Analysis of data from four courses shows that (1) female open learning students complete their courses more often than male counterparts, (2) open learning students who register at certain campuses complete their courses more often than students who register at others, and (3) students who complete and return assigned questionnaires are more likely to complete courses than students who do not. From these data two formulae are derived which predict the number of student withdrawals. After being tested and refined, the two formulae are unified into one weighted index. Subsequent trials of the unified index result in seven of eight predictions being within (+ or -) 10%. Predictions were accurate over different presentation media. (WBC)