Design:
Participants are randomly assigned to one of two groups. Treatement 1 or Treatment 2.
All participants are Tested on Matrix Reasoning and Analogies prior to Treatment. All participants are re-tested following treatment.
Due to limited participant numbers I have no control group.
Is an ancova still a suitable option is I do not have a control group, but instead have two treatment group (each designed to improve either Matrix Reasoning or Analogies)
For your study, I would conduct a 2 (between: treatment 1 vs. treatment 2) by 2 (within: pre to post) repeated-measure ANOVA to predict each dependent variable (Matrix, Analogies). A significant interaction would indicate that the experience of the treatment causes a different change. They could be ANCOVA if you have covariates. Your limitation is primarily in your interpretation because you don’t know how much, if any, change from pre to post is simply due to repeatedly measuring Matrix or Analogies. For example, if you only have a main effect of time, but no interaction, you might hope to conclude that both your treatments worked equally well, but maybe neither really worked and it’s nothing more than a practice effect. In this case, you might look at prior studies of test-retest reliability. Best wishes with your project, Bo Co. ~ Kevin
You may use pre and post test design. Depending on sample size and normality, you may then use parametric or non-parametric techniques for the test of mean difference to justify the effect of the treatment.
- Badges
- Science topic
- Similar topics
- Philosophy
- Philosophy Of Science
- Reasoning