Research Within Reach
CURRENTS Article Kevin MacDonell's "CoolData" blog focuses on predictive modeling in higher ed fundraising and the process of determining which variables to look at and which are most predictive.

Hidden Treasures
CURRENTS Article Sometimes the best information you have on the best potential donors is right under your nose. Some prospect researchers are sniffing out their best donors by using data mining techniques, and they've discovered some interesting things.

A Wealth of Knowledge
CURRENTS Article Fund raisers often fail to leverage the wealth of donor data they have at their fingertips, but the secrets to winning over their never-givers often lie in those numbers. Peter B. Wylie's 2004 CASE book "Data Mining for Fund Raisers" offers tips for data analysis to improve fund raising return on investment.

Talking Points: Grants Today, Gifts Tomorrow
CURRENTS Article To determine how student aid affects giving after graduation, researchers at Vanderbilt University analyzed factors that influence young alumni giving. Data on 2,822 graduates over eight years showed that receiving a need-based scholarship raised a graduate’s likelihood of giving by 12 percent, while need-based loans reduced the probability by the same degree. The amount of aid had no effect.

Model Behavior
CURRENTS Article Statistical modeling can make fund-raising efforts more productive by predicting which groups of donors are most likely to respond to appeals. The process, which involves segmenting donors based on behavior and traits, makes solicitations more cost-effective, produces a high return from a minimal investment, and shows its worth quickly. Statistical modeling requires input from a fund-raising expert (usually the annual fund director) and both a computer expert and a statistician (who may be consultants or campus employees). Wylie, a fund-raising consultant, describes the modeling process and provides examples. In a sidebar, he answers common questions about statistical modeling.