Using Eye Movements to Study Reading Processes: Methodological Considerations

Methodological Issues

Since the mid-1960’s, researchers from education and psychology interested in reading have focused not only on the outcome of reading comprehension, but also on the processes that must take place in order for comprehension to occur. These two aspects of reading comprehension – product and process – are naturally related but tend to be studied with different measures and focus on different aspects of reading comprehension theory (van den Broek & Espin, 2012). Research on reading comprehension outcomes, or the product of reading, largely uses off-line measures that are designed to capture the mental representations encoded in long-term memory after reading has occurred. These may include tests that are designed to assess readers’ overall understanding of the information in a text, or tests that specifically assess recall of text information (e.g., Hua & Keenan, 2014; Carlson, van den Broek, McMaster, Rapp, Boh-Gettler, Kendeou, & White, 2014). Studies that focus on reading processes, however, utilize online measures that record differences in processing speed over the course of a text. Information from these measures can then be mapped onto different sub-processes within reading from perception to word-level processing to syntactic analysis to discourse comprehension. These online measures have ranged from relatively general measures of processing time (e.g., reading rate; e.g., Bolanos, Cole, Ward, Tindal, Hasbrouk, & Schwanenflugel, 2013; Hiebert, Samuels, & Rasinski, 2012; Veenendaal, Groen, & Verhoeven, 2015), to more specific measures that allow for a fine-grained view of what occurs during reading (e.g., word or sentence reading time; Cook, Myers, & O’Brien, 2005; Mitchell, 2004; Taraban & McClelland, 1998). Eye tracking based measures of reading behavior, which will be the focus of our chapter, provide the most detailed view of moment-by-moment processing, mapping readers’ fixations and their durations and locations onto specific words and phrases within a text to allow researchers to draw inferences about the cognitive processes that occur during reading.

According to Rayner (1978; see also Rayner, Pollatsek, Ashby & Clifton, 2012), eye movement research in reading has experienced three eras. The first was from 1879 to 1920, when researchers investigated basic characteristics of eye movements. Early work was conducted by Huey (1908), who provided the foundation for much of what is known today about eye movements during reading. These early studies used tachistoscopes (affectionately termed t-scopes) to control the amount and duration of exposure of a stimulus to a participant. Through clever experimentation, Huey and others (e.g., Cattell, 1886; Erdmann & Dodge, 1898) revealed critical information about readers’ perceptual span and basic word recognition processes (for a review, see Rayner et al., 2012). The second era of eye movement research fell between 1920 and 1960, during the behaviorist revolution in psychology, and focused more on observable applications of eye movements, with less attention toward understanding the underlying processes involved in reading. Instead, research was oriented more toward use of eye movements as measures of reading ability (e.g., Anderson & Swanson, 1937; Luckiesh & Moss, 1941; Tinker, 1936).