EDUCATIONAL PSYCHOLOGIST, 42(3), 123–137 Copyright C 2007, Lawrence Erlbaum Associates, Inc. ⃝ Cognitive Load and Classroom Teaching: The Double-Edged Sword of Automaticity David F. Feldon Department of Educational Studies University of South Carolina Research in the development of teacher cognition and teaching performance in K–12 class- rooms has identified consistent challenges and patterns of behavior that are congruent with the predictions of dual-process models of cognition. However, cognitive models of information processing are not often used to synthesize these results. This article reviews findings from the research on teaching and teacher education through the lens of a dual-process model and emphasizes the role that cognitive load plays in driving teaching performance. Data reflecting the salience of automaticity and its relationship with cognitive overload are highlighted, and implications for teacher preparation and inservice training strategies are discussed. Specific suggestions for teacher training draw on empirical findings from cognitive approaches to train- ing that emphasize the development of automaticity in teaching skills to minimize extraneous cognitive load and maximize effective performance. Feeling overwhelmed by the amount of simultaneous activity is cognitive load’s modulation of the functional dynamic be- in a classroom is a common experience for novice teachers tween deliberate (i.e., conscious) and automatic (i.e., non- (e.g., Carre, 1993; Corcoran, 1981; Kagan, 1992; Olson & conscious) processing. Following a discussion of the effects Osborne, 1991; Veenman, 1984; Wideen, Mayer-Smith, & of cognitive load on general cognition and behavior, this per- Moon, 1998). The attempt to attend to the needs and behav- spective frames empirical findings from research on teaching iors of an entire classroom while also trying to remember and teacher education as manifestations of the same effects. and implement a lesson plan inundates their available cogni- Drawing on the compatibility of the findings in these dis- tive resources. Therefore, this cognitive overload limits the tinct bodies of literature to advocate this hypothesis, the abilities of novice teachers to adapt effectively to complex article concludes by discussing implications for preparing classroom dynamics (Doyle, 1986). preservice teachers and interpreting the classroom behav- Cognitive overload occurs when the total processing de- ior of teachers by applying training principles that address mands of external stimuli and internal cognitions exceed the development of automaticity for complex skills (Rogers, available attentional resources (Sweller, 1989). When con- Maurer, Salas, & Fisk, 1997; van Merrienboer,¨ Kirschner, & scious mental operations occur, they occupy some portion of Kester, 2003; van Merrienboer,¨ Kester, & Paas, 2006). limited working memory and constrain the attention available for other cognitive activities. Thus, levels of cognitive load THE DUAL-PROCESS MODEL OF COGNITION significantly affect both learning and performance in authen- tic contexts (e.g., Goldinger, Kleider, Azuma, & Beike, 2003; During the last century, both behaviorists and cognitive sci- Sweller, 1988; Sweller, van Merrienboer,¨ & Paas, 1998). entists have provided evidence that many mental and be- This article analyzes a broad cross-section of empirical havioral processes take place without conscious deliberation findings from studies of teaching through the lens of the (Wegner, 2002). The dual-process model of cognition indi- dual-process model of cognition to demonstrate the utility of cates that controlled and automatic processes operate inde- acognitivetheoreticalframeworkforunderstandingteacher pendently but intersect at certain points to produce human training and performance. A primary aspect of this model performance (Bargh & Chartrand, 1999; Barrett, Tugade, & Engle, 2004; Devine & Monteith, 1999; Schneider & Chein, Correspondence should be addressed to David Feldon, Department of 2003; Sun, Slusarz, & Terry, 2005). The mental events avail- Educational Studies, University of South Carolina, 1801 Shadowood Dr., able for conscious manipulation and control occur in working Columbia, SC 29212. E-mail: [email protected] memory, function more slowly, and require more effort than 124 FELDON other processes. Because the processing capacity of working reason is that some teaching situations call for immediate memory is severely limited (Cowan, 2000; Miller, 1956), ex- rather than reflective responses that probably preclude ratio- cessive cognitive load can prevent fully conscious, deliberate nal processing of information in making an informed judg- reasoning by forcing some goals to be either neglected or ment or decision. The second reason is that the capacity of the pursued predominantly through nonconscious mechanisms human mind for formulating and solving complex problems (Bargh, 2000). such as those presented in teaching is very small compared to the enormity of some “ideal” model of rationality (i.e., some In short, dual-process theories of cognition assert that in- normative model). In order to handle this complexity, a per- formation processing occurs simultaneously on parallel path- son constructs a simplified model of the real situation. The ways. The controlled pathway generates conscious, slow, and teacher, then, behaves rationally with respect to the simpli- effortful processing of perceptual and semantic information fied model of reality constructed. This conception of teachers that tends to represent accurately instance-specific informa- with “bounded rationality,” that is, rational within the con- tion. In contrast, the automatic pathway generates fast, effort- straints of their information processing capabilities, leads to less, nonconscious processing through pattern recognition- a modification of the first assumption: Teachers behave rea- based processes that rely on heuristics and generalized and sonably in making judgments and decisions in an uncertain, stereotypic schematic representations (Schneider & Shiffrin, complex environment. 1977; Sloman, 2002). When the two pathways generate con- flicting outcomes (e.g., the Stroop effect; Stroop, 1935), per- The second assumption is that a teacher’s behavior is guided formance slows while the conflict is consciously mediated by his thoughts, judgments, and decisions. If this is not true, “then teachers are automata of some kind” (Fenstermacher, in working memory (Botvinick, Cohen, & Carter, 2004). 1980, p. 36). Hence the question of the relationship between However, if the demands on conscious processing exceed thought and action in teaching becomes crucial yet problem- working memory capacity, cognition generated by the non- atic. Fortunately or unfortunately, researchers studying conscious pathway will manifest without the full benefit teachers’ thoughts, judgments, decisions and behavior do not of conscious monitoring or modification (Beilock & Carr, have . an easy out because in order to understand teaching, 2001; Beilock, Wierenga, & Carr, 2002; Brown & Bennett, we must understand how thoughts get carried into actions. 2002; J. W. Payne, Bettman, & Johnson, 1993; Rieskamp & (pp. 456–457) Hoffrage, 1999). To mitigate the limitations of working mem- ory span on controlled reasoning, people frequently rely on This discussion embodies several common assumptions “fast and frugal” reasoning strategies that reduce the quan- about research in this field that have shaped relevant studies tity of simultaneous information necessary to be attended to to date. First, it acknowledges bounded rationality in human through interactions with automatic processes (Gigerenzer, thought, but it asserts that teachers behave reasonably in re- Czerlinski, & Martignon, 2002, p. 559). Such adaptations lation to the beliefs they hold and the models of classroom may either help or hinder performance. They may help be- situations they construct. In other words, their intended ac- cause they permit other, task-relevant cognitive processing to tions are rational in relation to their consciously available occur. However, they may also hinder performance by lim- representations of teaching situations. However, their result- iting conscious monitoring that might otherwise detect and ing behaviors may not appear rational if their beliefs and correct performance errors. mental models are not accurate in their current situation. Second, Shavelson and Stern’s (1981) commentary implies Historical Assumptions of Bounded Rationality that theories which posit nonconscious decision making by in Teaching and the Dual-Process Model teachers as an explanatory mechanism are unacceptable, be- cause they position teachers as “automata” (Fenstermacher, Research on teaching and teacher education has predomi- 1980, p. 36). Further, they asserted that it is necessary to ac- nantly grappled with the implications of humans’ limited count for how (consciously-held) thoughts “get carried into capacity for consciously mediated cognition by restricting its actions” (Shavelson & Stern, 1981, p. 457). focus to the bounded rationality (Simon, 1957) of teach- Over the past 25 years, these two assumptions have func- ers’ beliefs, intentions, and reflections (O. Lee & Porter, tioned to restrict cognitively oriented inquiry in research on 1990; Rhine, 1998). The logic underlying this approach teaching to methods that rely on the verbal self-reports of emerged from the early research on teachers’ cognition (e.g., teachers’ decision-making processes. Shavelson and