ISSN 0439-755X
CN 11-1911/B

Acta Psychologica Sinica ›› 2020, Vol. 52 ›› Issue (8): 933-945.doi: 10.3724/SP.J.1041.2020.00933

• Reports of Empirical Studies • Previous Articles     Next Articles

The influence of foveal processing load on parafoveal preview of fast and slow readers during Chinese reading

ZHANG Manman, ZANG Chuanli(), XU Yufeng, BAI Xuejun(), YAN Guoli   

  1. Key Research Base of Humanities and Social Sciences of the Ministry of Education, Academy of Psychology and Behavior,Tianjin Normal University; Faculty of Psychology, Tianjin Normal University; Tianjin Social Science Laboratory of Students’ Mental Development and Learning, Tianjin 300387, China
  • Received:2019-11-21 Published:2020-08-25 Online:2020-06-28
  • Contact: ZANG Chuanli,BAI Xuejun;
  • Supported by:
    National Natural Science Foundation of China(31800920);National Natural Science Foundation of China(31571122);Tianjin Special Support Program for talent development;Outstanding Youth Innovation Team Project(52WZ1702);distinguished professor of the “Yangtze Scholar” award program of the Ministry of Education;Major project of the key research base of humanities and social sciences of the Ministry of Education(18JJD190001)


Parafoveal pre-processing contributes to highly efficient reading for skilled readers. Research has demonstrated that high-skilled or fast readers extract more parafoveal information from a wider parafoveal region more efficiently compared to less-skilled or slow readers. It is argued that individual differences in parafoveal preview are due to high-skilled or fast readers focusing less of their attention on foveal word processing than less-skilled or slow readers. In other words, foveal processing difficulty might modulate an individual’s amount of parafoveal preview (i.e., Foveal Load Hypothesis). However, few studies have provided evidence in support of this claim. Therefore, the present study aimed to explore whether and how foveal lexical processing load modulates parafoveal preview of readers with different reading speeds (a commonly used measurement of reading skill or reading proficiency).
By using a three-minute reading comprehension task, 28 groups of fast and slow readers were selected from 300 participants (234 were valid) according to their reading speed in the current study. Participants were then asked to read sentences while their eye movements were recorded using an Eyelink 1000 eye-tracker. Each experimental sentence contained a pre-target word that varied in lexical frequency to manipulate foveal processing load (low load: high frequency; high load: low frequency), and a target word manipulated for preview (identical or pseudocharacter) within the boundary paradigm.
Global analyses showed that, although fast readers had similar accuracy of reading comprehension to slow readers, they had shorter reading times, longer forward saccades, made fewer fixations and regressions, and had higher reading speeds compared to slow readers, indicating that our selection of fast and slow readers was highly effective. The pre-target word analyses showed that there was a main effect of word frequency on first-pass reading times, indicating an effective manipulation of foveal load. Additionally, there were significant interactions of Reading Group × Word Frequency, and Reading Group × Word Frequency × Parafoveal Preview for first fixation and single fixation durations, showing that the frequency effects were reliable for fast readers rather than for slow readers with pseudocharacter previews, while the frequency effects were similar for the two groups with identical previews. However, the target word analyses did not show any three-way or two-way interactions for the first-pass reading times as well as for skipping probability. To be specific, the first-pass reading times were shorter at the target word with identical previews in relation to pseudocharacter previews (i.e., preview benefit effects); importantly, similar size effects occurred for both fast readers and slow readers.
The findings in the present study suggest that lexical information from the currently fixated word can be extracted and can be used quickly for fast readers, while such information is used later for slow readers. This, however, does not result in more (or less) preview benefit for fast readers in relation to slow readers. In conclusion, foveal lexical processing does not modulate preview benefit for fast and slow readers, and the present results provide no support for the Foveal Load Hypothesis. Our findings of foveal load effects on parafoveal preview for fast and slow readers cannot be readily explained by current computational models (e.g., E-Z Reader model and SWIFT model).

Key words: fast readers, slow readers, parafoveal preview, foveal processing load, Chinese reading