What would have happened otherwise?

A town introduces a new evening bus service and later reports that more residents visit the central shopping area. The increase may be welcome, but it does not by itself show how much the bus service contributed. Other changes, such as longer shop opening hours, could also affect visits.

To evaluate the service, an analyst needs an estimate of what would have happened without it. That situation cannot be observed directly for the same town at the same time. One approach is to examine a similar town where no new bus service was introduced. If visits increase in both places, part of the change may reflect a wider trend.

The comparison still depends on judgment. The towns may differ in ways that affect the result, and a new event in only one town could make an earlier similarity less useful. Looking at patterns before the service began helps assess whether the comparison is reasonable, but cannot guarantee that the towns will develop identically afterward. Evaluation therefore involves testing an explanation against plausible alternatives. A rise following a policy is evidence worth investigating, not a complete demonstration that the policy produced the entire rise.