Are BISP Beneficiaries have Better Resilience and Food Security? A Critical Review of Regression Discontinuity Design Evaluation Approach
Author: Seemab Riaz

RDD is a quasi-experimental regression discontinuity design. When reviewing new policies and treatments, this strategy is utilized to reduce confounding bias. This method is used when a policy intervention has been assigned to people based on whether they are above or below a pre-determined cut-off on a continuous variable. The proposed study evaluated the assumption of RDD and also analyze the impact of BISP’s cash transfers on food security and nutrition in the HH by using Oxford Policy Management BISP’s impact evaluation panel survey (OPM) 2019. Primary data were collected by conducted qualitative research such as conducted in-depth interviews from BISP beneficiaries to assess what poor needs, proposed study were conducted in-depth interviews 30 BISP beneficiaries from the Islamabad and Muzaffarabad tehsil office. This study found that, where as in the context of our investigation, the RDD assumption was generally met. According to the findings, BISP unconditional cash transfers have no substantial influence on food security and nutrition. Finding of the study suggested that with the varying bandwidth leads to increase the value of standard errors. Results of the RDD analysis specify that when number of covariates increases then the value of coefficient also increase. Standard error is depend on number of regression coefficient, number of data points and deviation of data sets from assumed regression model. Thus, for a given data set, the standard error also increase when we increase no of regression co-efficient. So, the results of different round is differ to each other because of different data sets and different regression co-efficient, different covariates in the regression model and also by inclusion and exclusion of co-variate in the model. The necessity of examining the validity of RDD design assumptions, testing them, and making adjustments to promote reliable causal inference is demonstrated our findings. Supervisor:- Dr. Shujaat Farooq

Meta Data

Supervisor: Shujaat Farooq

Related Thesis​