Todd Stinebrickner on the Berea Panel Study

This week we have Professor Todd Stinebrickner from University of West Ontario as our ERID visitor. He talked about the Berea Panel Study  he initiated to investigate college dropout decisions.

Professor Stinebrickner first pointed out the difficulty to find data catering to the needs of economists. Labor economists cares not only about individual outcomes (education, earnings, health, etc.) but also the decision process governing these patterns. Unfortunately, the majority of government-organized longitudinal studies do not measure people’s decision process. For example, the census has data on whether an individual finished college but is silent on what happens between college entrance and dropout. Thus, the exact decision-making process remains a mystery.

This insufficiency of existing data motivated Stinebrickner to collect his own data that will answer the question he was interested in: determinants of college dropout decisions. He located his study in Berea College in Kentucky where his father Ralph Stinebrickner had been teaching. With the help of school administrators, they conducted a preliminary survey in 1997 which focused on students from low-income families. Their results suggest that factors other than direct costs are important in dropout decisions, thus justifying the need for systematic longitudinal surveys. The Stinebrickner’s desgined their survey with an emphasis how beliefs impact the decisions of students. Expectation is a central to their investigation. To track the dynamic decision making process, they surveyed students 10 to 12 times a year. Post-college surveys are conducted annually.

With a solid research design in place, finding administrative support to implement it turned out to be nontrivial. On their first attempt, they made the baseline survey an assignment in a mandatory course that all freshmen were taking. But due to the time commitment required for the course instructor, this didn’t work out. “That’s why we call it a ‘pilot’ study,” smiled Stinebrickner. They then shifted to surveying students in class with a decent amount of compensation (around $25). The participation rate was as high as 87%. For subsequent surveys in school, they put the questionnaires in mailboxs and students can get paid upon completing and returning the questionnaire. This method worked out well as the response rate was 80%-90%. Post-school surveys are much harder to monitor, as the school no longer has direct contact with the students. However, with the help from Berea College, the Stinebrickner’s were able to mail their surveys to the students’ most up-to-date address with a check. A respondent can redeem the check by filling out the survey and mailing it back to the researchers. This method also proved to be effective and the response rate is currently 80%.

Using the data collected, Stinebrickner conducted a series of analyses on college dropout decisions, college major choice, and networks. The value of the data lies in their detailed documentation of the expectations of the students on a dynamic basis. This allows researchers to link beliefs about certain variables to individual future choices and model the decision making process better. This elicitation of beliefs may also be used as instrumental variables for future fertility choices and job hunting behaviors.

Professor Stinebrickner’s talk offers me a fresh view on the role that academics can play on data collection and informing the public. While it is always rewarding to witness a research idea come to fruition, the intensive intellectual exercise and implementation work required is often beyond the capacity of many.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s