Application of classification tree approach to analysis youth marriage age gap

Background: During the last decades, the average gap between attitudinal and behavioral youths’ marriage age has increased due to the changes in Iranian society and family patterns. This paper is devoted to studying this increment. Methods: Classification and Regression Trees (CART) are applied for modeling the marriage age gap (MAG) of 12741 youths selected by a multi-stage cluster sampling method from 31 provinces in Iran. Results: Classification accuracies of fitted CART for females’ and males’ MAG were equal to .62 and .60, respectively. The most influential variables on females' and males' MAG were educational level and the number of siblings, respectively. Females with "university education," "diploma and less education with 5 and more siblings", and “employed diploma and less education with 3 or 4 siblings" married later than their desired time. Males with "3 and more siblings", "employed with 2 and fewer siblings and 3 and more ideal number of children", "employed university educated with 2 and fewer siblings and 1 or 2 ideal number of children", and " employed with 2 and fewer siblings and 1 or 2 ideal number of children with a diploma and less education and negative opinion towards childbearing" also married later than their desired time. Conclusion: If the inevitable experience of modernity doesn’t combine with the convenient policy and the economic and socio-cultural conditions of the community don’t change, the negative consequences of such developments would be more than its positive achievements on different social issues especially and more importantly youth’s marriage age.