Donald Kirkpatrick's 4-stage model of evaluation (Reaction - Learning - Behaviour -Results) is described. The author argues that many practitioners are more comfortable with the idea of evaluation rather than impact analysis and that Kirkpatrick's levels, particularly Level 4 - Results, relate to impact analysis.

There is an interesting example of analysis conducted after an Insight into the Media course which highlights the difficulty of finding suitable ways of measuring impact, particularly outcomes such as participants deciding a media career was not for them.

The problems caused by evaluation are described. These include the observation that data can be used defensively to bolster complacency rather than a prompt for examining the need for improvement.