I have the following experimental design and would like to know how to apply ANOVA to the data:
The experimental set up is as follows: Pretest, Treatment (or Training) and Posttest In each Pretest and Posttest, the subjects perform the following three tasks:
- $16$ times throwing a ball at a static (non-moving) target
- $16$ times throwing a ball at a target whose position is moved after each pitch horizontally
- $16$ times throwing a ball at a target whose position is moved after each pitch vertically
The order in which these tasks are performed is randomized for the different subjects
During treatment (training), the subjects are split up which either: 1. Trains only on static targets - only after each training block the target is either moved horizontally or vertically 1. Training on targets which during one training block are moved after each pitch either horizontally or vertically
In the training phase, each subjects performs $20$ pitches during $4$ blocks ($80$ pitches in total)
So we have also four groups in the training: Blocked/horizontal, blocked/vertical, Variable/horizontal, Variable/vertical
During pretest and posttest, the following quantities are measured: - For each of the exercises (1)-(3) the error (distance where the ball hit the wall from the target) and the duration of the last visual fixation of the target before a pitch (how long the subject stares at the target before a pitch).
I want to test now the following hypotheses:
- H1: The average error during posttest is bigger for the variable training/treatment group than for the blocked training/treatment group
How can I do that with ANOVA?
Would it be correct to average the errors for (1)-(3) for both pre- and posttest and then perform ANOVA on the four groups pretest/variable, pretest/blocked, posttest/variable and posttest/blocked?