Surveys done for health services and economics research purposes, including those taken as part of clinical trials, often suffer from poor response levels or attrition and this may prompt researchers to weight the data. Survey results may also not reflect situations where the interview subjects have similar characteristics, for example when the respondents are predominantly women.
In an Oct. 16 CHEPA seminar, economics PhD candidate Qing (Felix) Li will describe an approach that combines both population and survey data using a method of moments technique that matches auxiliary moments of the two data sources to estimate micro-econometric models. He provides Monte Carlo evidence to show that the approach can appreciably reduce both bias and variance.
Li will show how this technique can be used in a health human resource context by describing how it was used to analyze factors associated with students’ decisions to drop out of the Ontario Midwifery Education Program (MEP). The high dropout rate in this academically and professionally challenging program is causing a lack of access to midwifery care for an estimated 2,425 Ontario women each year.
A survey conducted in 2007 was administered to midwifery students, graduates and dropouts to determine the factors prompting students to quit the program and provided the data Li used for his research. He will explain how his analysis approach can be widely applicable in health economics and health services.
Li is a PhD student in economics at McMaster who has research interests in health economics, labour economics and econometrics. He has an MA in economics from the University of Waterloo and BA in economics from Western University.
The seminar will be held Wednesday Oct. 16 from 12:30 in CRL B119, from 12:30 p.m. to 1:30 p.m. All are welcome to attend. The seminar will be available remotely for those who are unable to attend.