I'm going to make an experimental research to test an specific phenomenon (a cognitive bias, to be more specific). Then I'm assigning 6 questions to the participants.
Let's call E1 and E2 the two effects I'm testing for, and NE the absence of the effect. And let's call C1 and C2 the two possible conditions in which the tests can be made. Thus, I've created 6 questions in a form combining E1, E2 and NE with both C1 and C2. The presence of E1, E2 and NE are randomized along the 6 questions, where C1 and C2 are fixed in their positions.
As the questions are of similar kind (or they wouldn't be comparable), should I care about the possible interference of maturation of the participant from one question to another? If so, how do I control for this?
To be more specific about the maturation, I mean: after answering, i.e., 2 questions, the participant might be more thoughtful about the next 4 questions and figure out better the problem he/she is facing.
Does my concerns make any sense?
As no one answered my question and I have already found a solution more than a year ago, I'll answer to share it.
First, maturation in cases like this can be mitigated by randomized or pseudorandomized order of the questions. It's the first possible strategy for that. If you offer all participants the order Q1, Q2, Q3, the measurement of Q3 might be distorted by maturation. But if instead you offer one participant this order, and another one the order Q1, Q3, Q2, and another Q3, Q2, Q1, and another Q3, Q1, Q2; Q2,Q1,Q3; and Q2, Q3, Q1, having an "equilibrium" (ensuring all orders are equally available with pseudorandomization) or simply being truly random (reducing effects to chance), you mitigate the effect.
Yet you can also measure the effect of maturation by collecting information about the answering order. Doing so you won't just really prevent the extraneous variable, but also measure it.
- Badges
- Science topic
- Similar topics
- Psychology
- Cognitive Science