Establish appropriate baseline information, then measure results
You need baseline information to assess whether change has happened, and the extent of that change. Measuring the difference between the groups you target and relevant control groups, before and after your intervention, lowers the risk that your results are biased by differences between these groups that are unrelated to your intervention.
Be smart about baselines
In considering establishing baselines, ask yourself the following three key questions:
Have you made the most of the information gathered during the diagnostic process?
Draw on your understanding of the present performance and under-performance of those players your interventions are aiming to effect changes in, to derive baseline information at the system-level. This saves effort and keeps strategy and measurement coherent: diagnose down and measure up.
Are you putting too much faith in a single programme-wide baseline study?
It is better to conduct several smaller baseline surveys specific to each system or intervention given how specific data needs are and how locations and target groups are likely to vary across interventions. In conducting a single baseline study to cover all interventions, you often find that the people assessed initially are not those who are ultimately covered by your interventions.
When are you planning your baseline?
Timeliness is important. Significant delays in capturing baseline information may make your results less accurate. The people you wish to assess may no longer recall their exact status before the intervention began.
If you are unable to establish a reliable baseline from the outset, you can still measure results accurately later. Sample sizes can be made larger and even more care should be taken to identify and isolate any factors that might differentiate targeted and control groups.
Use good measurement practices
You should adhere to accepted good measurement practices. The following key questions are particularly relevant to programmes employing the market systems development approach:
- Which measurement tasks is it safe to outsource?
- How reliable are secondary sources?
- What exactly do you need to measure?