According to a study published in the Journal of Sex Research, relationship characteristics like relationship satisfaction, relationship length, and romantic love are among the top predictors of cheating within a relationship. The researchers used a machine learning algorithm to pinpoint the top predictors of infidelity among over 95 different variables.
While a host of studies have investigated predictors of infidelity, the research has largely revealed mixed and often contradictory findings. Study authors Laura M. Vowels and her colleagues aimed to improve on these inconsistencies by using machine learning models. This approach would allow them to compare the relative predictability of various relationship factors within the same analyses.
“The research topic was actually suggested by my co-author, Dr. Kristen Mark, who was interested in understanding predictors of infidelity better. She has previously published several articles on infidelity and is interested in the topic,” explained Vowels, a principal researcher for Blueheart.io and postdoctoral researcher at the University of Lausanne.
Vowels and her team pooled data from two different studies. The first data set came from a study of 891 adults, the majority of whom were married or cohabitating with a partner (63%). Around 54% of the sample identified as straight, 21% identified as bisexual, 11% identified as gay, and 7% identified as lesbian. A second data set was collected from both members of 202 mixed-sex couples who had been together for an average of 9 years, the majority of whom were straight (93%).
Data from the two studies included many of the same variables — such as demographic measures like age, race, sexual orientation, and education, in addition to assessments of participants’ sexual behavior, sexual satisfaction, relationship satisfaction, and attachment styles. Both studies also included a measure of in-person infidelity (having interacted sexually with someone other than one’s current partner) and online infidelity (having interacted sexually with someone other than one’s current partner on the internet).
Using machine learning techniques, the researchers analyzed the data sets together — first for all respondents and then separately for men and women. They then identified the top ten predictors for in-person cheating and for online cheating. Across both samples and among both men and women, higher relationship satisfaction predicted a lower likelihood of in-person cheating. By contrast, higher desire for solo sexual activity, higher desire for sex with one’s partner, and being in a longer relationship predicted a higher likelihood of in-person cheating. In the second data set only, greater sexual satisfaction and romantic love predicted a lower likelihood of in-person infidelity.
When it came to online cheating, greater sexual desire and being in a longer relationship predicted a higher likelihood of cheating. Never having had anal sex with one’s current partner decreased the likelihood of cheating online — a finding the authors say likely reflects more conservative attitudes toward sexuality. In the second data set only, higher relationship and sexual satisfaction also predicted a lower likelihood of cheating.
“Overall, I would say that there isn’t one specific thing that would predict infidelity. However, relationship related variables were more predictive of infidelity compared to individual variables like personality. Therefore, preventing infidelity might be more successful by maintaining a good and healthy relationship rather than thinking about specific characteristics of the person,” Vowels told PsyPost.
Consistent with previous studies, relationship characteristics like romantic love and sexual satisfaction surfaced as top predictors of infidelity across both samples. The researchers say this suggests that the strongest predictors for cheating are often found within the relationship, noting that, “addressing relationship issues may buffer against the likelihood of one partner going out of the relationship to seek fulfillment.”
“These results suggest that intervening in relationships when difficulties first arise may be the best way to prevent future infidelity. Furthermore, because sexual desire was one of the most robust predictors of infidelity, discussing sexual needs and desires and finding ways to meet those needs in relationships may also decrease the risk of infidelity,” the authors report.
The researchers emphasize that their analysis involved predicting past experiences of infidelity from an array of present-day assessments. They say that this design may have affected their findings, since couples who had previously dealt with cheating within the relationship may have worked through it by the time they completed the survey.
“The study was exploratory in nature and didn’t include all the potential predictors,” Vowels explained. “It also predicted infidelity in the past rather than current or future infidelity, so there are certain elements like relationship satisfaction that might have changed since the infidelity occurred. I think in the future it would be useful to look into other variables and also look at recent infidelity because that would make the measure of infidelity more reliable.”
The study, “Is Infidelity Predictable? Using Explainable Machine Learning to Identify the Most Important Predictors of Infidelity”, was authored by Laura M. Vowels, Matthew J. Vowels, and Kristen P. Mark.