- Build evaluation in from the start of projects or programmes, for example
by making sure that the necessary data collection arrangements are in place.
- Consider how detailed the evaluation needs to be. Large or innovative
projects will justify more extensive evaluation than smaller projects based
on tried and tested approaches.
- Ensure there is a way of checking the standard of operation - i.e. that
the programme is being run as intended.
- Where the ultimate outcome is long term consider obtaining short or ‘intermediate’
as well as long term outcome measures. (In the case of offender programmes
these might include changes in attitude, substance abuse, qualifications
obtained etc, depending what the programme is trying to achieve.) These will
give more immediate results and help to illustrate how the programme
is working.
- Choose outcome measures to suit what the programme is designed to do.
- Make sure that sample sizes are sufficient to produce robust results.
- Aim to have a control or comparison group so robust analysis can be made
of the difference made by the programme.
- Make sure selection effects are accounted for. With offender programmes
this means looking at the outcomes for those not selected for the
programme or who drop out as well as those who complete it.
- Make sure that evaluators are not involved in running or evaluating the
programme being evaluated.
Based on a Home Office Guidance Note, Programmes for offenders: guidance
for evaluators, 1999