الملخص الإنجليزي
Propensity score (PS) is the probability of treatment assignment conditional on ob served baseline characteristics. PS is a balancing score, that is, conditional on the PS,
the distribution of observed baseline covariates can be similar between treated and un treated subjects. Propensity Score Matching (PSM) refers to the pairing process of
treated and untreated units with similar PS. There is a lack of systematic research in
the current literature comparing the matching methods in PSM. On the other side, there
are few applications of PSM methods to design and analyze marketing treatments, de spite its powerful ability in advertising to target particular customers that could led
to decrease the huge financial losses. This study primarily focuses on evaluating the
PSM mechanisms and its application to personalized digital marketing in Banking in stitutions within Oman and beyond. The six most popular PSM approaches (Nearest
Neighbor, Nearest Neighbor with Caliper, Optimal, Genetic, and Subclass) have been
evaluated based on three di↵erent schemes, using primary survey data that were col lected in the course of this study. Moreover, a new approach, the Paired-samples t
test (PStT) has been applied to assess the quality of matching for the PSM. Further more, matched data from PSM' preferable method has been utilized to develop statis tical model for banking customer conversion behavior. Findings of this study, based
on bank customers' survey data, showed that the use of PStT with treatment group
covariance gives a more precise conclusion in line with previously used assessment
methods. Further, using various performance indicators, the Nearest Neighbor with
Caliper (NNWC) outperformed other PSM methods. The resulting predictive model
using NNWC has a high accuracy, and can be used to predict bank customers' conversion behavior. We recommend the banks to use this method to analyze marketing
campaigns and to predict customers behaviors.
