Abstract

The sample size in a quantitative study does not only affect statistical power but also the precision with which effects can be estimated. Schönbrodt and Perugini (2013) have applied Monte-Carlo simulations to establish the point of stability of correlation coefficients (i.e., the sample size from which on they only fluctuate around the true value within acceptable margins). They reported that the sample size necessary to achieve stability depends on the size of the underlying correlation, the width of the margins within which fluctuations are tolerated (the corridor of stability), and the desired confidence that the correlation does not exceed these margins. According to their suggestion, a sample around 250 is desirable for typical scenarios in personality psychology. The present contribution aimed to replicate these findings and extend them by determining the point of stability for rho = .0 and very narrow corridors of stability. Results pertaining to the replication were virtually identical to those reported by the original authors. In addition, correlations of .0 to became stable within a corridor of .1 at around 260 participants. Considerably more data points should be collected if only correlations within a very narrow corridor of stability are to be accepted. Future research could extend these findings by investigating the point of stability in data with different types of non-normal distributions.