Trade-offs in Detecting Letters and Comprehending Text

Abstract

Critics of the alphabetic character detection task have questioned whether findings from that paradigm throw back normal reading processes. The at hand study addresses these questions using a novel computerized version of the alphabetic character detection task in which reading rate along with alphabetic character detection and comprehension accuracy are examined. Previous alphabetic character detection findings were replicated with this fresh computerized task. Different conditions were compared in which detection and comprehension instructions were manipulated as well as the salience of the target alphabetic characters The requirement to comprehend had small issues on letter detection accuracy and reading rate, and alphabetic character detection only modestly reduced comprehension. Thus, the processs developed in this study permit examination of the constituent processes contributing to performance in the alphabetic character detection task.

A simple task has been used to investigation reading processes in which participants are given a normal unromantic passage to read and asked to indicate each instance of a target alphabetic character in the passage. This task has been useful for revealing perceptual, syntactic, and semantic processe underlying skilled reading (see Greenberg, Healy, Koriat, & Kreiner, 2004; Healy, 1994; Koriat & Greenberg, 1994 for summaries). The validity of these arises depends on the assumption that the dual task methodology underlying the alphabetic character detection task does not distort the normal reading processe individual goal of the present research is to vary the task demands to identify in what way the component processes of the alphabetic character detection task affect reading rate and comprehension.

Investigations using the alphabetic character detection task have found trade-offs between spe and accuracy in detecting targets (eg Healy, 1976) However, timing has been uncooked in the earlier investigations. Hence, a next to the first goal of the present investigation is to develop a computerized version of the alphabetic character detection task that would allow for more precise timing as well as an examination of composing processes.

A study by Rayner and Raney (1996) lay the foundation of that word frequency effects upon eye movements for participants who read short passages for comprehension disappeared when participants instead searched the passages for target words. These findings are consistent with a claim by dint of Rayner and Pollatsek (1989) that the detection task disrupts normal reading processe thus that generalizations to normal reading are "somewhat dubious" (p 450) However, in the close attention by Rayner and Raney, participants were told to scan the body "as fast as they could" (p 246) in the way that their comparison of normal reading and visual scanning confuses two factors: the task demand to read for meaning and the neighborhood of a secondary task to scan for targets. A third goal of the not away study is to apply a componential analysis separating these sum of two units factors.

A computerized version of the alphabetic character detection task was developed by means of Healy, Oliver, and McNamara (1987) (see also Proctor & Healy, 1985; Saint-Aubin & Klein, 2001 2004; Saint-Aubin, Klein, & Roy-Charland, 2003) true copy segments are shown for fixed intervals of time, and participants reply with a key press whenever they lay open a target letter. This task is limited because reading time offers at a fixed rate by means of segment, because it is impossible to know to which specific word participants are responding, and because the amount of information in the parafovea is reduc In sum of two units experiments, Saint-Aubin et al. (2003) addressed the first sum of two units of these limitations in a self-paced variant of this task. However, the third limitation, concerning parafoveal information, remained. Parafoveal processing has been shown allowing to play an important character in reading (e.g., Blanchard, Pollatsek, & Rayner, 1989) and specifically in the alphabetic character detection task (Hadley & Healy, 1991; on the other hand see Saint-Aubin & Klein, 2004)

To subdue all three limitations of the previous computerized version of the alphabetic character detection task, we developed a novel version in which participants view an entire passage at one time on a computer screen and make their detection reply by mouse clicking on words containing the target alphabetic character This version allows participants to vary their reading rate to suited the demands of the task. It eliminates the riddle of response ambiguity, and, because passages are neared fully, parafoveal processing is not disrupted. Further, it allows the experimenter to check for detection regressions (i.e., instances when participants find two targets with the next to the first detection occurring earlier in the passage than the first). similar regressions are not desirable because they violate the instructions to the participants; these detection regressions cannot be revealed in the paper-and-pencil version of the alphabetic character detection task. We compared performance in this alphabetic character detection task under varying instructions and secondary task demands. Unlike earlier studies of alphabetic character detection, the passages and comprehension trials were taken from standardized reading trials Thus, the comprehension measure is likely to be more valid and sensitive than that used previously.