expt with repaginating

Author: benwhalley

_output.yml

@@ -1,13 +1,14 @@
 bookdown::gitbook:
   includes:
     in_header: hypothesis.html
+  css: style.css
   config:
     toc:
       collapse: section
       before: |
         <li><a href="./">Just enough R</a></li>
     download: ["pdf", "epub"]
-  split_by: rmd
+  split_by: section
 
 bookdown::pdf_book:
   includes:

airquality-r-values.csv

@@ -0,0 +1,7 @@
+"","Ozone","Solar.R","Wind","Temp","Month","Day"
+"Ozone",0,0.00197241942881554,1.2981615782337e-11,0,0.561787014869315,1
+"Solar.R",0.000179310857165049,0,1,0.00751772924010297,1,0.561787014869315
+"Wind",9.27258270166931e-13,0.495955206815127,0,3.43407631220316e-08,0.247105980658337,1
+"Temp",0,0.000751772924010297,2.6415971632332e-09,0,7.23144291114863e-07,0.645698579067462
+"Month",0.0776000963996033,0.366353350873138,0.0274562200731485,6.02620242595719e-08,0,1
+"Day",0.887942543669527,0.0702233768586644,0.738746589753025,0.107616429844577,0.92218998575754,0

anova-cookbook.Rmd

@@ -8,12 +8,12 @@ output:
 knitr::opts_chunk$set(echo = TRUE, collapse=TRUE, cache=TRUE)
 library(tidyverse)
 library(pander)
+
 library(lmerTest)
 ```
 
 
 
-
 # Anova 'Cookbook'
 
 This section is intended as a shortcut to running Anova for a variety of common types of model. If you want to understand more about what you are doing, read the section on [principles of Anova in R](anova.html).
@@ -25,9 +25,8 @@ This section is intended as a shortcut to running Anova for a variety of common
 
 ### Factorial anova, no bigger than 2x2
 
-```{r}
 
-```
+TODO
 
 
 ### Factorial anova, where one factor has > 2 levels.
@@ -54,6 +53,9 @@ eysenck %>%
 ```
 
 
+[Visual inspection of the data (see Figure X) suggested that older adults recalled more words than younger adults, and that this difference was greatest for the intention, imagery, and adjective conditions. Recall peformance was worst in the counting and rhyming conditions.]{.apa-example}
+
+
 Or alternatively if we wanted to provde a better summary of the distribution of the raw data we could use a boxplot:
 
 ```{r, fig.cap="Boxplot for recall in older and young adults, by condition."}

correlations.Rmd

@@ -72,10 +72,10 @@ mycorrelations
 One thing to be aware of is that by default `corr.test()` produces p values that are adjusted for multiple comparisons in the top right hand triangle (i.e. above the diagonal). If you want the uncorrected values use the values below the diagonal (or pass `adjust=FALSE` when calling the function). 
 
 
-## Making correlation tables for publication
+## Tables for publication
 
 
-### `apaTables`
+### Using `apaTables` {-}
 If you want to produce output tables for publication the `apaTables` package might be useful. This block saves an APA formatted correlation table to an [external Word document like this](Table1_APA.doc).
 
 ```{r}
@@ -84,7 +84,7 @@ apa.cor.table(airquality, filename="Table1_APA.doc", show.conf.interval=F)
 ```
 
 
-### By hand
+### By hand {-}
 
 If you're not bothered about strict APA foramt, you might still want to extract the *r* and *p* values as dataframes which can then be saved to a csv and opened in excel, or converted to a table by some other means. 
 
@@ -108,7 +108,7 @@ mycorrelations$ci %>%
 ```
 
 
-## Other types of correlation
+## Other correlation methods
 
 By default `corr.test` produces Pearson correlations, but You can pass the `method` argument `psych::corr.test()`:
 

docs/a-real-example.html

@@ -8,7 +8,6 @@ output:
 
 
 
-
 # Anova 'Cookbook'
 
 This section is intended as a shortcut to running Anova for a variety of common types of model. If you want to understand more about what you are doing, read the section on [principles of Anova in R](anova.html).
@@ -21,6 +20,7 @@ This section is intended as a shortcut to running Anova for a variety of common
 ### Factorial anova, no bigger than 2x2
 
 
+TODO
 
 
 ### Factorial anova, where one factor has > 2 levels.
@@ -40,7 +40,10 @@ eysenck %>%
   ylab("Recall (95% CI)") + xlab("")
 ```
 
-<img src="anova-cookbook_files/figure-html/unnamed-chunk-4-1.png" width="672" />
+<img src="anova-cookbook_files/figure-html/unnamed-chunk-3-1.png" width="672" />
+
+
+[Visual inspection of the data (see Figure X) suggested that older adults recalled more words than younger adults, and that this difference was greatest for the intention, imagery, and adjective conditions. Recall peformance was worst in the counting and rhyming conditions.]{.apa-example}
 
 
 Or alternatively if we wanted to provde a better summary of the distribution of the raw data we could use a boxplot:
@@ -54,8 +57,8 @@ eysenck %>%
 ```
 
 <div class="figure">
-<img src="anova-cookbook_files/figure-html/unnamed-chunk-5-1.png" alt="Boxplot for recall in older and young adults, by condition." width="672" />
-<p class="caption">(\#fig:unnamed-chunk-5)Boxplot for recall in older and young adults, by condition.</p>
+<img src="anova-cookbook_files/figure-html/unnamed-chunk-4-1.png" alt="Boxplot for recall in older and young adults, by condition." width="672" />
+<p class="caption">(\#fig:unnamed-chunk-4)Boxplot for recall in older and young adults, by condition.</p>
 </div>
 
 
@@ -110,8 +113,8 @@ car::qqPlot(eysenck.model)
 ```
 
 <div class="figure">
-<img src="anova-cookbook_files/figure-html/unnamed-chunk-8-1.png" alt="QQ plot to assess normality of model residuals" width="672" />
-<p class="caption">(\#fig:unnamed-chunk-8)QQ plot to assess normality of model residuals</p>
+<img src="anova-cookbook_files/figure-html/unnamed-chunk-7-1.png" alt="QQ plot to assess normality of model residuals" width="672" />
+<p class="caption">(\#fig:unnamed-chunk-7)QQ plot to assess normality of model residuals</p>
 </div>
 
 
@@ -131,8 +134,8 @@ data_frame(
 ```
 
 <div class="figure">
-<img src="anova-cookbook_files/figure-html/unnamed-chunk-9-1.png" alt="Residual vs fitted (spread vs. level) plot to check homogeneity of variance." width="672" />
-<p class="caption">(\#fig:unnamed-chunk-9)Residual vs fitted (spread vs. level) plot to check homogeneity of variance.</p>
+<img src="anova-cookbook_files/figure-html/unnamed-chunk-8-1.png" alt="Residual vs fitted (spread vs. level) plot to check homogeneity of variance." width="672" />
+<p class="caption">(\#fig:unnamed-chunk-8)Residual vs fitted (spread vs. level) plot to check homogeneity of variance.</p>
 </div>
 
 
@@ -315,7 +318,7 @@ lme4::sleepstudy %>%
   geom_label(aes(y=450, x=9, label="imagine how bad\nyou feel by this point"), color="red") 
 ```
 
-<img src="anova-cookbook_files/figure-html/unnamed-chunk-13-1.png" width="672" />
+<img src="anova-cookbook_files/figure-html/unnamed-chunk-12-1.png" width="672" />
 
 
 If we want to test whether there are significant differences in RTs between `Days`, we could fit something very similar to a traditional repeat measures Anova using the `lme4::lmer()` function, and obtain an Anova table for the model using the special `lmerTest::anova()` function: