Assessing the Structure of Verbal Protocols Stacey A. Todaro (shaberk1@niu.edu) Joseph P. Magliano (jmagliano@niu.edu) Keith K. Millis (kmillis@niu.edu) Department of Psychology, Northern Illinois University DeKalb, IL 60115 USA Danielle S. McNamara (d.mcnamara@mail.psyc.memphis.edu) Department of Psychology, University of Memphis Memphis, TN 38152 USA Christopher A. Kurby (ckurby@artsci.wustl.edu) Department of Psychology, Washington University St. Louis, MO 63130 USA Olson, Duffy, & Mack, 1984; Pressley & Afflerbach, 1995; Trabasso & Magliano, 1996; Wolf & Goldman, 2005). Procedurally, this requires readers to either verbally report, or type whatever thoughts come to mind after reading sentence (s) of a text. Thus, one commonly held assumption regarding verbal protocols is they reveal what information from the prior text of the passage and the reader’s world knowledge is consciously available in working memory and ‘codeable’ in language (Ericsson & Simon, 1993). Therefore, if a reader is making a causal connection between the current sentence and a prior one, evidence of it could be found in his or her verbal protocol. Because verbal protocols do not reflect a veridical description of the processes per se, researchers must develop qualitative coding schemes that can be used to classify the inferential and strategic processes presumed to be present. More frequently than not, this requires parsing a reader’s utterance into clauses and then classifying each clause as belonging to one of the categories specified by a particular coding scheme without considering the overall structure of the verbal protocol. By structure, we mean the preceding and proceeding clauses. Proponents of verbal protocol methodologies (Ericsson & Simon, 1993; Pressley & Afflerbach, 1995) contend that the structure of verbal protocols per se may be systematic and meaningful of the covert mental processes involved in the phenomena of interest. Unfortunately, in the context of reading comprehension research, the structure of verbal protocols is rarely considered. Consequently, much less is known is about the structure of verbal protocols and how that structure might be reflective of the conscious processes involved in comprehension. Thus, the goal of this study was to examine the structure of verbal protocols and in particular to determine whether there was an identifiable sequence to the inferential and strategic processes readers engage in when constructing meaning from text. Abstract One approach that has been used to study processes that give rise to comprehension is to have readers verbally report their understanding during reading. The thoughts that readers produce usually contain multiple clauses that convey different kinds of processing strategies, such as paraphrasing, bridging, or elaborations. The goal of this study was to examine the structure of verbal protocols and to identify whether there was a sequence to the information sources that comprised a given verbal report. Participants produced think-aloud protocols while reading scientific texts. With respect to whether there was an identifiable sequence to the information contained in a verbal protocol, the results indicated that participants tended to begin their protocols with information from the current sentence before adding information from world knowledge and the prior text. As additional clauses were produced, there was a shift in focus from the current sentence to elaborative information. Introduction It has been well established that to comprehend text, readers must go well beyond the information explicit in it (e.g., Graesser, Millis, & Zwaan, 1997). This requires readers to actively participate in the comprehension process by engaging in inferential and strategic processes that serve to connect text events to one another and to the reader’s general knowledge about the world (Graesser & Kruez, 1993). These types of processes presumably aid in the construction of coherent mental representation of what the text is about, a fundamental aspect of most theories of text comprehension (Gernsbacher, 1990; Graesser, Singer, & Trabasso, 1994; Kintsch, 1988; van den Broek, Young, Tzeng, & Linderholm, 1999; van Dijk & Kintsch, 1983). One methodological approach that has been used to study inferential and strategic processes that give rise to comprehension is to have readers verbally report (i.e., think- aloud, self-explain) their understanding as they proceed through text (Chi, Bassok, Lewis, Reimann, & Glaser, 1989; Cote & Goldman, 1999; Ericsson & Simon, 1993; Magliano & Millis, 2003; Magliano, Trabasso, & Graesser, 1999; Information Sources and Verbal Protocols It has been well established that in order for inferential and strategic processes to occur during reading, readers must be