How might one define “relevant” in this context? I can think of a statistical definition and a pragmatic one.

Tom Goddard is a lawyer for a large HMO. He was concerned about the fact that government regulators assessed a fine of $500,000 against his company as a result of an audit of its practices. The reason for the concern was that the largest previous fine assessed to ANY HMO had been $200,000. Tom, therefore, sought to demonstrate that the half-million dollar fine was not consistent with auditing agency policy. He intended to do this by 1) generating a regression equation to predict amount of fine from relevant variables, and 2) using this equation to estimate the size of the fine that should have been levied against his company given its standing on the relevant variables. Discussion points are as follows. Your task is to respond to two of the following questions. If you choose a question to which others have responded, then you must give some thought to how your contribution is unique.

1. How might one define “relevant” in this context? I can think of a statistical definition and a pragmatic one.

2. Speculate about the sorts of variables that might be “relevant”.