What matters when? A multiversal investigation of pubertal timing measurement and mental health

Natasha Chaku,Dominic Kelly

DOI: https://doi.org/10.1016/j.psyneuen.2023.106701

IF: 4.693

2024-02-01

Psychoneuroendocrinology

Abstract:Background Pubertal timing is an important marker of future wellbeing. Yet, how researchers use and interpret pubertal measures is a matter of active debate. The current study uses multiversal analyses to begin to unravel this question. Methods Data were drawn from the Study of Early Childcare and Youth Development (N=932, 51% female). Puberty was assessed yearly by clinicians using Tanner staging and mother questionnaires (Tanner line drawings, Pubertal Development Scale). Youth reported on their internalizing behaviors. For each measure, estimates of pubertal timing were indexed via logistic growth curves, by categories (“early”, “on-time” or “late” at each time point), and from residuals (regressing pubertal stage onto age at each time point). Specification curve analysis (90 conditions) was used to assess how associations between pubertal timing and internalizing behaviors were impacted by these decisions. Results Each pubertal measure had consistent associations with internalizing behaviors - for both sexes, earlier timing was associated with more internalizing behaviors. Critically, this was only true when timing was indexed using logistic growth curves. When categorical and residual estimates of timing were used, findings were domain- specific. Indeed, when puberty was categorized, it was only associated with internalizing behaviors at age 12 for girls and not at all for boys. Conclusions Parent-reported timing and clinician-derived estimates were similar; thus, more accessible measures may be appropriate for many research questions. The age at which that data is collected and how it is analyzed, however, has a bigger impact on inferences. These implications will be discussed during the talk.

neurosciences,endocrinology & metabolism,psychiatry