Background
Although bio-psycho-social health research is an ideal, samples adequate for complex modeling require biomarker specimens from hundreds of participants. Ecological sampling departs from laboratory study norms, with implications for analysis.
Objective
This paper compares salivary cortisol levels and effect sizes of ‘focal’ psychiatric factors, such as trauma history, posttraumatic stress diagnosis, comorbidity, and chronic stress, and ‘nuisance’ factors, including endocrine disorders, medications, physiological factors, such as gestational age, and smoking, to inform ecological study designs.
Study Design
This is a descriptive analysis of ecologically collected cortisol specimens, assayed in an on-going perinatal psychobiological study, addressing methodological considerations.
Results
Focal and nuisance factors are often interdependent with similar effect sizes. Careful specimen deletion decisions and model specification are needed to achieve the hoped-for external validity while maintaining internal validity.
Conclusions
Results of multivariate models support the validity and usefulness of an ecological approach to incorporating biomarkers in health research.