In personality (psychology) there exist multiple questionnaires around the same domains. The domains are the well known big five domains: neuroticism, extroversion, openness, agreeableness and conscientiousness. Suppose I am interested in the characteristic personality of mathematicians and have found multiple studies reporting the mean and standard deviation (SD) for people having this occupation. Some of these studies report only the mean and SD, while other studies report also comparisons with different groups. For example, the mathematicians are compared with civilians or economists.

I would like to aggregate the existing knowledge on the personality of mathematicians. This, to me, sounds like a meta-analysis. I could compare effect sizes or means and SDs. By the ladder of causation (Pearl 2009), it would make sense to focus on effect sizes, since differences between groups in one study are probably the result of a "do"-action. To me it seems that summarizing many questionnaires can also be insightful. The aggregation will mitigate confounding variables. However, both Borenstein et al. (2011) and Cooper, Hedges, and Valentine (2019) seem to focus on effect sizes and not on comparing means and SDs.

So, my questions are:

Borenstein, M., L. V. Hedges, J. P. T. Higgins, & H. R. Rothstein. 2011. Introduction to Meta-Analysis. John Wiley & Sons. https://doi.org/10.1002/9780470743386.

Cooper, H., L. V. Hedges, & J. C. Valentine. 2019. The Handbook of Research Synthesis. Russell Sage Foundation. https://doi.org/10.7758/9781610448864.

Pearl, J. (2009). Causality. Cambridge university press. https://doi.org/10.1017/CBO9780511803161.

  • 2
    $\begingroup$ The study designs based on your description are observational in nature and therefore have inherent biases and confounding factors that should have been adjusted for in the primary studies. Even so, not all will have adjusted for the same confounding factors (or adjusted at all) making pooling the evidence quantitatively difficult. You can look into systematic reviews with network meta-analysis as an option but it's not something you can pick up over a weekend. It may be best to find researchers already doing this type of research and collaborate with them. $\endgroup$
    – abousetta
    Commented Jun 3, 2020 at 5:38


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Browse other questions tagged or ask your own question.