Check session
Impact of session on the main DV
A) All valid trials
1. Time to cross the corridor
Without controlling for multiple comparisons, and with the standard alpha level in psychology (5%), the time taken to cross the corridor is higher for session 2 than for session 1. This effect is not statistically significant with the correction for multiple comparisons and is not significant at the 1% alpha level used for the main hypotheses. It should be noted that, since all valid trials are taken into account (including trials with a return inside the corridor), the distribution of the DV (time) has a very long tail in the direction of the highest values. Hence, the performance of the statistical model is not entirely adequate even with the log-transformed DV.
2. Return
Without controlling for multiple comparisons, and with the standard alpha level in psychology (5%), the probability of (at least) one return inside the corridor is higher for session 2 than for session 1. This could explain the result reported above, more returns inside the corridor leading to longer times on average. This effect is not statistically significant with the correction for multiple comparisons and is not significant at the 1% alpha level used for the main hypotheses.
3. Time to leave the starting square
There is no significant difference between the two sessions with regard to the time taken to leave the starting square.
4. Time spent in the apparatus
There is no significant difference between the two sessions in terms of time spent in the apparatus (time taken to leave the starting square + time taken to cross the corridor).
B) Straight trials only
1. Time to cross the corridor
Keeping only the trials with a straight trajectory inside the corridor, there is no significant difference between the two sessions with regard to the time taken to cross the corridor.
2. Time to leave the starting square
Keeping only the trials with a straight trajectory inside the corridor, there is no significant difference between the two sessions with regard to the time taken to leave the starting square.
3. Time spent in the apparatus
Keeping only the trials with a straight trajectory inside the corridor, there is no significant difference between the two sessions with regard to the time spent in the apparatus (time taken to leave the starting square + time taken to cross the corridor).
Main results on session 1 only
1. Impact of waiting time on returns
The impact of waiting time on the probability of (at least one) return inside the corridor remains significant, even if we focus solely on session 1.
2. Time to leave the starting square
If only session 1 is retained, the impact of waiting time on the time taken to leave the starting square remains significant at the alpha 5% level but not at alpha = 1%, in the same way as the analyses carried out on the two sessions.
3. Time to cross the corridor
If only session 1 is retained, the impact of waiting time on the time taken to cross the corridor is not significant, in the same way as the analyses carried out on the two sessions.
4. Time spent in the apparatus
If only session 1 is retained, the impact of waiting time on the time spent in the apparatus (time taken to leave the starting square + time taken to cross the corridor) remains significant at the alpha 5% level but not at alpha = 1%.
Distribution of the IV as a function of session
The distribution of IV is not significantly different between the two sessions.