Desktop version

Home arrow Philosophy

  • Increase font
  • Decrease font

<<   CONTENTS   >>

Learning strategies fuse domain knowledge and cognition

The chapters covering learning strategies in disciplinary domains are explicit that disciplinary knowledge is a key component of learning strategies. Without knowledge such as schemas about key features of an argument, signals about the trustworthiness of online information, allowable operations for manipulating fractions, and so forth, learners lack critical Ifs and Thens they need to metacognitively monitor content or targets to aim for in applying cognitive operations that generate knowledge. For example, tasks like self-explaining or locating a main idea are possible because a learner knows what an explanation is and what constitutes a main idea. Put simply, tactics comprising learning strategies are fusions of cognitive operations and information to which operations are applied.

This view of learning strategies and their component If-Then tactics makes explicit that learners need disciplinary knowledge as one key to acquiring and applying learning strategies. Taboada Barber et al. (this volume) note working memory plays a role like a valve governing students’ opportunity to use learning strategies. While this might appear to emphasize a view of strategies as pure cognition, it actually makes the point that strategies fuse knowledge with cognition. Working memory is where operations are performed on information. Information learners need is curtailed in proportion to how heavily working memory is taxed. Without critical information, cognitive operations are starved for the raw material a strategy needs. This analysis has a direct and important implication, also emphasized in multiple chapters of this section. Students need to be well taught in the subjects they study because disciplinary knowledge is fuel for learning strategies.

Concerns about experiments on learning strategies

Features of experimental designs strongly affect the validly of inferences about whether and how learning strategies affect achievement. The “gold standard” design - the randomized controlled trial - randomly assigns to a treatment group a proportion of a sample of learners randomly drawn from a clearly and explicitly defined population. These learners receive instruction about a strategy before engaging in a session where content can be learned using the just-trained strategy. One or more other partitions of the sample form control or comparison groups (randomly assigned if there are more than two groups). These learners may experience a form of placebo “training” or proceed directly to the learning session where they learn using whatever methods they bring with them. The learning session all groups experience may be designed with bias if it includes cues or affordances to use the learning strategy being investigated. This would give an advantage to the treatment group trained to use that strategy for this experiment but greatly limit generalizability beyond the experimental setting. After the learning session ends, all participants take a measure of achievement. Analyses of achievement data examine a hypothesis about the effect of the strategy training.

Interpreting findings from randomized controlled trials is problematic (see Winne, 2017). For example, variables a research reports as “defining” a population commonly have little or no empirical backing as validated moderator variables. Also, the sample is almost never randomly drawn from some population. And, there is rarely a genuine incentive for learners to build knowledge. Beyond these shortcomings, most primary studies reviewed for this section of the Handbook suffer other shortcomings. These could be, but rarely are, reduced in ways that strengthen the validity of inferences about effects of learning strategies.

In studies probing effects of learning strategies, it is helpful to know which students use the strategy being investigated. While it strains the logic of random assignment, inferences about strategy effects could be sharpened by removing achievement data for learners (a) trained to use the strategy but who do not demonstrate skill in using it before the learning session, (b) trained to use the strategy but who do not demonstrate they apply the skill during the learning session, and (c) in the comparison group(s) who already use the strategy or a sufficiently close cousin.

To act on these proposals requires gathering trace data. Trace data are observations of a learners behavior that are operationally defined to provide a strong signal about which particular cognitive operations a learner applies to which particular information. For instance, a strategy to identify main ideas and relate them to everyday life could be traced by having learners highlight content they judge to be a main idea. This traces metacognitive monitoring for main ideas and can be scored for whether the highlighted content actually is a main idea. Stronger inferences about students’ use of the learning strategy can be supported if learners also create a note in which they record a personally generated example of the main idea that was not presented in the source the learner studied, the note traces assembling the learners prior knowledge, or a schema-relevant instance of the main idea with the main idea developed in the source. Trace data also can be used to examine whether trained students’ applications of the learning strategy “wobble” as they engage with content in the learning episode. Such data could be useful in adjusting interpretations of strategy effects as a function of trained participants’ fidelity of strategy implementation.

Another feature that can bolster inferences about the effects of learning strategies is to design the measure of achievement to include items in two categories. Strategysensitive items require learners to have applied the trained strategy to respond correctly. Scores on these items afford testing a hypothesis about whether the strategy is necessary. If trained students and non-trained students can both answer strategy-sensitive items, the strategy is not necessary. Strategy-insensitive items have no theoretical relation to students’ use of the strategy during the learning session. These items afford testing whether the strategy undermines learning in some way or has side effects beyond what current theory forecasts. As best I can judge, primary studies using outcome measures like this were not available to authors of the chapters in this section of the Handbook.

<<   CONTENTS   >>

Related topics