David K. Guilkey
Ph.D., Cary C. Boshamer Professor,
CPC Office: 137 E Franklin St, Room 506A
CPC Phone Number: (919) 962-6140
Campus Office: Gardner Hall, Room 208C
Campus Phone Number: (919) 966-5335
Dr. Guilkey's Curriculum Vitae
Dr. Guilkey's Personal Home Page
Dr. Guilkey's publications in PubMed
Dr. Guilkey's CPC publications
Guilkey is an applied econometrician with a microeconomics focus. Much of his work has involved the use of large survey data sets that involve limited dependent variables and the presence of endogenous right-hand-side variables. A well-known example of his work is a paper that examines the impact of Indonesia’s family planning program on fertility. This paper takes into account two complicating factors. First, the introduction of Indonesia’s family planning program was targeted, which means that methods that control for endogenous program placement must be used. Second, the paper treats the woman’s level of education as endogenous as well and allows family planning effects to operate through their effect on female education. A semi-parametric maximum likelihood estimation method was used to estimate equations for the timing of marriage, husband’s and wife’s completed years of education, and fertility. The results showed that standard methods understate program impact while overstating the impact of female education on fertility. These types of endogenity corrections have proved important in collaborations with Rindfuss and Morgan where the availability of childcare centers in Norway was endogenous with women’s first birth timing and overall childbearing. Controlling for this endogenity reversed a counterintuitive antinatalist effect to reveal a strong pronatalist one. The contribution of this work was recognized by the 2011 Award for Distinguished Contribution to Scholarship in Population (given by the ASA Population Section).
Guilkey contributes to the Population Health PRA through a set of collaborations. For instance, with Popkin and Gordon-Larsen, Guilkey conceptualized and implemented a structural equations modeling approach to measure the impact of physical activity on obesity, and ultimately health, in several different settings. In China, for example, drawing on almost two decades of data from the China Health and Nutrition Survey, and using generalized method of moments estimation methods to control for the endogeneity of key explanatory variables, Guilkey and his colleagues show that 30% of the weight gain among adult Chinese men was due to declines in physical activity, while 20% was due to higher fat intake.
Finally, as PI of the Gates Foundation Measurement, Learning, & Evaluation (MLE) Project, Guilkey contributes to the research area of reproductive health. The MLE project evaluates the Gates’ Urban Reproductive Health Initiative, which will introduce reproductive health programs in the urban areas of four countries (one in Asia and three in Africa) targeting the urban poor. Speizer, Stewart, Angeles, and Curtis collaborate on this effort, which has involved the design, collection, and analysis of longitudinal datasets in all four countries.
Guilkey will continue to exploit the high quality longitudinal data gathered in four countries for the Gates Project to explore a variety of topics. The impact evaluation strategy for the Gates Project is limited to evaluating the impact of country programs on current modern contraceptive use and straightforward fixed effects methods are used since they approximate a pretest/post-test experimental design. However, a more interesting line of research would be to estimate structural models that follow the pathways of program activities to an ultimate effect on fertility. The data in Nigeria and Kenya can also be coupled with other surveys to examine if program had diffuse effects outside of the urban areas where they were implemented. Guilkey also plans to continue his collaborations with Popkin and Gordon-Larsen to estimate models related to obesity using Cardia and Add Health data and his work with Morgan and Harris that is exploring a variety of topics using Add Health data.
Primary Research Areas:
Current Research Projects:
Information updated on 1/19/2017