Figure 5. Influence of less-than-perfect reliability of oxytocin measurements on the sample sizes required to detect significant endogenous oxytocin-outcome associations in neurobehavioral human oxytocin research.

In this figure, we show that the results of a set of simulations illustrating the impact less-than-perfect reliabilities of endogenous oxytocin measurements in peripheral fluids might have on the sample size required to detect significant oxytocin (A) – outcome (B) correlations of varying effect sizes in research studies. We conduct calculations for a minimally acceptable statistical power of 80% in a two-tailed parametric test. The outcome measure (B) was assumed to present perfect reliability. The figure was generated using the ‘pwr.r.test’ function of the ‘pwr’ R package. We specified ‘r’ according to the attenuation formula below (Nunnally, 1970). ICC -. Intraclass correlation coefficient; r – Pearson’s correlation coefficient.