[Editor’s note: This post is part of a series devoted to tools and frameworks for researchers to plan better projects right from the start. Read the first blog for an introduction to conceptual frameworks]
What is the formal hypothesis framework?
The formal hypothesis model is probably the best known of these models, since it is closely link with the scientific method. In this framework there is a clear hypothesis that will be proven. The usual model of stating such a hypothesis is to establish a relationship between an independent variable that has an impact on a dependent variable. What comes to my mind, and probably yours, are impact evaluations. These evaluations are meant to evaluate to what extent a given intervention achieved its goal, independently from other factors.
When should you use it?
A proof is necessary. Impact evaluations have become the gold standard of social research, but they are not always needed. Given the complexity and the extensive resources they need, one should probably use this method only when a proof is really needed. Do we always need a proof? Furthermore, one should probably explore the extent to which proving that a program works or does not work will have an opportunity to change a given program.
When good data is available. Evaluations and other formal hypothesis require particularly good data, and results are very sensitive to change in numbers.
When a good design is possible. The evaluations require a good design from the start. Some programs allow for this
Knowledge about the impact of any given program is valuable information. Evaluations and other type of research with a formal hypothesis are long, complex and many times expensive. Researchers would benefit considering these aspects before initiating such a project.
Also, it is worth asking to what extent successful policies are actually those that are proven to be successful, or those that without such guarantee, take a leap of faith. Sometimes I tend to think that it is more the latter than the first.