Scholars connected to Bocconi’s Dondena Centre for Research on Social Dynamics and Public Policy recently used a machine learning (ML) technique called a machine learning (ML) technique to analyze data on 2,038 married or cohabiting couples who participated in the German Socio-Economic Panel Survey. The big news? It was able to predict the top reasons for couple dissolution: the woman’s percentage of housework and the life satisfaction of both partners.

According to the article published online on Demography, the couples who participated in the research were observed, on average, for 12 years. This translates to a total of 18,613 observations, which is a massive set of data that can be difficult to manage. 

“A clear-cut example of the potential difficulties of considering all variables and their possible interactions concerns the ‘big five’ personality traits,” said Professor Mencarini. “To account for both partners’ traits (10 variables) and all their two-way interactions (25 variables), one would need to include 35 independent variables, which would be very problematic in a regression model.”

Fortunately, Bruno Arpino (University of Florence), Marco Le Moglie (Catholic University, Milan), and Letizia Mencarini (Bocconi) were able to successfully manage a large number of independent variables in conventional models through RSF. According to them, ML tools can detect complex patterns in relatively small datasets.

To prove the superior predictive power of ML compared to conventional models, the authors split up the sample in their study into two parts. And using the results of the first half, they were able to predict the outcomes of the second half with accuracy. According to them, this indicates how effective the predictive accuracy of RSF is. However, it was also mentioned that the predictive accuracy of RSF was limited despite the use, as input variables, of all the most important predictors of union dissolution identified in the literature.

In the end, the study states that some of the variables with the greatest predictive ability are the man’s level of extraversion, woman’s percentage of housework, woman’s working hours, woman’s level of openness, marital status of the couple, and the life satisfaction of both partners.