Data quality affects meta-analytic conclusions: a response to Miner and Raju (2004) concerning entrepreneurial risk propensity.

J. B. Miner and N. S. Raju (2004) contended that there are additional studies that compare the risk propensity of entrepreneurs and managers, which, when added to the data from W. H. Stewart and P. L. Roth (2001) and meta-analyzed, produce relationships between risk propensity and entrepreneurial status that are substantially weaker than previously believed. This conclusion was evaluated by identifying and examining methodological problems associated with their inclusion of effect sizes (i.e., ds) from studies with variables not relevant to research questions, dependent samples, extraneous variance in outcome variables, and confusion of constructs. When these methodological issues were addressed, a new meta-analysis indicated an overall conclusion consistent with that of Stewart and Roth, but the results varied according to instrumentation, particularly objective instrumentation (observed d = 0.31) versus the Miner Sentence Completion Scale-Form T (observed d = -0.35).