Steps

1.  Examine, one at a time, the relations between:

1. Pathways and domains of change
2. Domains of change
3. Domains of change to impact goal

Guided by these questions:

• What are some of the major assumptions you are making here?
• How are they supported or not by empirical evidence in your programming or other literature?
• Is the evidence substantive enough for you to continue to make this assumption without presenting a risk to your program?  If not, then continue with the process.

2.  List the hypotheses from step 1 that are critical to the success of your program and which ought to be tested.

3. Break each hypothesis down further to reveal any sequence of cause-effect, making sure you are not simply drawing on your own logical thought process but on your knowledge of what has been attempted and/or observed in programs – this can apply to CARE’s own or others.

Example from Mali

Supposition 2: The capacity of collective action by women increases their social and political status

Hypothesis: If a critical mass of networks are strong, while women are mobilizing deliberately in favor of their rights, if women mobilize deliberately for their rights, then they will influence the socio-cultural barriers and standards, if they influence barriers and socio cultural norms, then they have more [space] in the spheres of decision making and local community.

4. Do another review of your hypotheses using this set of questions:

1. Are you certain that the hypothesis in all its constituent parts has not been tested before?  List sources of information that would furnish evidence. If you are not sure it has been tested, you may need to first start with a literature review and come back to this hypothesis later with greater precision of what needs to be tested.  When you consult other sources, make sure to include sources from within your country context, as what is true in one locality may not be true for another.
2. When you are sure the hypothesis has not been tested before, consider again whether the hypothesis has strategic importance to your theory of change.  Imagine the effect it might have on your program if the hypothesis is disconfirmed when tested, e.g., if “collective capacity of women did not increase their social and political status.”
3. In the scenario that evidence does exist to confirm your hypothesis, you may actually be more concerned with the magnitude and pace of change. In this case, you may want to simply monitor progress as you would in tracking change for “within-pathway” hypotheses. To wit, it may not require a piece of research to test your hypothesis.

5.  Once decided that a set of hypotheses are indeed “strategic” and need to be tested, do another final review to make them ready for testing:

1. Be specific in your hypothesis, as it applies to your empirical context – what institutions, what issues, which policies, etc.  Specificity is necessary to be able to measure your hypothesis.
2. Identify for whom within your impact group this is a relevant hypothesis, i.e., to whom does it apply? (this is also part of being specific)
3. Does it need to be more context-specific – a population belonging to a particular geographic location?
4. Does your entire team agree on the formulation of the hypothesis? Do you now have a sound rationale for testing this?

6. Drop any hypotheses from your list that do not meet the review criteria above. With your final list of hypotheses that ought to be test, you will now prioritize these according to these guiding questions:

1. Which are urgent for the program, having relevance to the next five years?
2. What are opportunities to conduct the research (in a project evaluation or baseline; or a special study funded out of multiple initiatives)?
3. What are then potential sources of information for this research (e.g., your own initiatives)?
4. Do you have the funds, skills, and resources to do this OR is another organization better-placed?  Will you pay them for conducting the study?

7. Once you conclude which hypotheses are feasible and meet the criteria, then you need to develop a funding / resourcing strategy and a timeline for the research. It is also wise to try and interest other organizations in co-funding and benefiting from the endeavor. Decide whether you need additional expertise to develop the research framework.

Below is an excerpt from CARE Tanzania’s workshop to select hypotheses.  Out of a list of 5 candidate strategy hypotheses, the team prioritized this one for testing:

Hypothesis 1

WE and Economic Development: Women and adolescent girls in VSLA groups build the self-esteem, social and economic capital that enables/shifts in gender roles in HH and community, which makes them to raise their voices and participate in the community decision-making.

Discussion

This is an active assumption in much of CARE’s work – but has it been proven?

Seems to qualify as a research hypothesis.

In this context, we’re looking at the generation of cohesion that leads to women’s ability to make decisions in community and HH. (e.g. “which gender roles” are we looking at? Division of labour, control over assets?) Will need to unpack this further – what are the critical roles/norms that we’re shifting?  Need to narrow it down to items that are comparable and compatible. We may actually be talking about “gender relations” than “gender roles” since we’re talking about the relationships within the HH/community.

In Burundi, VSLA groups are an entry point for other work – we start with the VSL groups to mobilize women/bring them together

Another dimension: VSL groups are not uniformly focused – e.g. a focus on money, a focus on building solidarity, a focus on social issues - does the difference in focus lead to difference in outcomes?  EG in Bdi, they call the group a Solidarity Group.

This might be a reason to focus on “decision-making” because the decision may be context-driven, but it’s the change in women’s decision-making power/role that we’re interested in.

The hypothesis asks: does participation of women & girls in VSL build social capital and self-esteem?  Does participation enable them to play a different role in HH, community, society?  In some of the VSL groups, there’s been increase in GbV – so, if participation in VSL builds self-esteem, social capital, economic capital, we need to understand why.  If there are negative effects (e.g. GbV), we also need to understand why.

There might be a need for us to look out for other organizations collecting information on similar hypotheses. There’s an opportunity to join with others.

If we’re able to collect useful data and analyze it, one element we want to look at is: so what??  How does this help build impact at scale?  How widely is that information shared – how much influence will it have on our work and the work of others.

After another workshop, the team formulated the hypothesis, as follows, and then proceeded to test it as part of a project evaluation:

Women and girls in VSL groups build the self-confidence, social and economic capital that enables changes in gender relations in household and community, which makes them to voice their concerns and opinions and participate in the community decision making.