Da5Ax

New Member
#1
Hello everyone,

I am trying to see how the motor time of a response to a stimulus (DV: quantitative, continuous) evolves based on difficulty level of the task (IV: 8 levels, categorical).
My hypothesis is that motor time will increase, then decrease (inverted U shape).

I conducted a RM ANOVA (using the Greenhouse-Geisser correction, as the scphericity assumption was violated) and polynomial contrasts using JASP, but the RM ANOVA is significant without the GG correction, but non-significant with it. However, the T-test for the quadratic contrast is significant.

Questions:
  1. I've read in papers that the overall ANOVA can be insignificant and the contrast significant, but why ? What is the relation between the contrast T-test and the ANOVA F-test ?
  2. What would the null hypothesis of the contrast in this experiment's case be ? That the 8 means are equal ?
  3. Does the quadratic contrast also test for inverted U shapes ? Or only U shapes ?
  4. Should I report the results of the ANOVA ? Or just the contrast ?
Thank you !