Skip to main content
. 2021 Feb 8;30(6):1693–1704. doi: 10.1007/s11136-021-02778-5

Fig. 2.

Fig. 2

Conditional effects, simple intercepts and slopes. Estimated regression coefficients for three item-level predictors (number of syllables, imageability, response format), conditional on the moderator variable (cognitive function), are shown in the left-hand column. Increased cognitive function was associated with decreased (toward zero) effects for number of syllables and response format, whereas the effect of imageability was increased (away from zero). Associated simple intercepts and slopes—regression lines between each predictor and response time (RT) and plotted at different values of cognitive function (mean and ± 2 standard deviations)—are shown in the right-hand column. These plots show how response times increase with increasing numbers of syllables and when using the difficulty vs. frequency response format, and how these increases are augmented at higher levels of cognitive function. Conversely, for imageability, the positive effect only emerges at higher levels of cognitive function, though the effect is confounded with number of syllables (see Fig. 3)