Action Effect Anticipation in Saccades1

Supplementary Material

  1. Results of the percentile analyses for manual R-E compatibility effects

Recently, it has been shown that manual R-E compatibility effects increase with increasing RT level (Kunde, Lozo, & Neumann, 2011). To investigate whether the R-E compatibility effect in manual RTs depended on RT level, we conducted 2 x 5 repeated measures ANOVAs with the within-subject factors R-E compatibility (R-E compatible vs. R-E incompatible) and RT percentile (10% vs. 30% vs. 50% vs. 70% vs. 90%). Greenhouse-Geisser corrections were applied where appropriate. Paired t-tests were subsequently conducted to investigate the R-E compatibility effect separately for each RT percentile.

1.1Experiment 1

Both the main effects of R-E compatibility, F(1,18) = 22.21, p < .001, = .55, and RT percentile, F(4,72) = 52.09, p < .001, = .74, were significant with longer RTs being associated with R-E incompatible trials and higher RT percentiles as compared to R-E compatible trials and lower RT percentiles. Moreover, R-E compatibility and RT percentile significantly interacted, F(4,72) = 129.41, p < .001, = .88. Subsequent paired t-tests conducted separately for each percentile showed that the R-E compatibility effect was significant for all percentiles, but varied between percentiles, 10% - 13ms for the incompatible-compatible difference,t(18) = 3.39, p = .003, d = 0.78; 30% - 16 ms,t(18) = 3.18, p = .005, d = 0.73; 50% - 20 ms,t(18) = 2.84, p = .011, d = 0.65; 70% - 27 ms,t(18) = 3.24, p = .005, d = 0.74; 90% - 48 ms,t(18) = 4.35, p = .001, d = 1.00.

1.2Experiment 2

In Experiment 2,where we observed no overall R-E compatibility effect, the main effects of RT percentile, F(4,76) = 57.55, p < .001, = .75, and the interaction between manual R-E compatibility and RT percentile, F(4,76) = 130.85, p < .001, = .87, were significant. Expectedly, participants RTs increased with RT percentile. Subsequent paired t-tests conducted separately for each percentile to investigate the interaction of R-E compatibility and RT percentile showed that the R-E compatibility effect failed to reach significance for each percentile, but the RT differences between R-E incompatible and R-E compatible trials increased with percentile, 10% - 11 ms,t(19) = 0.89, p = .387, d = 0.20; 30% - 9 ms,t(19) = 0,97, p = .344, d = 0.22; 50% - 10 ms, t(19) = 1.10, p = .285, d = 0.25; 70% - 16 ms,t(19) = 1.43, p = .168, d = 0.32; 90% - 27 ms,t(19) = 1.45, p = .162, d = 0.33. The main effect of R-E compatibility failed to reach significance, F(1,19) = 2.08, p = .166, = .10.

1.3Experiment 3

In Experiment 3,where we also did not observe an overall R-E compatibility effect, the main effect of RT percentile, F(4,80) = 342.26, p < .001, = .95, was significant.RT increased with percentile. Furthermore, R-E compatibility and RT percentile significantly interacted, F(4,80) = 465.75, p < .001, = .96. Subsequent paired t-tests conducted separately for each percentile showed that the R-E compatibility effect failed to reach significance for each percentile, but the differences between R-E incompatible and R-E compatible trials increased with percentile, 10% - -1ms,t(20) = 0.32, p = .749, d = 0.07; 30% - 1 ms,t(20) = 0.31, p = .758, d = 0.07; 50% - 3 ms,t(20) = 1.54, p = .139, d = 0.34; 70% - 5 ms,t(20) = 1.88, p = .074, d = 0.41; 90% - 6 ms,t(20) = 1.56, p = .135, d = 0.34. The main effect of R-E compatibility failed to reach significance, F < 1.

  1. Discussion

The analysis of manual R-E compatibility effects in RTs by RT percentile again showed that, when accounting for RT percentile, the overall R-E compatibility effect only reached significance in Experiment 1, whereas in Experiments 2 and 3 the main effect of R-E compatibility in manual RTs was not significant. However, in parallel with the findings of Kunde et al. (2011), we found evidence for a significant increase of the RT difference between R-E incompatible and R-E compatible trials with increasing RT percentile in all experiments. This may be tentative evidence for the emergence of a small manual R-E compatibility effect in the right (slow) tail of the RT distribution.

  1. References

Kunde, W., Lozo, L., & Neumann, R. (2011). Effect-basedcontroloffacialexpressions: Evidencefromaction–effectcompatibility. Psychonomic Bulletin & Review, 18, 820-826.