1. Tools for M&E
Choice of tools for M&E is partly dependent on the nature of specific markets and target groups, but should also
be guided by two considerations:
1.
Consistency between tools used for market assessment and intervention and those used for M&E
For reasons of coherence and efficiency, market assessment and M&E need to be closely linked. Tools
used for market assessment can also be used to provide evidence of overall market change and, in
particular, of the specific dimensions of the market changed by intervention. Tools might have to be
selected and adapted to suit different dimensions of the market system. This is the priority focus for
M&E in market systems – and relatively more rigour and resources are justified here.
2.
Whether information is needed for prove and improve purposes
The issue of attribution is important in market development when trying to ‘prove’ impact, particularly in
terms of poverty reduction. There is no magical fix for this issue. Methodologies for proving final impact
need to be able to determine strong:
Causality. Where the link between an intervention and market level (cause) and consequent higherlevel change (effect) is plausible.
Generalisability. Where it is reasonable to draw wider conclusions on the basis of analysis of an
observed change (i.e. to extrapolate).
Tools for measuring attribution
As can be seen from the figure above there are very few measurement methods that are able to ascertain both
causality and generalisability. Experimental and quasi-experimental approaches are difficult, requiring expertise
and resources. Yet, if programmes are serious about proving final impact, there is no real alternative to these
approaches.
A mix of methodologies is therefore usually required to estimate attribution:
Quasi-experimental methods
Comparative assessments (eg surveys) of affected and control groups should be used when practical and when
the scale of intervention justifies the cost of assessment. Quasi-experimental methods allow for the comparison
of changes between a target group affected by programme intervention and a group that has not been targeted.
Differences in the level of change between these two groups can be used to estimate the degree of change
resulting from programme intervention.
Quasi-experimental methods are not valid when the targeted group is unique and does not have a suitable
comparison group or when interventions can influence the control group as well as the targeted group.
Methodologies using control groups work best for specific interventions and well-defined target groups within a
homogenous population rather than across entire markets or areas.
2. Trend analysis
Programmes should compare change in areas where the programme is active with changes in other areas as
well as national and historical trends. Again, any differences between the changes identified in programme areas
and non-programme areas (or national and historical trends) can be used to estimate the change attributable to
programme intervention.
Qualitative methods
Qualitative methods are useful to investigate change processes (e.g. intermediate changes in terms of behaviour
or practices). Qualitative assessment can help understand the role of programme interventions in contributing to
observed changes and identify other contributing factors unrelated to the programme. For instance:
Field diaries of observations during staff field visits can be used to document what changes are really
happening and why.
In-depth interviews or focus group discussions can be used to capture the opinions of relevant market
players in order to explore why changes have occurred and the factors that contributed to change,
including the role of programme intervention in any change.