Diagnostic, Prescriptive, Predictive, and Evaluative: Types of Learning Questions for Projects and Activities
Learning from evidence and experience has always been part of the discipline of international development. In recent years, though, USAID Staff and USAID implementing partners have been asked to intentionally develop learning questions as part of their project and activity Monitoring, Evaluation, and Learning (MEL) plans. These learning questions are intended to help prioritize the efforts of USAID and its partners to fill knowledge gaps related to our programming.
But where to begin in developing learning questions for a USAID project or activity? It might help to consider learning questions as categorized into four distinct types:
Diagnostic questions: These are questions that ask about the nature of the development challenge that is the focus of a USAID project of activity. “Why are the incomes of smallholder farmers stagnating?” is an example of a diagnostic question.
Prescriptive questions: These are questions that ask what we should do about a particular development challenge. “What intervention would be most appropriate in this context for increasing grade level 3 child literacy rates?” is and example of a prescriptive question.
Predictive questions: These are questions that ask about the possible outcomes of USAID interventions to address a particular development challenge. “What are the potential consequences of providing community improvement grants to the village councils in this region?” is an example of a predictive question.
Evaluative questions*: These are questions that ask what has happened as a result of USAID interventions to address a particular development challenge. “Are nurses who receive USAID funded training on postnatal care using the newly learned methods?” is an example of an evaluative question.
There are, of course, many possible ways to categorize learning questions for projects and activities; these four broad types certainly do not exhaust all the different possibilities. So, what is the use of these categories? There are a couple of ways these categories may be of help in developing and responding to learning questions.
First, if you are struggling as to where to just begin with developing learning questions, then considering each of the categories in relation to a development challenge can potentially help in generating learning questions. Whether you are in the early design stage or already implementing a USAID activity, ask yourself if there are any high priority gaps in your knowledge concerning:
- What the development challenge is or why it exists (diagnostic)
- The best way to address the development challenge (prescriptive)
- What might happen if you implement a particular intervention to address the development challenge (predictive)
What happened as a result of an intervention to address the development challenge (evaluative)
Second, categorizing a learning question that you develop into one of these four types may help in determining when you should expect to answer the learning question. Each learning question should ultimately lead to some type of research or learning activity that will be undertaken to answer the question. However, some learning questions can be ambiguous about when such research or learning activities should occur. For instance, the question “Does teacher training improve student test scores?” could be prescriptive, predictive, or evaluative depending on how and when it will be answered relative to implementing a teacher training intervention. Considering which of the four types you intend your learning question to be can provide clarity as to when a research or learning activity should be conducted relative to the intervention. For example:
If your learning question is diagnostic, then you will plan to conduct a research or learning activity to answer the question prior to considering the possible interventions to address the development challenge. This may occur at the earliest stages of project design or even during activity implementation as new development challenges are identified.
Similarly, if your learning question is prescriptive, then you will plan to conduct a research or learning activity to answer the question prior to designing the specific intervention to address the development challenge. Prescriptive questions may also be asked after evaluative questions. If the answer to an evaluative question reveals a challenge or opportunity in an ongoing project or activity, then you might want to ask prescriptive question about the best way to address the challenge or take advantage of the opportunity.
If your learning question is predictive, then you will plan to conduct a research or learning activity to answer the question after you have designed the intervention (to at least some degree), but prior to implementing your designed intervention to address the development challenge.
If your learning question is evaluative, then you will plan to conduct a research or learning activity to answer the question that coincides with or take place after you have implemented an intervention to address the development challenge.
Finally, categorizing learning questions that you develop into one of these four types can help in determining how you will answer these learning questions. The crucial first step to answering all learning questions is to ask if someone else might have already answered the question or if there are existing sources of evidence that answer the question. If not, then categorizing your learning question into one of the four types can help in narrowing the choices of possible research and learning activities. For example:
If your learning question is diagnostic, then you may want to conduct a research or learning activity to help understand the challenge in its context. This might include a sector assessment, an organizational capacity assessment, a baseline assessment, a systems analysis, a needs assessment, etc. (For USAID staff, be sure to check out the Compilation of Analyses and Assessments (ADS 201 Additional Help) guidance available on USAID’s ProgramNet).
If your learning question is prescriptive, then you may want to research possible interventions from local experience, past evaluations, and systematic reviews of the intervention across many contexts. You might also brainstorm with local partners and stakeholders and consult with potential beneficiaries.
If your learning question is predictive, then you may want to engage in scenario planning of different possible outcomes, develop a simulation model, conduct a predictive impact assessment or analysis (such as an environmental impact assessment), or conduct a cost-benefit analysis. You might even design prototypes or small scale experiments that can be assessed to help predict potential outcomes when the activity or project is implemented at full scale.
If your learning question is evaluative, then you may want to tailor performance monitoring or an evaluation to help answer the question. Research and learning activities to answer evaluative questions may also include less formal methods, such as after-action reviews; participatory methods, such as collecting beneficiary feedback; and learning events where partners and stakeholders can pause and reflect on the consequences of project and activity interventions.
USAID provides a variety of further resources to help in developing and refining learning questions. You might what to check out the learning questions checklist, the tips for developing good evaluation questions, and the guidance for establishing a learning agenda.
* Evaluative questions are not the same as evaluation questions. Evaluative questions may be answered through a variety of research or learning activities as noted later, while Evaluation questions are evaluative questions that are asked as part of an evaluation. Moreover, evaluations often include prescriptive questions. For example, a question that asks what an evaluation team recommends should be done to address any problems identified in an evaluation is a prescriptive question.