The focus of this book, which is based on Bowles’s doctoral dissertation, is on a specific type of verbal report provided during the completion of a task—namely, think-aloud. Such think-alouds in principle would exclude verbal report involving retrospection, such as through stimulated recalls (Gass & Mackey, Reference Gass and Mackey2000) and self-report (Cohen, Reference Cohen2011). Bowles notes that, at the present time, think-alouds are a popular means for studying first- and second-language learners’ linguistic knowledge beyond the invariably incomplete data associated with learners’ language production.
Bowles offers a generally positive treatment of think-alouds. For example, she is quick to point out that language researchers with differing research perspectives (e.g., conversational analytic, sociocultural, and cognitive) have all embraced verbal report, albeit for different purposes. It is fair to say that the principal aim of the book is to stimulate the responsible use of think-alouds as a research tool.
Chapter 1 provides a brief historical perspective on the use of verbal report in the 1930s and 1940s and then focuses on applications since the 1970s. Bowles notes that think-alouds have been applied to studies in numerous skill areas, such as reading, writing, strategy use, translation, conversational interaction, interlanguage pragmatics, and language testing. Chapter 2 presents the model for verbal report developed by cognitive psychologists Ericsson and Simon and discusses the controversy regarding the reactivity of the approach. Findings from studies comparing participants who verbalized and those who did not—on both nonverbal and verbal tasks—lead Bowles to conclude that verbal reports are largely nonreactive in nature. In Chapter 3 she uses a meta-analysis to provide insight into the conditions under which think-alouds may in fact be reactive. Of note to other researchers, she found that verbal report does increase time on task slightly and that there is no simple answer regarding the reactive effects of the measure. For instance, the effect of the measure was found to depend in part on what respondents need to do (e.g., providing metacognitive feedback is more demanding). A call is made for future research to determine how reactive the measures actually are.
Chapter 4 provides guidelines for collecting verbal report data, such as the recommendations that respondents be told why they are asked to think aloud, be given instructions as to how to do so, and be given a warm-up task and a chance to ask any questions that they might have. Chapter 5 focuses on factors that need to be taken into account when analyzing think-aloud data, such as how to transcribe the think-alouds (e.g., conversation-analytic approaches, which use a detailed transcription system, as opposed to studies in sociocultural theory and cognitivist approaches to SLA, which do not). Other factors include making sure the data are representative and determining how to code the data. The author concludes by commenting on the debate over think-alouds. Her position, based on the findings from her metastudy, is that the one-size-fits-all approach is inappropriate because think-alouds may have different functions under different circumstances. In the same way, she concludes that reactivity is not unidimensional but rather depends on a host of factors and can be either positive or negative with regard to task performance.
A strong feature of the volume is that it focuses on just one aspect of verbal report—namely, think-alouds. Another is that the author herself has conducted think-aloud studies, enabling her to draw on her first-hand research experience with think-alouds to lend credibility to the book. An area of possible concern is that of the line between think-alouds and other forms of verbal report. In reality, when does verbalization “while performing the task” lapse into verbalization “some time after performing the task” (p. 13)? The issue of whether subjects are truly describing their current mental states or actually drawing on their memory of what transpired is often finessed in the research literature. This fuzziness at the boundary of categories can help to explain some of the error variance in verbal report studies. Despite this quibble over terminology, the volume is extremely valuable for its willingness to tackle head-on the issue of the reactivity of think-aloud measures. Bowles does an impressively rigorous job of accomplishing this, which makes the book worthy of careful scrutiny by research practitioners interested in this form of verbal report.
