Does text contrast mediate binocular advantages in reading?

Abstract

Humans typically make use of both of their eyes in reading and efficient processes of binocular vision provide a stable, single percept of the text. Binocular reading also comes with an advantage: reading speed is high and word frequency effects (i.e., faster lexical processing of words that are more often encountered in a language) emerge during fixations, which is not the case for monocular reading (Jainta, Blythe, & Liversedge, 2014). A potential contributor to this benefit is the reduced contrast in monocular reading: reduced text contrasts in binocular reading are known to slow down reading and word identification (Reingold & Rayner, 2006). To investigate whether contrast reduction mediates the binocular advantage, we first replicated increased reading time and nullified frequency effects for monocular reading (Experiment 1). Next, we reduced the contrast for binocular but whole sentences to 70% (Weber-contrast); this reading condition resembled monocular reading, but found no effect on reading speed and word identification (Experiment 2). A reasonable conclusion, therefore, was that a reduction in contrast is not the (primary) factor that mediates less efficient lexical processing under monocular reading. In a third experiment (Experiment 3) we reduced the sentence contrast to 40% and the pattern of results showed that, globally, reading was slowed down but clear word frequency effects were present in the data. Thus, word identification processes during reading (i.e., the word frequency effect) were qualitatively different in monocular reading compared with effects observed when text was read with substantially reduced contrast. (PsycINFO Database Record

Reading Single Words Aloud With Monocular Presentation: The Effect of Word Frequency

“…They found that when English sentences were read with one eye, the processing benefit for high frequency words was eliminated in both gaze duration and fixation times (Jainta et al, 2014, jointly manipulated the preview reading conditions and the target reading conditions-preview benefits are not relevant to our study, so we focus only on the viewing conditions when reading the target words). Jainta et al (2017) observed the same elimination of the frequency effect under monocular reading conditions for gaze duration and first fixation times in German. The total reading time measure continued to show a word frequency effect (one-tailed) under monocular conditions, but it was less than half the magnitude of the effect under binocular conditions (see Figure 1).…”

“…Indeed, we are unable to conceive how such models could even show a reduction of the word frequency effect due only to slowing of high frequency items. In short, the results reported by Jainta et al (2014Jainta et al ( , 2017 should they generalize to single word reading experiments, would strongly challenge current thinking about one of the most widely accepted benchmarks used to evaluate accounts of single word reading aloud.…”

“…A word frequency effect is observed in single word reading aloud, lexical decision, sentence reading, same-different matching, and using many measures: reaction time, errors, gaze duration, fixation times, detection thresholds and so on (e.g., see Howes and Solomon, 1951;Broadbent, 1967;Morton, 1968;Rayner and Duffy, 1986; among many others). A notable exception are the results seen in Jainta et al (2014Jainta et al ( , 2017. They report, in two eye-tracking studies, that the word frequency effect is eliminated (or reduced) when participants read sentences with only one eye (monocular reading).…”

“…Jainta et al (2014Jainta et al ( , 2017 report an exception in that, when reading sentences with only one eye, the word frequency advantage disappeared. If this same pattern were seen in single word reading it would strongly challenge all current theoretical accounts of reading aloud currently on the table.…”

“…Due to how the participants arrived in the lab, we had collected 22 subjects by the time the first analysis point occurred, at which point the stopping criteria were met and data collection was terminated. We also note that rather than using the word frequency effect in the monocular condition as our stopping rule, we used the presence or absence of the word frequency by reading condition interaction since this would clearly support the Jainta et al (2017) account even without a complete elimination of the word frequency effect. In the end, both stopping rules were achieved at the same point so there would have been no difference had we remained with the pre-registered rule.…”