The present study adopts a meta-analysis approach. It examines the effect of the research design characteristics on reliability coefficients: Cronbach’s alpha and Jöreskog’s rhô. Two independent variables are examined; namely, sample size and item number. The effect of these variables on reliability coefficients is also examined by using five moderator variables, namely: sample type, respondent type, scale type, category number and construct type. The results suggest that the number of items affects the two aforementioned reliability coefficients while sample size has an effect only on alpha. The results also prove that sample type, respondent type, scale type and category number have a moderator effect between the dependent and independent variables. The type of construct, however, does not have a moderation effect between the predictor and criterion variables.

reliability, alpha coefficient, rhô coefficient, sample size, number of items.