1- Department of Psychology, Sav.C., Islamic Azad University, Saveh, Iran
2- Department of Educational Psychology, Isl.C., Islamic Azad university, Islamshahr, Iran ,zrabbani@iau.ir
3- Department of Clinical Psychology, Isl.C., Islamic Azad university, Islamshahr, Iran
4- Department of Educational Psychology, Isl.C., Islamic Azad university, Islamshahr, Iran
2- Department of Educational Psychology, Isl.C., Islamic Azad university, Islamshahr, Iran ,
3- Department of Clinical Psychology, Isl.C., Islamic Azad university, Islamshahr, Iran
4- Department of Educational Psychology, Isl.C., Islamic Azad university, Islamshahr, Iran
Abstract: (23 Views)
Objective: This study aimed to examine the role of personality traits based on the Five-Factor Model in predicting marital relationship quality, and to compare linear regression and machine learning approaches for this prediction.
Methods: This descriptive–correlational study was conducted with 392 couples (784 individuals) aged 25 to 55, who were married and had referred to counseling centers in Tehran. Participants were selected through convenience sampling and completed the NEO-FFI and the Revised Dyadic Adjustment Scale (RDAS). Dyadic data were analyzed using the Actor–Partner Interdependence Model (APIM), and linear regression models were compared with several machine learning algorithms to predict relationship quality.
Results: Personality traits collectively explained approximately 6% of the variance in relationship quality. Among the personality traits, neuroticism and agreeableness were the strongest predictors, and the interaction effect between these two traits also made a significant contribution to predicting relationship quality. In the Actor–Partner Interdependence Model analysis, actor effects (related to individuals' own characteristics) were significant, whereas partner effects (related to their partners' characteristics) were not significant. In comparing prediction models, machine learning approaches showed performance similar to linear regression, with no significant superiority observed over the traditional statistical model.
Conclusions: Personality traits — particularly neuroticism and agreeableness, along with their interaction — contribute modestly but meaningfully to predicting marital relationship quality, with individuals' own traits playing a more consistent role than their partners' traits. The comparable performance of machine learning and linear regression approaches suggests that, for this prediction task, more complex modeling techniques do not necessarily offer added value over traditional statistical methods.
Methods: This descriptive–correlational study was conducted with 392 couples (784 individuals) aged 25 to 55, who were married and had referred to counseling centers in Tehran. Participants were selected through convenience sampling and completed the NEO-FFI and the Revised Dyadic Adjustment Scale (RDAS). Dyadic data were analyzed using the Actor–Partner Interdependence Model (APIM), and linear regression models were compared with several machine learning algorithms to predict relationship quality.
Results: Personality traits collectively explained approximately 6% of the variance in relationship quality. Among the personality traits, neuroticism and agreeableness were the strongest predictors, and the interaction effect between these two traits also made a significant contribution to predicting relationship quality. In the Actor–Partner Interdependence Model analysis, actor effects (related to individuals' own characteristics) were significant, whereas partner effects (related to their partners' characteristics) were not significant. In comparing prediction models, machine learning approaches showed performance similar to linear regression, with no significant superiority observed over the traditional statistical model.
Conclusions: Personality traits — particularly neuroticism and agreeableness, along with their interaction — contribute modestly but meaningfully to predicting marital relationship quality, with individuals' own traits playing a more consistent role than their partners' traits. The comparable performance of machine learning and linear regression approaches suggests that, for this prediction task, more complex modeling techniques do not necessarily offer added value over traditional statistical methods.
Keywords: Marital quality, Five-Factor Model of personality, Actor–Partner Interdependence Model (APIM), Machine learning, Linear regression
Type of Study: Original |
Subject:
Evolutionary Psychology
Received: 2026/01/23 | Accepted: 2026/04/7
Received: 2026/01/23 | Accepted: 2026/04/7
Send email to the article author
| Rights and permissions | |
 |
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License. |
