Citation

Agerbo E, Stack S, Petersen L. Soc. Sci. J. 2011; 48(4): 630-640.

Copyright

(Copyright © 2011, Elsevier Publishing)

DOI

10.1016/j.soscij.2011.06.004

PMID

unavailable

Abstract

Research on the relationship between social integration and suicide rates has neglected a historical perspective. Analyses are often based on relatively short time periods where there may not be enough variation in integration to affect the national suicide rate, or where overall integration levels are not low enough to buttress a link between a specific index of low integration, such as divorce, and suicide. The present investigation contributes to the literature by testing a hypothesis on domestic integration and suicide over a century, the longest period studied to date, encompassing periods of low and high overall social integration, and emergent risk and protective factors. Data are available for core variables for Denmark, 1906-2006. Annual data include those on suicide, and indicators of integration (divorces, births, marriages), as well as for economic strain (unemployment) and time trends. A log linear Poisson model is estimated, which explores the central divorce-suicide relationship under controls for confounders. Controlling for confounders, a one percent increase in divorce increases male suicides by 0.52% and female suicides by 1.12%. As anticipated, marriages decrease suicide: a one percent increase in marriages reduces suicide by 0.77% for men and by 0.63% for women. The trend in divorce, in particular, offers accurate predictions of suicide throughout the century. The relationship between divorce and suicide over a century is robust. The results offer the strongest support to date in support of a social integration model based on long term historical data on suicide and divorce.