A comparison of meta-analytic approaches on the consequences of role stressors
Hoover, Edward Rickamer
MetadataShow full item record
The use of conducting meta-analytic work is on the rise, yet analysts tend to overlook differences between approaches and simply select one that is common to the field of study. This study contrasts two popular meta-analytic approaches based on work by Hedges and Olkin (1985) and Hunter, Schmidt, and Jackson (1982) and compares them both in theory and in application. Broadly, while Hedges and Olkin (1985) first corrects for statistical biases, Hunter et al. (1982) uses the biased estimates but corrects for statistical artifacts prior to integration. Conceptual and statistical differences between these approaches lead to numerous disparities between estimates. Ultimately, it is difficult to provide guidelines for which meta-analytic approach is the best as it differs by scenario and the information available. This decision must be made holistically by examining the assumptions behind the nature of the constructs in question, the characteristics of the data set, and possible types of inaccuracies that must most be guarded against.