"Crossover" in Depressive Symptoms Among Older Couples: Are Previous Findings Artifactual?

Objective: This study examined effects of a partner's depressive symptoms on those of a focal respondent. Method: Data were from the 2008-through-2014 waves of the Health and Retirement Study. Linear growth modeling was used to distinguish within- from between-person variation, and genetic instrumental variables to take endogeneity into account. Results: A partner's 2008 depressive symptoms had no associations with within-person change in those of a respondent between 2010 and 2014. Corresponding linkages with their between-person variation in 2010 were positive and significant for both genders. Among women, these last estimates remained intact in genetic instrumental variable models restricted to white couples. Discussion: Results indicate that contrary to previous literature, late life relationships are not marked by "crossover" of depressive symptoms. Women's affect may, however, be influenced by such transmission in the relationship's history. Beyond depression, findings suggest current theories of dyadic health effects may partly rest on flawed evidence.