On the launch-site effect for skipped words during reading

Abstract

The launch-site effect, a systematic variation of within-word landing position as a function of launch-site distance, is among the most important oculomotor phenomena in reading. Here we show that the launch-site effect is strongly modulated in word skipping, a finding which is inconsistent with the view that the launch-site effect is caused by a saccadic-range error. We observe that distributions of landing positions in skipping saccades show an increased leftward shift compared to non-skipping saccades at equal launch-site distances. Using an improved algorithm for the estimation of mislocated fixations, we demonstrate the reliability of our results.

A model of saccadic landing positions in reading under the influence of sensory noise

“…Landing-position distributions are decomposed into cases after skipping saccades (black) and onestep saccades (grey) (from . With respect to the random variability of saccadic landing positions (Figure 2b), Krügel and Engbert (2010) observed that landing positions in skipping saccades are associated with a generally greater variability, confirmed by consistently larger standard deviations of the fitted normal distributions. Moreover, the standard deviations of the distributions of landing positions in word-skipping saccades markedly increased with increasing launch-site distances, whereas the standard deviations of the landing distributions in one-step saccades only slightly increased with longer target-word distances.…”

“…Interestingly, within-word landing positions 336 KRÜGEL AND ENGBERT after skipping saccades differ from landing positions after one-step saccadeseven if they are launched from the same distances to the target words. Krügel and Engbert (2010); demonstrated that word skipping strongly modulates where the eyes land within the target word (see also Drieghe, Pollatsek, Staub, & Rayner, 2008;Krügel, Vitu, & Engbert, 2012;Radach, 1996;Radach & McConkie, 1998). Figure 1 presents experimental data for the distributions of landing positions within words of different length for skipping saccades and one-step saccades, both launched from equal distances.…”

“…In a more detailed analysis, Krügel and Engbert (2010) provided separate plots and regression analyses for average within-word landing positions and the standard deviations of the corresponding landing-position distributions (see Figure 2), with each plot contingent on factors launch-site distance (in letters) and saccade type (one-step saccades vs. skipping saccades). For both measures (average landing position and standard deviation), a main effect of saccade type and an interaction of saccade type and launch-site distance was found.…”

“…Rather these models typically sample fixation positions based on the purely statistical model by McConkie et al (1988). However, it has been pointed out earlier by Krügel and Engbert (2010) that the simplification of modelling fixation positions without considering word skipping will induce a number of inaccuracies and problems. On the one hand, taking saccade type into account will result in substantially different fixation position within the target words, which is important because where readers fixate within words is among the most important predictors of fixation duration (Vitú et al, 2001).…”

Linked linear mixed models: A joint analysis of fixation locations and fixation durations in natural reading

“…Furthermore, they argued that fixation locations are subject to a saccadic range error that results in deviations from the intended fixation locations. Finally, Krügel and Engbert (2010) discovered that the launch-site effect is strongly modulated by word skipping. If the word before the target is skipped, the distribution of fixation locations is shifted leftwards compared to nonskipping saccades.…”