Abstract
In reading, the metrical properties of saccadic eye movements are traditionally measured in letters. This approach, supported by several empirical findings showing that the mean length of saccades, when measured in letters, does not vary as a function of viewing distance and character size, may however bias our interpretation of basic eye-movement patterns along the lines of text. In the present study, we investigated the contribution of character size to one of these rather robust phenomena, the launch site effect; this reveals that the eyes land further into a word as the distance of the eyes to the beginning of the word decreases. The eye movements of five participants were recorded while they read lines of words in an animal-name search task. Words were presented in two different font sizes (.2° and .4°) and the eccentricity of the first target word (or launch site) was manipulated. As in previous studies, the launch site was defined in number of character spaces (1 to 9) from the beginning of the target word. Results confirmed a launch site effect in both font sizes, but the effect tended to be stronger for small-printed target words; as the launch-site distance became smaller, the eyes tended to land closer to the end of small- compared to large-printed words. As revealed in further analyses, this trend extended beyond the word boundaries as the likelihood of skipping the target word was greater for small- compared to large-printed words in the close launch site conditions. Inconsistent with classical accounts of the launch site effect, the present findings provide new benchmark data for models of eye-movement control in reading. They also suggest that we may need to reconsider the use of letters as a metric to measure eye behaviour in reading.