Good evening all! I am beyond appreciative of all your knowledge.
I am trying to establish sample size for an experiment yet lost on a few of the factors inside of G*Power.
Pre-Test = 2 different surveys to test for self-esteem and need to belong additionally a writing prompt (using them as a covariate).
Intervention = Two different methods of receiving an affirmation
Post-Test = Writing prompt, and the two scales (all DV for my study). I am predicting that due to affirmations, the scores will change yet want to see the difference between the two delivery methods.
I was trying to use G*Power and ANCOVA unless you suggest Manova. I am lost on the numerator df and what to place here, it is set for 10. I placed 2 groups instead 5. What might I be doing wrong.
I have a previous post, and of course, things changed.
Thank you in advance for any guidance.
Hello Jason,
For univariate ancova (using g*power option of "ANCOVA: Fixed effects, main effects, and interactions"),
For power of the intervention/treatment:
Numerator df = 1 (2 levels of treatment condition)
Number of groups = 2
Number of covariates = 1 (for a given DV)
And that's that.
For multivariate ancova (if you decide to go that route):
For the power of the intervention/treatment:
Number of groups = 2
Number of response variables = 3
And understand, that: (a) the total sample size given will be too low by an amount equal to the number of covariates; and (b) you'll be estimating power to detect ES values equal to your chosen threshold on residualized scores (after adjustment for covariate/s), not the raw scores.
Good luck with your work.
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Sampling Design
- Sample Size