Efficiency: Time To Learn, Cognitive Capacity, and Task Demands
We combine multiple topics here because attempting to isolate factors that interact avoids the difficult challenge of confronting the complexity of human behavior. The background for this discussion might be found in our comments on the limits of working memory and our description of the controversy between advocates of direct instruction and more experiential learning activities.
Learning by definition places a "load" on the cognitive system. Think of load as work. If the cognitive system is not engaged in work, learning does not happen. It takes work to explore the connections between new information and what you already know. It is easier, but less effective to memorize because rote learning makes less use of existing knowledge. It is also important to recognize that the work needed to accomplish meaningful learning is not a constant across individuals. Less work is required when a learner already has a good background relevant to what she is trying to learn.
There are practical limits on the load that the cognitive system can tolerate and the work of meaningful learning must fit within what an individual is capable of accomplishing. Here, we are most interested in helping educators identify the multiple sources of load that influence learner achievement. In addition to work or load of learning there are other sources of load educators should recognize. These multiple sources sources are additive and if the total load exceeds the cognitive capacity of working memory, the system becomes less effective. When possible, the idea is to control what can be controlled to avoid situations in which the combined sources of load exceed possible.
The various sources of load involved in learning situations have been studied by cognitive scientists (e.g., Paas, Renkl, & Sweller, 2004). Our somewhat simplified version suggests that the sources of load in educational settings might be identified as (1) the cognitive activity involved in meaningful learning, (2) the cognitive activity to accomplish any external tasks intended to facilitate learning, and (3) the cognitive activities not directly relevant to learning. You might attempt to apply this system to situations familiar to you. Perhaps you commonly study a textbook by highlighting important passages while you read and you like to study in your room while listening to music. Highlighting adds an external task to your attempts at comprehension and retention that may or may not be beneficial. You think about what to highlight and this adds a cognitive activity to basic comprehension activities. Whether highlighting is beneficial or not is not the issue here, the point is that highlighting involves an additional cognitive demand. In contrast, listening to music has nothing to do with understanding the content of the book. It may be an experience that will keep you working longer, but within the moment, any attention captured by the music cannot be focused on understanding and retention or on making decisions about highlighting. This system may be effective if the combination does not exceed the capacity of working memory. It may also be ineffective for some learners. If ineffective, it would seem reasonable to try eliminating the background music before making modifications to study strategies.
It can be helpful to analyze instructional/learning tasks from this perspective. For example, consider why flight simulators are useful at certain points in pilot training. Simulations have multiple advantages which we address elsewhere. One way to describe some of these advantages would be as a way to control load. The use of a simulator in the early stages of pilot training is helpful because it simplifies the learning task by controlling the variables students need to attend to (task associated with the learning task) and because it reduces distractions and anxiety (extraneous load).
Now, with this background, the question becomes what does this perspective have to do with the controversy regarding direct instruction versus more experiential learning activities? We will begin by briefly summarizing the controversy. The dispute between those advocating direct instruction and those advocating more hands-on, experiential methods goes back at least a half-century. We focus on a specific analysis (Kirschner, Sweller, & Clark, 2006), but this analysis is but one of several careful reviews of research comparing learning experiences. Simply put, the argument is that strategies such as problem-based learning, inquiry learning, and other experiential strategies involve greater cumulative cognitive load than direct instruction. We would suggest this difference in load could involve either relevant or irrelevant cognitive activities and find it helpful to use Ausubel's (1963) concept of how learners get to the point of learning. With direct instruction, less learner cognitive activity is required to get to the point of learning and a larger proportion of learners would be likely to have sufficient remaining cognitive capacity to have successful learning experiences.
We find a potential limitation in this argument. The concept of working memory capacity describes a bottleneck in the cognitive system that is largely a problem for simultaneous activities. This concern would fit our previous example of reading comprehension, highlighting, and listening to music. Experiential tasks, however, do not necessarily require simultaneous activities. So, for example, a simulation task intended to demonstrate a principle to be learned does not necessarily overload cognitive capacity if the simulation task and the learning of whatever principle emerges from the exploration of the simulation do not happen simultaneously. So, the significance of the concern that certain types of learning tasks overload working memory may depend on whether there is an opportunity to separate opportunities to think carefully about important ideas from the activities responsible for generating these ideas. We do not see known limitations of working memory as a "fatal flaw" in arguments against project based learning and other forms of experiential learning.
However, there is another and possibly related issue we think educators must consider. We have taken to labeling this as the challenge of educational efficiency. One reason learners are sometimes not allowed the necessary time for reflection following experiential tasks is that time within educational settings is always in short supply. Analyses of the time available in school (for a brief summary see Berliner, 2007) make the points that the time available for learning is limited and the increasing complexity of the demands placed on schools (e.g., meeting the needs of learners from increasingly diverse backgrounds, teaching skills that were previously not the responsibility of schools) only adds to the expectations for what teachers and learners must accomplish within a limited amount of time. This reality exists despite evidence "deep learning" in comparison to fact retention is disproportionally improved by increased learning time. In other words, schools are put in a difficult situation by a combination of expectations and constraints on how schools traditionally operate. Ask an educator you know and you are likely to get an ear full on this topic. Suggestions to remedy this issue have included lengthening the school day or the school year (e.g., Matthews, 2009). In fact, one of the arguments for 1:1 technology initiatives (each student has a computer) is that educators can be assured that students would then have technology resources available at home offering a possibility for extending the school day. However the practical issue of time for learning is addressed, educators do need to be sensitive the issue of efficient use of school time.
There are two efficiency issues we propose teachers should consider in evaluating possible technology-supported, classroom activities. These might be described as the efficiency of the learning task and the efficiency of the learning tool. The issue of task efficiency is similar to issues raised by Ausubel or Kirchner and colleagues. We do not see this as a working memory load concern, but rather an issue of the time it takes to complete the task relative to the amount of time in "deep learning" the task encourages. We do not want to give the impression that the variables we propose have or maybe even can be accurately measured, but we describe the issue in this way because we think if teachers consider this perspective they will understand the issues we are addressing.