Mining philanthropic data: Models for predicting alumni/us giving at a medium-sized public master's university

The cost of providing a quality higher education in America has risen faster than the rate of inflation. State legislative appropriations and tuition have not provided the funds necessary to meet these costs. Increasingly, public institutions have had to turn to private funding sources to fill the gap. Alumni are the greatest source of voluntary support for higher education. Identifying which alumni are most likely to support their alma mater is an essential component of an efficient and successful development effort.;The purpose of this study was to develop models that predict alumni/us giving at a medium-sized, public master's university by using data mining techniques on existing databases. Although higher education institutions are typically proficient in collecting data, they have not been as effective in mining the data from different sources on campus.;This study combined data from Winona State University's alumni and donor databases with additional data provided by the Bernard C. Harris Publishing Company, Inc. (N = 37,393). One component of the Harris data is the Composite Donor Index (CDI), a proprietary system that ranks alumni on their ability to give based on demographic and biographic information collected by Harris. The study found that CDI and six additional variables were significant predictors of alumni giving.;The results of the analysis were used to develop two new models to predict whether an alumnus would give, and a third to predict the level of gift. The logistic model was able to predict 78.85% of alumni non-donors, and the discriminant model was able to predict 61.16% of alumni donors. A multiple regression model was able to target alumni who were likely to give in greater amounts.;The findings of the study show that mining existing data can significantly increase understanding of the giving potential of alumni. By more effectively segmenting the alumni database Winona State University will be able to focus limited development dollars on those alumni who are most likely to give. Practical applications for more effective direct mail campaigns, phone solicitations, additional prospect research, major gifts, and others exist because of database segmentation.