Premium
Note About Possible Bias Resulting When Under‐Statisticized Studies are Excluded from Meta‐Analyses
Author(s) -
Stanley Julian C.
Publication year - 1987
Publication title -
journal of educational measurement
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.917
H-Index - 47
eISSN - 1745-3984
pISSN - 0022-0655
DOI - 10.1111/j.1745-3984.1987.tb00263.x
Subject(s) - meta analysis , statistics , econometrics , simple (philosophy) , publication bias , correlation , computer science , psychology , mathematics , medicine , epistemology , philosophy , confidence interval , geometry
Reviews and meta‐analyses of research on a given topic may exclude' a sizable percentage of reports because they do not lend themselves to the type of summarizing procedures used. If the excluded articles contain relevant information, this may bias the conclusions of the analysis. It seems likely that, when computing statistics from their data, researchers will need to consider this aspect. A simple illustration of how that can sometimes be done readily is presented. A robust correlation coefficient easily computable from published data is shown to indicate a sizable relationship that is contrary to the main conclusion of a meta‐analysis.