Search code examples
rsweave

How to produce APA style statistical output in Sweave?

I am trying to find a way to integrate statistical output from the apa R package in Sweave.

I need to use the ez package for my statistical analysis (as I need type 3 sum of squares). This works perfectly with Rmarkdown, but I am having problems with the format of the output in Sweave. 

\documentclass{article}
\begin{document}
\SweaveOpts{concordance=TRUE}

\section*{Results}
<<echo=FALSE, results=hide>>=
library(knitr)
library(ez)
library(apa)

subject<- c(1:40)
reward<- rep(c("p", "s", "p", "s", "s", "p", "s", "p", "s", "p"), 4)
class<- rep(c("ST", "GT", "ST", "GT", "GT", "ST", "GT", "ST", "GT", "ST"), 4)
value<- runif(40, min=0, max=5)
df<- data.frame(subject, reward, class, value)
df$subject<- as.factor(df$subject)

analysis<- ezANOVA(data= df, dv=.(value), wid= .(subject),between = .(class), detailed= TRUE, type= 3)
apa.format<- apa(analysis, format= "text")
@

The results, \Sexpr{(apa.format[2,2])}, indicate...

\end{document}

The result it should produce is:
F(1, 38) = 0.93, p = .341, ηp² = .02 I have tried to change the format from the apa function, and every output produces a different issue:
* latex: produces textit (and it does not make the 'F' italic or the 'p')
* markdown: makes the the '*' instead of producing italic.
* text: does not produce the italic or the partial eta squared symbol.
This can be achieved easily in Rmarkdown, however I had many other issues. Suggestions? This may include: different package for APA reporting in Sweave, a different way to integrate Latex and Rcode (and use packages) or a fix to the previous output. Thanks!

Solution

  • The \Sexpr macro appears to "eat" the backslashes that apa() puts into the output, so you need to double them. So choose "latex" format, and then double the backslashes:

    apa.format <- apa(analysis, format= "latex")
apa.format <- gsub("\\", "\\\\", as.matrix(apa.format), fixed = TRUE)

    This fixes all entries of apa.format so they are compatible with \Sexpr. The as.matrix is necessary, because apa.format is a tibble, and it wouldn't work nicely with gsub(). (I imagine you could do the same using mutate or something like that, if you know what you're doing with tibbles. I don't.)

    Output is

    screenshot