I am trying to assemble a meta-analytic dataset for computing yi and vi for metafor. For this dataset, I am including paired data (i.e., one sample of participants that completed upright and inverted face recognition task (% correct)). The problem is that half of the studies include the M, SD, and N, which can be used to compute the "standardized mean change" score in escalc() as shown below.

escalc(measure = "SMCR", m1i = upright_mean, sd1i = upright_sd,
                 m2i= inverted_mean, sd2i = inverted_sd, ri = ri, ni= N, data = rawData)

The other half of the studies only include the paired t-statistic and/or p-value. I would like to know if it is possible to assemble all of these studies into the same measure of Cohen's d, yi, and sampling variance, vi. and if so, how to accomplish that? I have not been able to figure out how to compute both yi and vi when only the paired t-statistic is available.

  • $\begingroup$ If a reasonable number of studies have both raw and derived estimates of effect you could try multivariate meta-analysis (or as a last resort missing data imputation). $\endgroup$ – Joe_74 Jul 12 at 11:56

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.