Hello,
I am currently in the middle of a project and cannot figure out what statistical test I should use?
In this hypothetical study, my group and I are looking at the effects of a meditation intervention on anxiety levels, EMG activity and golf putting performance. These 3 variables would be measured at two time-points: baseline and post-intervention. To assess whether the meditation program was effective, a control group would be used that doesn't do the meditation program.
So to sum up, we would have:
2 groups: Intervention and control
2 time-points: Pre- and post-intervention
2 different anxiety conditions: Low and high-anxiety
2 dependent variables: EMG activity and putting performance
I understand that the test needs to combine within-subject elements (due to the timepoints and anxiety conditions) and between-subject elements (putting performance and EMG activity). So would a mixed-mode ANOVA be used?
I would appreciate any help anyone can give me.
Kind regards,
T.S.
Thomas Stirling I would suggest you look upon Two-way Analysis of Variance with different anxiety conditions (low and high) as the blocking variables provided that both groups are selected randomly from a homogenous population. If the sample you selected is biased between groups than you can add the pre-test measures as covariates, making it a two-way ANCOVA.
This would allow you to find the effect of the intervention on the dependent variable with and without the blocking effect.
Although, I am a little confused with your statements above for anxiety levels. Did you randomly assigned subjects with low and high anxiety condition to both groups?
Hi Mr Thapa,
Thank you for your help. No, so basically we need a condition with high anxiety to test whether a particular phenomenon occurs due to anxiety. So, both groups would perform the putting task in a control ("normal anxiety" situation) and perform it again (on a different day) in a created "high anxiety" situation.
Thomas Stirling If that's so, then you are good to go with mixed-design ANOVA (2x2 ANOVA). I suppose you are interested in the interaction effect too (time/ anxiety x group interaction) for your experiment.
You will need to carry out two different analyses for "putting performance" and "EMG activity" separately. The within-subject factor will be anxiety level and the between-subject factor will be the two groups.
But there is a concern with this method, you are not exposing both the groups to both anxiety levels, and hence, I feel you should look back into your research question, and frame your methodology again. If only the effect of two different levels of anxiety is your concern, then better go with paired t-test.
For e.g. when you take time as a within-subject factor, the difference between pre-test and post-test time (magnitude) have to be same, you can't include two different time period for different groups, and in your case, both the group are not exposed to similar magnitude (high anxiety) in post-test.
- Badges
- Science topic
- Similar topics
- Computer Science
- Program