Table 2.
Effect of sequence structure and temporal parameters on discrimination performance. Effect is measured as estimated change in the percentage of correct responses. Significance is assessed by analysis of deviance of linear mixed models (χ2 statistic with one degree of freedom). The first two lines refer to a linear mixed model with % correct responses as the dependent variable, two-fixed effects variables (identity of the last stimulus in the to-be-discriminated sequences, and identity of the last two stimuli) and experiment as a random effect to control for differences in experimental tasks. The analysis is restricted to studies employing sequences of two or three stimuli. The following lines estimate the effect of temporal parameters. These effects refer to a 1 s increase in the independent variable. For example, an increase in stimulus duration of 1 s is estimated as improving correct responses by 2%. All models are linear mixed models restricted to studies in which the variable of interest was varied, with experiment as a random effect.
| effect | χ2(1) | p< | |
|---|---|---|---|
| sequence content: | |||
| same last stimulus | −8% | 43.4 | 10−9 |
| same last two stimuli | −18% | 83.7 | 10−9 |
| duration of: | |||
| stimuli | 2% | 26.0 | 10−6 |
| interval between trials | 0.2% | 7.7 | 0.01 |
| gap between stimuli | −1% | 162.4 | 10−9 |
| gap between end of sequence and response opportunity | −2% | 14.9 | 0.001 |