The most incisive guide to issues facing the American family today . . . An invaluable resource for anyone wishing to stay on the cutting edge of research on family trends.
-W. Bradford Wilcox
Associate Professor of Sociology, University of Virginia
What is perhaps one of the most effective ways to avoid divorce? Attend counseling? Read self-help books?
Or, perhaps, have more children?
A recent study coauthored by researchers at Auburn University at Montgomery in Montgomery, Alabama, indicates that the larger the household, the less likely it is that a couple experiences divorce. The authors argue that where other studies on divorce tend to focus on “spouses’ characteristics,” their study “goes beyond that conventional couple-level analysis and looks at the impact of macro-social variables, especially the association between household size and divorce rate at the county level.”
The researchers begin with the hypothesis that “counties with larger average household size will have lower rates of divorce.” A larger household size, they speculate, would serve as a form of social capital. They glean their data from the 1990 and 2000 census results, using a random sampling of 20% of counties in each state.
The researchers measure such factors as “average household size, religious homogeneity, percent population change, percent urban residence, percent unemployment, percent employment in manufacturing occupations, median family income, percent population aged 15-34 years, and region of residence in the United States, i.e. Northeast, South, Midwest, Plains and West.” All of these have been shown to have some correlation with the variability of divorce rates.
A multiple regression analysis of the 1990 data reveals exactly the correlation that the researchers expected: “When the 1990 divorce rate is regressed on the social and economic factors, the average household size is a statistically significant correlate (beta = -.23).” The larger the household, the less likely it is that household has experienced divorce. An analysis of the 2000 data revealed a similar pattern. Seven of the eight variables (with the exception of median family income) are significantly related to the divorce rate, and “average household size gain shows a significant, inverse association with divorce rate.”
When adding region into the equation for the data from 1990, the researchers found that three of the nine independent variables are positively associated with divorce rate—increase in percent population size, percent urban, and percent unemployed. Household size and religious homogeneity, however, were inversely associated with divorce rate. That is, the larger the households, and the greater the religious homogeneity within a county, the less likely it is that households in that community experience divorce. An analysis of the 2000 data revealed similar results. The researchers found that “Average household size, percent population age 15-34 years, and religious homogeneity index were negatively associated with divorce rate in all regions.”
The results were strong and clear: “Overall, average household size was one of the most important, and in one instance, the single most important, correlate of divorce rate at the county level. As hypothesized, our results showed that counties with larger household size had significantly lower rates of divorce. This finding was consistent across time and place.” As the researchers expected, the more people in the household, the more resources that couple had to stick together.
While this study offers valuable insight into the importance of large families in sustaining marriage, it stops short of naming children as the reason married couples stay together. While “larger household” could indeed involve grandparents, cousins, and other extended family, in this age of the nuclear family it seems more likely that children are the primary source of increased household size. Yet another reason to have lots of them.
(Yanyi K. Djamba et. al., “Household Size as a Correlate of Divorce Rate: A County-Level Analysis,” Sociological Spectrum 32.5 [July 2012]: 436-448.)