1
$\begingroup$

I am conducting a meta-analysis with metafor in R. For the majority of articles, I managed to extract means and standard deviations, which I need for the escalc function to calculate an effect size.

However, many of the articles only provide means and within-subject standard error.

My questions: 1. is there a way to calculate standard deviation or standard error of the mean from the within subject standard error? 2. is there a way to calculate an effect size from only means and within subject standard error in metafor?

thanks!

$\endgroup$
  • $\begingroup$ Can you clarify what the design of the studies is so that it is clearer what within subject standard error means? $\endgroup$ – mdewey Nov 14 '18 at 16:26
  • $\begingroup$ Instead of standard error you could weigh studies by sample size, as long as the measurement approach is similar across studies. $\endgroup$ – Joe_74 Nov 16 '18 at 11:42
  • $\begingroup$ answer to @mdewey the studies are all experiments (cognitive psychology) with 2 levels of manipulations (high-reward distractors vs low reward distractors), the dependent measure is response time. within subject standard error: because basically no articles provide basic descriptives such as M and SD, I used a plot digitizer to read descriptives. in these plots, sometimes, the error bars represent within subject standard errors (meaning, standard error of the mean, corrected for between-subject variability -- I think). $\endgroup$ – DRusz Nov 19 '18 at 8:16
  • $\begingroup$ Well, if you have the standard error and the sample size you can back-calculate the standard deviation so I am still not sure what you are missing here. $\endgroup$ – mdewey Nov 19 '18 at 13:54

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.