Stouffer’s meta-analysis with weight and direction effect in R

May 15, 2018

(This article was first published on R – Fabio Marroni's Blog, and kindly contributed to R-bloggers)

I often have to perform meta-analysis of past experiments for which the only info I have is the fold change and the p-value (of any measure you may imagine: species richness, gene expression, depth of coverage, plates of carbonara eaten in 5 minutes, everything).
The hardest thing to find out for me was how to take the direction of the changes into account. E.g. if monday I eat 10 carbonara more than my brother (p=0.01), on tuesday 10 more (p=0.01), on wednesday 5 more (p=0.1), on thursday 10 less (p=0.01), on friday ten less (p=0.01) and on saturday 5 less (p=0.1), a standard Stouffer meta-analysis would return a highly significant p-value, completely disregarding the fact that the significant changes were half in my favor, and half in favor of my brother.
How can I take into account the information on the direction?
I had no idea on how to proceed, and I wasn’t able to find any reference (I am not an expert of meta-analysis), but then I was so lucky to stumble upon the manual page of the software package METAINTER for performing meta-analysis.

The authors described there a method to perform Stouffer meta-analysis accounting for the direction of the effects. Their explanation was so clear that it was easy for me to write a simple function in R: I paste it below.

signed.Stouffer.meta <- function(p, w, sign) { # p is a vector of p-values
  if (missing(w)) {
    w <- rep(1, length(p))/length(p)
  } else {
    if (length(w) != length(p))
      stop("Length of p and w must equal!")
  if(length(p)==1) Zi1)
  Z  <- sum(w*Zi)/sqrt(sum(w^2))
  p.val <- 2*(1-pnorm(abs(Z)))
  return(c(Z = Z, p.value = p.val))

To leave a comment for the author, please follow the link and comment on their blog: R – Fabio Marroni's Blog. offers daily e-mail updates about R news and tutorials on topics such as: Data science, Big Data, R jobs, visualization (ggplot2, Boxplots, maps, animation), programming (RStudio, Sweave, LaTeX, SQL, Eclipse, git, hadoop, Web Scraping) statistics (regression, PCA, time series, trading) and more...

If you got this far, why not subscribe for updates from the site? Choose your flavor: e-mail, twitter, RSS, or facebook...

Comments are closed.

Search R-bloggers


Never miss an update!
Subscribe to R-bloggers to receive
e-mails with the latest R posts.
(You will not see this message again.)

Click here to close (This popup will not appear again)