qr_kkd7n.png b/SSR/2024/20. correlation_and_simple_regression/qr_kkd7n.png
diff --git a/SSR/topics/block3_opening/block3_opening.html b/SSR/topics/block3_opening/block3_opening.html
@@ -6,6 +6,7 @@
 
 <meta name="viewport" content="width=device-width, initial-scale=1.0, user-scalable=yes">
 
+<meta name="author" content="Sharon Klinkenberg &amp; Johnny van Doorn">
 
 <title>block3_opening</title>
 <style>
@@ -39,26 +40,43 @@
 
 </head>
 
-<body class="fullcontent">
-
-<div id="quarto-content" class="page-columns page-rows-contents page-layout-article">
+<body>
 
+<div id="quarto-content" class="page-columns page-rows-contents page-layout-article toc-left">
+<div id="quarto-sidebar-toc-left" class="sidebar toc-left">
+  <nav id="TOC" role="doc-toc" class="toc-active">
+    <h2 id="toc-title">Table of contents</h2>
+
+  <ul class="collapse">
+  <li><a href="#some-opening-mantras" id="toc-some-opening-mantras" class="nav-link active" data-scroll-target="#some-opening-mantras">Some Opening Mantra’s</a>
+  <ul class="collapse">
+  <li><a href="#always-plot-your-data" id="toc-always-plot-your-data" class="nav-link" data-scroll-target="#always-plot-your-data">Always plot your data</a></li>
+  <li><a href="#always-plot-your-data-1" id="toc-always-plot-your-data-1" class="nav-link" data-scroll-target="#always-plot-your-data-1">Always plot your data</a></li>
+  <li><a href="#predicting-the-dv" id="toc-predicting-the-dv" class="nav-link" data-scroll-target="#predicting-the-dv">Predicting the DV</a></li>
+  <li><a href="#keep-your-models-simpleinterpretable" id="toc-keep-your-models-simpleinterpretable" class="nav-link" data-scroll-target="#keep-your-models-simpleinterpretable">Keep your models simple/interpretable</a></li>
+  <li><a href="#statistics-in-block-3" id="toc-statistics-in-block-3" class="nav-link" data-scroll-target="#statistics-in-block-3">Statistics in block 3</a></li>
+  </ul></li>
+  </ul>
+<div class="quarto-alternate-formats"><h2>Other Formats</h2><ul><li><a href="block3_opening.html"><i class="bi bi-file-slides"></i>RevealJS</a></li></ul></div></nav>
+</div>
+<div id="quarto-margin-sidebar" class="sidebar margin-sidebar">
+</div>
 <main class="content" id="quarto-document-content">
 
 
 
-<section id="what-have-we-learnt-so-far" class="level1">
-<h1>What have we learnt so far?</h1>
+<section id="some-opening-mantras" class="level1">
+<h1>Some Opening Mantra’s</h1>
 <section id="always-plot-your-data" class="level2">
 <h2 class="anchored" data-anchor-id="always-plot-your-data">Always plot your data</h2>
-<p><img src="trex.png" class="img-fluid" width="800"></p>
+<p><img src="raincloud.png" class="img-fluid" width="800"></p>
 </section>
 <section id="always-plot-your-data-1" class="level2">
 <h2 class="anchored" data-anchor-id="always-plot-your-data-1">Always plot your data</h2>
-<p><img src="raincloud.png" class="img-fluid" width="800"></p>
+<p><img src="trex.png" class="img-fluid" width="800"></p>
 </section>
-<section id="explaining-variance" class="level2">
-<h2 class="anchored" data-anchor-id="explaining-variance">Explaining variance</h2>
+<section id="predicting-the-dv" class="level2 smaller">
+<h2 class="smaller anchored" data-anchor-id="predicting-the-dv">Predicting the DV</h2>
 <div class="incremental">
 <ul class="incremental">
 <li>The dependent variable (DV) varies</li>
@@ -71,27 +89,25 @@ <h2 class="anchored" data-anchor-id="explaining-variance">Explaining variance</h
 <li>Accurate approximation/prediction of the DV? -&gt; high explained variance</li>
 <li>Total variability of DV = explained variance + unexplained variance
 <ul class="incremental">
-<li><span class="math inline">\(\text{DV} = \text{model} + \text{error}\)</span></li>
+<li><span class="math inline">\(\text{DV} = \text{model prediction} + \text{error}\)</span></li>
 </ul></li>
 <li>More explained variance than unexplained variance (e.g., <span class="math inline">\(F &gt;&gt; 1\)</span>) -&gt; success!</li>
 </ul>
 </div>
 </section>
-<section id="predictions" class="level2">
-<h2 class="anchored" data-anchor-id="predictions">Predictions</h2>
-<div class="incremental">
-<ul class="incremental">
-<li>The dependent variable (DV) varies</li>
-<li>We try to recreate/predict the observed values by looking at predictor variables</li>
-<li>Total variability of DV = explained variance + unexplained variance</li>
-<li>More explained variance than unexplained variance (<span class="math inline">\(F &gt;&gt; 1\)</span>) -&gt; success!</li>
-</ul>
-</div>
-</section>
 <section id="keep-your-models-simpleinterpretable" class="level2">
 <h2 class="anchored" data-anchor-id="keep-your-models-simpleinterpretable">Keep your models simple/interpretable</h2>
 <p><img src="manyvars.png" class="img-fluid" width="800"></p>
 </section>
+<section id="statistics-in-block-3" class="level2">
+<h2 class="anchored" data-anchor-id="statistics-in-block-3">Statistics in block 3</h2>
+<ul>
+<li>Simple regression (single predictor)</li>
+<li>Multiple regression (two predictors)</li>
+<li>Mediation analysis (i.e., modeling dependence)</li>
+<li>Moderation analysis (i.e., modeling interactions)</li>
+</ul>
+</section>
 </section>
 
 </main>

diff --git a/SSR/topics/block3_opening/block3_opening.qmd b/SSR/topics/block3_opening/block3_opening.qmd
@@ -24,3 +24,9 @@
 ![](manyvars.png){width="800px"}
 
 
+## Statistics in block 3 
+
+- Simple regression (single predictor)
+- Multiple regression (two predictors)
+- Mediation analysis (i.e., modeling dependence)
+- Moderation analysis (i.e., modeling interactions)
diff --git a/SSR/topics/correlation/correlation_albumsales.html b/SSR/topics/correlation/correlation_albumsales.html
@@ -385,7 +385,6 @@ <h2 class="subsection anchored" data-anchor-id="significance-of-partial-correlat
 <p><img src="correlation_albumsales_files/figure-html/unnamed-chunk-22-1.png" class="img-fluid" width="672"></p>
 </div>
 </div>
-<p>```</p>
 </section>
 </section>
 

diff --git a/SSR/topics/correlation/correlation_albumsales.qmd b/SSR/topics/correlation/correlation_albumsales.qmd
@@ -387,4 +387,3 @@ arrows(x0 = t.pr, x1 = t.pr, y0 = 0 , y1 = 0.25, col = "darkred", lwd = 2)
 text(x = t.pr, y = 0.3, paste("t =", round(t.pr, 2)), cex = 1.2)
 text(x = t.pr, y = 0.35, paste("p-val < 0.001"), cex = 1.2)
 ```
-```
diff --git a/SSR/topics/regression_one_predictor/regression_one_predictor.html b/SSR/topics/regression_one_predictor/regression_one_predictor.html
@@ -112,6 +112,7 @@ <h2 id="toc-title">Table of contents</h2>
   <li><a href="#predicted-values" id="toc-predicted-values" class="nav-link" data-scroll-target="#predicted-values">Predicted values</a></li>
   <li><a href="#y-vs-haty" id="toc-y-vs-haty" class="nav-link" data-scroll-target="#y-vs-haty"><span class="math inline">\(y\)</span> vs <span class="math inline">\(\hat{y}\)</span></a></li>
   <li><a href="#error" id="toc-error" class="nav-link" data-scroll-target="#error">Error</a></li>
+  <li><a href="#homoscedasticity-1" id="toc-homoscedasticity-1" class="nav-link" data-scroll-target="#homoscedasticity-1">Homoscedasticity</a></li>
   <li><a href="#model-fit" id="toc-model-fit" class="nav-link" data-scroll-target="#model-fit">Model fit</a></li>
   <li><a href="#explained-variance" id="toc-explained-variance" class="nav-link" data-scroll-target="#explained-variance">Explained variance</a></li>
   <li><a href="#explained-variance-visually-n-10" id="toc-explained-variance-visually-n-10" class="nav-link" data-scroll-target="#explained-variance-visually-n-10">Explained variance visually (<span class="math inline">\(n = 10\)</span>)</a></li>
@@ -142,30 +143,14 @@ <h2 class="smaller anchored" data-anchor-id="regression">Regression</h2>
 <p><span class="math display">\[\LARGE{Y_i = \beta_0 + \beta_1 X_i + \epsilon_i}\]</span></p>
 <p>In linear regression, the relationships are modeled using linear predictor functions whose unknown model parameters are estimated from the data.</p>
 <p>Source: <a href="https://en.wikipedia.org/wiki/Linear_regression">wikipedia</a></p>
-<!-- ## Outcome vs Model -->
-<!-- ```{r, fig.align='center', fig.asp=.5} -->
-<!-- set.seed(123) -->
-<!-- error = c(2, 1, .5, .1) -->
-<!-- n = 10 -->
-<!-- layout(matrix(1:4,1,4)) -->
-<!-- for(e in error) { -->
-<!--   x = rnorm(n) -->
-<!--   y = x + rnorm(n, 0 , e) -->
-<!--   r   = round(cor(x,y), 2) -->
-<!--   r.2 = round(r^2, 2) -->
-<!--   plot(x,y, las = 1, ylab = "y", xlab = "x",  -->
-<!--        main = paste("r =", r," r2 = ", r.2), ylim=c(-2,2), xlim=c(-2,2)) -->
-<!--   fit <- lm(y ~ x) -->
-<!--   abline(fit, col = "red") -->
-<!-- } -->
-<!-- ``` -->
 </section>
 <section id="assumptions" class="level2">
 <h2 class="anchored" data-anchor-id="assumptions">Assumptions</h2>
-<p>A selection from Field:</p>
+<p>A selection from Field (Section 8.4):</p>
 <ul>
-<li>Sensitivity</li>
-<li>Homoscedasticity</li>
+<li>Sensitivity (Section 6.4)</li>
+<li>Homoscedasticity (Section 6.7)</li>
+<li>Normality (Section 6.7)</li>
 </ul>
 </section>
 <section id="sensitivity" class="level2">
@@ -174,7 +159,7 @@ <h2 class="anchored" data-anchor-id="sensitivity">Sensitivity</h2>
 <ul>
 <li>Extreme residuals</li>
 <li>Cook’s distance (&lt; 1)</li>
-<li>Check Q-Q and residuals plots</li>
+<li>Check Q-Q, residuals plots, casewise diagnostics</li>
 </ul>
 </section>
 <section id="sensitivity-1" class="level2">
@@ -194,8 +179,8 @@ <h2 class="anchored" data-anchor-id="homoscedasticity">Homoscedasticity</h2>
 <div class="columns">
 <div class="column">
 <ul>
-<li>Variance of residual should be equal across all expected values</li>
-<li>Look at scatterplot of standardized: predicted values <span class="math inline">\(\times\)</span> residuals. Roughly round shape is needed</li>
+<li>Variance of residuals should be equal across all expected values</li>
+<li>Look at scatterplot of standardized: predicted values vs.&nbsp;residuals. Roughly round shape is needed</li>
 <li>After the analysis is complete because it’s based on the residuals</li>
 </ul>
 </div><div class="column">
@@ -213,6 +198,9 @@ <h2 class="anchored" data-anchor-id="homoscedasticity">Homoscedasticity</h2>
 </section>
 <section id="the-data" class="level2 smaller">
 <h2 class="smaller anchored" data-anchor-id="the-data">The data</h2>
+<blockquote class="blockquote">
+<p>Predict album sales (x 1,000 copies) based on airplay (no. plays).</p>
+</blockquote>
 <div class="cell">
 <div class="cell-output-display">
 
@@ -357,6 +345,21 @@ <h2 class="smaller anchored" data-anchor-id="error">Error</h2>
 </div>
 </div>
 </section>
+<section id="homoscedasticity-1" class="level2">
+<h2 class="anchored" data-anchor-id="homoscedasticity-1">Homoscedasticity</h2>
+<blockquote class="blockquote">
+<p>Is there any systematic error?</p>
+</blockquote>
+<div class="cell" data-layout-align="center">
+<div class="cell-output-display">
+<div class="quarto-figure quarto-figure-center">
+<figure class="figure">
+<p><img src="regression_one_predictor_files/figure-html/unnamed-chunk-17-1.png" class="img-fluid figure-img" width="480"></p>
+</figure>
+</div>
+</div>
+</div>
+</section>
 <section id="model-fit" class="level2 smaller">
 <h2 class="smaller anchored" data-anchor-id="model-fit">Model fit</h2>
 <ul>
@@ -387,7 +390,7 @@ <h2 class="smaller anchored" data-anchor-id="explained-variance-visually-n-10">E
 <div class="cell-output-display">
 <div class="quarto-figure quarto-figure-center">
 <figure class="figure">
-<p><img src="regression_one_predictor_files/figure-html/unnamed-chunk-19-1.png" class="img-fluid figure-img" width="1440"></p>
+<p><img src="regression_one_predictor_files/figure-html/unnamed-chunk-20-1.png" class="img-fluid figure-img" width="1440"></p>
 </figure>
 </div>
 </div>
@@ -460,7 +463,7 @@ <h2 class="anchored" data-anchor-id="section"></h2>
 <h2 class="anchored" data-anchor-id="p-values-of-b_1">P-values of <span class="math inline">\(b_1\)</span></h2>
 <div class="cell">
 <div class="cell-output-display">
-<p><img src="regression_one_predictor_files/figure-html/unnamed-chunk-24-1.png" class="img-fluid" width="672"></p>
+<p><img src="regression_one_predictor_files/figure-html/unnamed-chunk-25-1.png" class="img-fluid" width="672"></p>
 </div>
 </div>
 </section>
@@ -469,7 +472,7 @@ <h2 class="anchored" data-anchor-id="visualize">Visualize</h2>
 <p>Instead of obtaining the p-value by locating the t in the t-distribution, we can locate the F in the F-distribution</p>
 <div class="cell">
 <div class="cell-output-display">
-<p><img src="regression_one_predictor_files/figure-html/unnamed-chunk-25-1.png" class="img-fluid" width="672"></p>
+<p><img src="regression_one_predictor_files/figure-html/unnamed-chunk-26-1.png" class="img-fluid" width="672"></p>
 </div>
 </div>
 </section>

diff --git a/SSR/topics/regression_one_predictor/regression_one_predictor.qmd b/SSR/topics/regression_one_predictor/regression_one_predictor.qmd
@@ -18,46 +18,22 @@ In linear regression, the relationships are modeled using linear predictor funct
 
 Source: [wikipedia](https://en.wikipedia.org/wiki/Linear_regression)
 
-<!-- ## Outcome vs Model -->
-
-<!-- ```{r, fig.align='center', fig.asp=.5} -->
-<!-- set.seed(123) -->
-<!-- error = c(2, 1, .5, .1) -->
-<!-- n = 10 -->
-
-<!-- layout(matrix(1:4,1,4)) -->
-<!-- for(e in error) { -->
-
-<!--   x = rnorm(n) -->
-<!--   y = x + rnorm(n, 0 , e) -->
-
-<!--   r   = round(cor(x,y), 2) -->
-<!--   r.2 = round(r^2, 2) -->
-
-<!--   plot(x,y, las = 1, ylab = "y", xlab = "x",  -->
-<!--        main = paste("r =", r," r2 = ", r.2), ylim=c(-2,2), xlim=c(-2,2)) -->
-<!--   fit <- lm(y ~ x) -->
-<!--   abline(fit, col = "red") -->
-
-<!-- } -->
-
-<!-- ``` -->
-
 
 ## Assumptions
 
-A selection from Field:
+A selection from Field (Section 8.4):
 
-* Sensitivity
-* Homoscedasticity
+* Sensitivity (Section 6.4)
+* Homoscedasticity (Section 6.7)
+* Normality (Section 6.7)
 
 ## Sensitivity
 
 Outliers
 
 * Extreme residuals
 * Cook's distance (< 1)
-* Check Q-Q and residuals plots
+* Check Q-Q, residuals plots, casewise diagnostics
 
 ## Sensitivity
 
@@ -111,8 +87,8 @@ text(0.5, y[1], bquote("r ="~ .(r) ~ "," ~ r^2 ~ " = " ~ .(r.2)), pos = 2, col =
 
 ::: {.column}
 
-* Variance of residual should be equal across all expected values
-* Look at scatterplot of standardized: predicted values $\times$ residuals. Roughly round shape is needed
+* Variance of residuals should be equal across all expected values
+* Look at scatterplot of standardized: predicted values vs. residuals. Roughly round shape is needed
 * After the analysis is complete because it's based on the residuals
 :::
 
@@ -296,6 +272,19 @@ data$error <- round(error, 2)
 datatable(data, rownames = FALSE, options = list(searching = FALSE, scrollY = 300, paging = F, info = F))
 ```
 
+## Homoscedasticity
+
+>Is there any systematic error? 
+
+```{r, fig.height=5, fig.width=5, fig.align='center', echo=FALSE}
+plot(prediction, error, xlim = c(0,360), ylim = c(-200,200), main = "Predictions vs. Residuals",
+     xlab = "Predicted Sales", ylab = "Model Error (Residuals)",
+     las = 1, pch = 21, bg = "lightgreen")
+abline(lm(error ~ prediction), 
+       col='darkred')
+text(50, -150, bquote("cor = " ~.(round(cor(prediction, error), 5))))
+```
+
 ## Model fit {.smaller}
 
 - The fit of the model can be viewed in terms of the correlation ($r$) between the predictions and the observed values: if the predictions are perfect, the correlation will be 1. 

diff --git a/...one_predictor/regression_one_predictor_files/figure-html/unnamed-chunk-17-1.png b/...one_predictor/regression_one_predictor_files/figure-html/unnamed-chunk-17-1.png
diff --git a/...one_predictor/regression_one_predictor_files/figure-html/unnamed-chunk-20-1.png b/...one_predictor/regression_one_predictor_files/figure-html/unnamed-chunk-20-1.png
diff --git a/...one_predictor/regression_one_predictor_files/figure-html/unnamed-chunk-25-1.png b/...one_predictor/regression_one_predictor_files/figure-html/unnamed-chunk-25-1.png
diff --git a/...one_predictor/regression_one_predictor_files/figure-html/unnamed-chunk-26-1.png b/...one_predictor/regression_one_predictor_files/figure-html/unnamed-chunk-26-1.png