Erratum to: Understanding of and reasoning about object–object relationships in long‐tailed macaques?
- 568 Downloads
Erratum to: Anim Cogn (2013) 16:493–507 DOI 10.1007/s10071-012-0591-x
A re‐inspection of the data of Experiment 2 (“reducing the computational demands—flat board versus cup”) revealed that for two subjects, one trial, each in the control condition A in which the reward was hidden underneath of one of two cups, had been coded erroneously as incorrect choices; the two subjects had indeed selected the baited object in both trials. We re‐calculated the repeated-measures ANOVA, which we had reported as a test of subjects’ performances in the four different conditions. The corrected results demonstrate, as before, that performance differed between conditions (F 3,21 = 6.778, P = 0.002; previously: F 3,21 = 7.143, P = 0.002). With one exception, the post hoc comparisons yielded equivalent results [even though the P values changed slightly (see below)]. The only qualitative change was found in the comparison between condition C (one board and one cup, with the food underneath the board) and the control condition A (two cups). Whereas we had originally reported a significant difference between these conditions (P = 0.035), this contrast now marginally fails to reach significance (P = 0.051).
The corrected P values for the other post hoc comparisons are as follows:
Condition B (two boards) versus Condition A (two cups): P = 0.006 (previously: P = 0.004).
Condition B versus Condition C (one board and one cup, reward underneath the board): P = 0.492 (previously: P = 0.494).
Condition B versus Condition D (one board and one cup, food underneath the cup): P = 0.012 (unchanged).
Condition C versus Condition D: P = 0.082 (previously: P = 0.083).
Condition D versus Condition A: P = 0.724 (previously: P = 0.604).
Furthermore, we had compared the performance within each condition against the hypothetical chance level of 50 %. For the corrected data set of the control condition A, this comparison is now: paired t test: t = −1.214, df = 7, P = 0.264 (previously: t = −1.57, P = 0.16).
Moreover, we had reported that the performance over the course of the six blocks (all conditions combined) did not change. Our new analysis confirms this finding (Pearson correlation: r = −0.266, P = 0.611; previously: r = −0.204, P = 0.698).
Lastly, for each condition, we had compared the performance in the first three blocks with the performance in the last three blocks. We repeated this analysis for the control condition A with unchanged results (paired t test: N = 8, df = 7, t = 0, P > 0.999).
We apologize for this error. In our view, the conclusions drawn from the data are still valid.
Performance of the monkeys in Experiment 2. Capital letters below the x‐axis denote the conditions. The objects used per condition are shown below the x‐axis. The illustration shows the objects after the hiding of the reward. For conditions B–D, the top object is rewarded, and the lower object is not rewarded. In the control condition A, each object was rewarded in 50 % of the trials. Boxplots show median, 25th and 75th quartile, whiskers show 10th and 90th percentiles, and dots represent outliers. The solid horizontal line represents the 50 % chance level, and the vertical dotted line separates control and test conditions. Boxes marked with small letters above the x‐axis differ significantly from each other, based on a repeated-measure ANVOVA with post hoc Holm‐Sidak tests. Asterisk shows significant deviation from chance according to a one‐sample t test