Academic journal article College and University

A New Look at Solving the Undergraduate Yield Problem: The Importance of Estimating Individual Price Sensitivities

Academic journal article College and University

A New Look at Solving the Undergraduate Yield Problem: The Importance of Estimating Individual Price Sensitivities

Article excerpt

Although colleges and universities interested in the most cost-effective way of diversifying their entering class or meeting enrollment targets have an array of analytical techniques to choose from, issues of accuracy, scope, and usability continue to slow their widespread use. To address these issues, this article develops an institutional enrollment probability model using logistic regression that explores the use of an alternative measure of average price sensitivity, one which produces and takes into account each individual student's price sensitivity. When the model was applied to the more than 13,000 students admitted over a four-year period to a private university located in the southwestern United States, the resulting average price sensitivity was significantly lower than that found in the empirical literature, which is hardly surprising given the non-linearity of the logistic distribution. In the final section, the full set of student-specific price sensitivities is used to describe how financial aid and admissions professionals can work together to maximize the value of each additional scholarship dollar.

For the last 30 years, the rapid growth in informational technologies has led to an absolute explosion in the production and diffusion of statistical software, which in turn has spawned the development of a set of analytical techniques known as enrollment management. For colleges and universities interested in the most cost-effective way of diversifying their entering class or meeting enrollment targets, this set of techniques-which includes the ability to predict the likelihood of an admitted individual enrolling in the institution-requires institutions to either construct their own predictive models or base their admission packaging decisions on the coefficients derived from some national or regional model. Although these models offer great promise when used effectively, issues of accuracy, scope, and usability continue to slow their widespread use (St. John and Somers 1997).

For example, for institutions forced to rely on estimates of price sensitivities derived from other institutions, there is always the issue of how relevant or accurate a national or regional estimate is for that particular institution. Furthermore, since there are a wide range of price sensitivities represented in the empirical literature (Heller 1997; Jackson and Weathersby 1975; Leslie and Brinkman 1987), the choice of any particular one as a basis for institutional policy is a decision wrought with peril-use an estimate too high and end up with empty seats, use an estimate too low and end up with too many students. Similarly, for those constructing their own institution-specific model, issues of aggregation and model specification make estimating price sensitivity a difficult and time-consuming exercise.

In addition to the difficulty in deciding which particular price sensitivity to use, there is also the very real problem of using this information easily and efficiently within an institution. Although the unit of analysis may vary from different groups of students (e.g., all early admits or all African-Americans males) to the individual students themselves, these price sensitivities need to be available to institutional policymakers in an easy-to-use manner that allows them to effortlessly target their limited supply of scholarship dollars. Unfortunately, since many administrators that work in the financial aid and admissions offices are often the most overworked on campus, developing a highly coordinated and easy to use interface between these two offices is rarely, if ever, accomplished. As a result, many admissions and aid decisions are made in a less than an optimal fashion.

In an effort to address these issues, we present results from an institution-specific predictive model that was built to solve the undergraduate yield problem at a private, religiously-affiliated university located in southwestern United States. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.