A direct marketing firm selling anti-virus subscriptions for personal computers has obtained a list of 1,000 potential customers along with an index that measures the propensity of each customer in terms of their "readiness to purchase".The list provider has offered the names at a price of $1 per name.The buying index ranges from 0-100 with higher index scores indicating a higher "readiness to purchase".The firm has sent an offer to all 1,000 names on the list to join service.The solicitation costs are $5 per customer.We observe whether or not each of the 1,000 consumers accepted the invitation: the value ofyis 1 if the person joined the service and the value is 0 otherwise.The firm creates an estimation and holdout sample byrandomlyassigning 500 consumers to each condition.The estimation sample is used to estimate a scoring model while the holdout sample is used to evaluate any model based target selection.A binary logistic regression model is fit to the estimation sample.The results are as follows:

\f