Selective attention towards infants in nulliparous women across the menstrual cycle.

Research in women showed that testosterone is associated with decreased selective attention towards infant stimuli, which can be compensated for by oxytocin administration. In theory, caregiving behavior is thought to be mediated by oxytocin. Oxytocin binds to dopaminergic neurons and thus supposedly motivates aspects of caregiving through its influence on dopaminergic transmission. Most previous studies on caregiving behaviors were thereby performed in women under hormonal contraception to avoid hormonal fluctuations. However, recent studies repeatedly demonstrated decisive influences of the hormonal changes across the female menstrual cycle on dopamine-mediated behaviors, suggesting that estradiol acts as dopamine agonist in the follicular phase and progesterone as dopamine antagonist in the luteal phase. In the present study, we investigated selective attention towards infants as one central aspect of caregiving behavior over the natural menstrual cycle and in relation to interindividual differences of estradiol and progesterone. As expected, we found that women with higher estradiol in the follicular phase also showed higher selective attention towards infant faces among adult distractors, whereas the correlation disappeared in the luteal phase. In contrast, progesterone did not correlate with selective attention towards infants. The present findings collectively support the assumption that estradiol may act as dopamine agonist in the follicular phase, thereby supposedly promoting an important aspect of caretaking behavior.

Illusory inconsistencies in judgment: Stimulus-evoked reference sets and between-subjects designs.

Asked to judge the subjective size of numbers in a between-subjects design, participants rated 9 as larger than 221 (Birnbaum, 1999). The 9 > 221 effect seems to indicate that different stimuli evoke different contexts for comparison, and sounds a warning for the interpretation of between-subjects comparisons. We show that, contrary to appearances, the effect is not a result of stimulus-evoked reference sets. Instead, it is an artifact of the original 1-10 response scale and task instructions, which encourage a conflation of the response scale and the reference set. When ratings are expressed on a 1-1000 scale, or on a non-numerical slider scale, the effect reverses. However, we also show that stimuli can evoke their own comparative contexts, generating illusions of inconsistency in between-subjects designs. We report two novel findings - a 9 > 009 effect and a -2 > 2 effect - which are best explained by stimulus-evoked reference sets. Thus, while revealing that the 9 > 221 effect is an artifact of the original response scale, our study ultimately affirms Birnbaum's warning about the comparison of between-subjects ratings.

A cautionary note on the use of the Analysis of Covariance (ANCOVA) in classification designs with and without within-subject factors.

A number of statistical textbooks recommend using an analysis of covariance (ANCOVA) to control for the effects of extraneous factors that might influence the dependent measure of interest. However, it is not generally recognized that serious problems of interpretation can arise when the design contains comparisons of participants sampled from different populations (classification designs). Designs that include a comparison of younger and older adults, or a comparison of musicians and non-musicians are examples of classification designs. In such cases, estimates of differences among groups can be contaminated by differences in the covariate population means across groups. A second problem of interpretation will arise if the experimenter fails to center the covariate measures (subtracting the mean covariate score from each covariate score) whenever the design contains within-subject factors. Unless the covariate measures on the participants are centered, estimates of within-subject factors are distorted, and significant increases in Type I error rates, and/or losses in power can occur when evaluating the effects of within-subject factors. This paper: (1) alerts potential users of ANCOVA of the need to center the covariate measures when the design contains within-subject factors, and (2) indicates how they can avoid biases when one cannot assume that the expected value of the covariate measure is the same for all of the groups in a classification design.