Aimed at researchers familiar with the simple scoring method for predictive models (a la Peter Wylie's "Data Mining for Fundraisers"), this session looks at some of the limitations of the simple-score method and suggests the use of multiple linear regression instead. Intended only as an introduction to more advanced techniques, this session will cover topics such as data preparation, correlation, and computing predictive scores. Examples will use Data Desk stats software but the concepts are not software-specific. Minimal knowledge of statistics required.