ABSTRACT
In the past decade the question of unit of analysis or grain size has forced a rethinking of methodological and analytical techniques used to conduct research in educational psychology. Although the term unit of analysis is used in many research traditions to describe the type of thing being observed and analyzed (e.g., Barab, Evans, & Baek, 2004; Chi, 1997), it also has a specific technical meaning in research designs using statistical inference, where it describes the smallest unit that can be treated as statistically independent. To avoid the constraints of the more specific meaning, but retain the sense of increased resolution or observational power yielded by smaller units, we use the term grain size. Grain size has emerged as a concern in the measurement of self-regulated learning (SRi), metacognition, and motivation, where it has been identified by writers on methodological and assessment issues as a central theme in contemporary research (pintrich, Wolters, & Baxter, 2000; Schraw, 2000; Winne, JamiesonNoel, & Muis, 2001; Winne & Perry, 2000). There seems to be consensus that shifts in grain size are important because they require shifts in the methodological and analytical tools used, and "grain size affects the degree to which measurements provide a useful test of theory or interventions" (Schraw, 2000, p. 300). However, discussions about grain size often deal with quite different issues. Some suggest moving from studying aptitudes to studying situated learning events. Others stress the importance of focusing the research lens to sample multiple events occurring over time. Finally, there is discussion of shifting from the individual as the unit of analysis to sociocultural interactions. Each of these grain size shifts is discussed next.
