Identify the primary problem or question that needs answered. According to Jack R. Fraenkel and Norman E. Wallen of Tusculum College, once you have identified the problem, you then need to clarify the problem. This means you need to more narrowly define the problem so that a specific experiment can be conducted to obtain the answer to a problem in education.
List the variables that will be a factor in the outcome of the experiment. For instance, if you are conducting a testing session to determine how female students test versus male students, you will need to list these plus any other factors that may determine the outcome of student performance such as if they test in the morning or the evening, socioeconomic background of the children and even whether the test is computer- or paper-based. All factors need to be listed.
Determine whether the research design you intend to use is adequate to answer the question at hand. According to Horn, this is one of the first steps in putting together a design model that works effectively. You may already have a research design model that you intend to use, like quantitative research, and discover that the data can only tell you so much.
Conduct your initial experiment to determine whether any adjustments or modifications need to be made to the basic research design. For instance, you may decide that median test scores prove a more accurate assessment of student abilities than average scores, which you may have intended to use initially. This is the tweaking phase of the research process.
Repeat the experiment. During this phase, gather as much data as possible to determine whether the research design has been adequate to answer your initial inquiry. By repeating the experiment multiple times you can assess whether or not the initial outcome is repeatable. Once the outcome is repeated, you can then analyze and assess the implications of your research for education.