Tdeveloping A Two-Phase Post-Stratified Inverse Sampling To Reduce The Nonresponse Bias. Case Study: Students’ Satisfaction Survey In Qatar University

Abstract

This research aims to develop a Two-phase Post-stratified Inverse Sampling (TPIS) to avoid or reduce the bias of nonresponse for students’ satisfaction survey of Qatar university. In the first phase, the students are partitioned into two groups of response and non-response. The second phase method of data collection is an intensified method to get information from nonresponse. An unbiased estimator of the mean of overall satisfaction and its variance estimator are developed. The efficiency of the estimator is assessed using a simulation study. Moreover, the efficiency of (TPIS) is assessed. The cased study results showed that the mean estimate of overall satisfaction was higher for the nonresponse groups than the response groups but the different was not statistically significant at level of 𝛼=0.05 but significant at level of 𝛼=0.10. However, the overall satisfactions were significantly different at 𝛼=0.05 for 13 majors out of 51 (25%). This results indicate that the non-response bias is exists. In addition, simulation study results show that the TPIS is more efficient than a SRS with equal effective sample size for TPIS. Based on the results some recommendation and suggestion are provided.