Imagining instructions: mental practice in highly cognitive domains.

Ginns, Paul

This article reviews recent empirical investigations of imagination or mental practice in highly cognitive, realistic educational domains such as mathematics or learning computer applications. While mental practice has been a standard tool in training schedules devised by sports psychologists for several decades, with its efficacy studied experimentally in a multitude of sports, there has been little corresponding research in the education or training research literature. Recent research has demonstrated that mental practice can be incorporated effectively when learning non-motor, complex cognitive skills. Experimental studies are reviewed showing 'imagining' worked examples, paired with practice questions, enhances learning for more experienced learners, but study activities are more appropriate for students less experienced in a given domain. Interactions of the imagination effect with cognitive load effects are also discussed. Possible directions for mental practice research in education are proposed.

Introduction

Imagery, visualisation and mental practice are terms for a class of cognitive processes which may enhance learning under certain circumstances. They involve quasi-sensory conscious experiences, or as Anderson (1981, p. 150) notes, 'they have in common the awareness of sensory qualities in the absence of appropriate external stimuli'. He goes on to note that while the term 'imagery' has a visual connotation, it need not be restricted to this modality, and recommends using the term 'imaginary' over 'imaginal' to reduce the visual connotation when discussing such phenomena. The mnemonic use of such processes is ancient, in particular, the Method of Loci ascribed to Simonides the orator (Yates, 1966). Imaginary processes have provided the basis for several more recently developed and empirically validated mnemonics, such as the keyword method (Atkinson, 1975) and the face-name method (Carney & Levin, 2000).

Whereas the focus of the above-mentioned mnemonics has traditionally been learning of declarative information (e.g., facts, vocabulary), the term 'mental practice' has traditionally been used in the context of developing procedural knowledge, such as physical or cognitive skills. Considerable evidence exists that mental practice, '... the symbolic, covert, mental rehearsal of a task in the absence of actual, overt, physical movement' (Driskell, Copper, & Moran, 1994, p. 481), assists learning a broad range of motor, perceptuo-motor, and cognitive skills. Synonyms used by researchers for 'mental practice' have included 'symbolic rehearsal' (Sackett, 1934; 1935), 'imaginary practice' (Perry, 1939), 'covert rehearsal' (Corbin, 1967; Peynircioglu, 1995), and 'mental rehearsal' (Dunbar, 2000; Rawlings & Rawlings, 1974).

Mental practice may be related to other instructional effects besides imagery and visualisation. Cooper, Tindall-Ford, Chandler, and Sweller (2001) note that the processes underlying effective mental practice may be similar to those underlying effective self-explanation (e.g., Chi, Bassok, Lewis, Reimann, & Glaser, 1989; Wong, Lawson, & Keeves, 2002). The latter body of research has found effective problem solvers are more likely to explain to themselves the relations between elements in to-be-learned material (e.g., parts of a worked example). One form of self-explanation identified by Renkl (1997), anticipative reasoning, occurs when a learner--typically with a relatively high level of prior knowledge--anticipates upcoming steps in a worked example (e.g., the next intermediate step in a multi-step probability calculation) in their mind, rather than simply studying the solution. Cooper et al. (2001, p. 80) argue this process can be considered 'a natural form of mental practice'.

Mental practice has been widely studied and advocated by sports psychologists as a means of enhancing performance at both an amateur and professional level, and many top sportspeople use the technique, often in conjunction with other sports psychology techniques like biofeedback and relaxation (Murphy & Jowdy, 1992). Evaluations of mental practice with complex instructional materials, particularly of major elements in school, higher education, or organisational training curricula, are currently in the minority of published mental practice studies. The overall goal of this article will be to highlight ways in which mental practice can be incorporated effectively into learning programs. A brief overview of the history of mental practice research is given, including the most recent meta-analysis of the research literature. More recent studies, which investigate mental practice of worked examples in realistic educational domains, will then be discussed. Opportunities for further research into both the design and application of mental practice activities are also suggested.

Empirical investigations of mental practice

Empirical studies of mental practice date from the 1930s. Initial experiments involving mental practice investigated its effect on tasks with a high cognitive component. Sackett (1934, 1935) investigated the effect of what he called 'symbolic rehearsal' on the retention of a 'maze habit', while Perry (1939) investigated whether 'imaginary practice' affected performance on a tapping task, a card-sorting task, a peg board task, a symbol digit substitution task, and a mirror tracing task. Relatively little research was carried out on mental practice in the 1940s and 1950s, but with the advent of sports psychology in the 1960s and 1970s interest in its application to sporting performance increased considerably. However, the voluminous body of research that resulted from this rediscovery was marked by considerable theoretical debate as to the precise nature of mental practice effects. Initial reviews of the mental practice research literature were contradictory. Richardson (1967) concluded that mental practice was associated with improved motor performance in the majority of studies reviewed, whereas Corbin (1972) argued that the situation was not so clear-cut as Richardson had proposed, with individual, task and methodological factors all moderating the impact of mental practice on performance. More recently, the development of meta-analysis (Glass, 1976) has provided a quantitative method for reviewing and assessing both overall effects and moderators of mental practice.

Meta-analysis of mental practice studies

The most recent and comprehensive meta-analysis of the mental practice literature was conducted by Driskell et al. (1994). They selected studies which had used precise, uni-dimensional operationalisations of mental practice, and only those studies which compared the performance of participants engaging in mental practice with a no practice control group (studies including a no practice control group as well as other comparison conditions were also retained). Included studies were coded for five factors which were hypothesised to moderate the effectiveness of mental practice: type of task (whether the task was primarily cognitive or physical), the time interval between practice and assessment, the experience level of trainees, the length of practice, and the type of control group used.

On the basis of these selection criteria, Driskell et al. (1994) meta-analysed 100 significance tests from 35 different studies. The average effect size of the 62 hypothesis tests examining mental practice was described as small to moderate in magnitude (d = 0.53). The 39 hypothesis tests in which physical practice was examined had an average effect size which was moderate to strong (d = 0.78). The difference between the overall magnitudes for mental and physical practice was found to be significant at the 0.01 level; thus, Driskell et al. concluded that while mental practice had a moderate and significant impact on performance, this effect tended to be weaker on average than physical practice. The following variables were found to moderate the overall effect at a significance level of 0.01:

* Type of Task: mental practice was more effective the more a task required cognitive operations (d = 0.69) compared to physical tasks (d = 0.38). The smaller impact of mental practice on physical performance tasks was particularly determined by the extent to which the task had a high strength component, and to a lesser extent by the presence of a coordination component.

* Length of Retention Interval: a significant negative relationship was detected between the length of the retention interval and the mental practice effect magnitude (r = -0.22); thus, longer delays between practice and performance were associated with weaker effects of mental practice on performance.

* Experience Level: both complete novices and participants with task experience were found to benefit from mental practice to a moderate degree (d = 0.46 and 0.54, respectively), with the difference in magnitude between the two groups not significantly different. However, this basic moderator was moderated by another variable, type of task. Thus, for novice subjects, mental practice resulted in stronger effects for cognitive tasks than physical tasks. In contrast, experienced subjects benefited from mental practice at a similar level for both physical and cognitive tasks.

* Duration: no significant effect was found between the number of practice trials and the mental practice effect size. A significant negative relationship was found between the duration of mental practice and the effect magnitude (r = -0.19); thus, 'the longer someone mentally practices, the less beneficial it becomes' (Driskell et al., 1994, p. 488).

Driskell et al.'s (1994) results indicate task nature is an important predictor of mental practice efficacy, supporting the symbolic learning theory of mental practice first proposed by Sackett (1934). According to this theory, '... mental practice gains are more due to the opportunity to practice the symbolic elements of a motor task than to muscle activation itself' (Suinn, 1997, p. 195). Examples of potential cognitive elements include temporal or spatial elements of the skill, possible courses of action, or sequences. There is, however, little extant research investigating the effectiveness of mental practice for skills that are almost entirely cognitive in nature, wherein any physical components are often limited to highly overlearned or trivial aspects, such as writing down the results of arithmetic mentation. The development of such skills forms a large part of educational and workplace training curricula, and mental practice may therefore be a simple yet powerful adjunct to existing methods for learning such skills. The following sections describe a program of research into a novel approach to mental practice, using the instructional mechanism of worked examples.

Mental practice of worked examples

Over the past two decades, a considerable body of research has amassed demonstrating the educational advantages and limitations of worked examples. Appropriately designed worked examples are argued to support construction and automation of schemas because they 'focus attention on problem states and associated operators (i.e., solution steps) enabling learners to induce generalized solutions or schemas' (Sweller, van Merrienboer & Paas, 1998, p. 273). These instructional devices 'typically include a problem statement and a procedure for solving the problem; together, they are meant to show how other similar problems could be solved' (Atkinson, Derry, Renkl, & Wortham, 2000, p. 181). An example of a geometry worked example used in Ginns, Chandler and Sweller's (2003) Experiment 2 is given in Figure 1. The worked example includes the instructions given to students in the mental practice experimental condition.

[FIGURE 1 OMITTED]

A typical approach taken by students with worked examples is to study them: that is, to read over the worked example and attempt to understand the steps involved in reaching a solution. When students have sufficient prior knowledge, and worked examples are designed in ways which minimise extraneous cognitive load, studying one or two worked examples paired with practice questions has been found to support schema construction compared with solving an equivalent number of problems (Sweller & Cooper, 1985). Moreover, studying an extended series of worked examples paired with paired practice questions has been found to favour both schema construction and automation compared to solving an equivalent number of problems (Cooper & Sweller, 1987). Might alternative instructions to studying worked examples fast-track the automation process?

An extended research program is currently being conducted on mental practice in highly cognitive domains. Reporting the initial results of this program, Cooper et al. (2001) flamed mental practice of worked examples within the broader context of deliberate practice (Ericsson, Krampe, & Tesch-Romer, 1993). The theory of deliberate practice holds that practice of a given task in isolation is not guaranteed to be effective in developing high levels of expertise. Rather, conscious intention to enhance the skill (or set of skills) is required on the part of the student, in turn motivating a mentor-supported, systematic, varied and extended practice schedule. By comparison, practice schedules consisting of simple practice, general exposure or play fall outside the realm of deliberate practice and are thus not expected to support high-level learning. Given its intrinsically conscious and effortful nature, mental practice, Cooper et al. (2001, p. 68) argue, may therefore be '... an effective and efficient form of deliberate practice'.

Taking Driskell et al.'s (1994) first finding of a stronger mental practice effect for tasks with a higher cognitive component, Cooper et al. (2001) investigated its effectiveness in highly cognitive domains such as geometry and using spreadsheets. In such domains, the physical component of successful performance is relatively minor, generally being restricted to paper-and-pencil operations or keyboard operations. In the first study, students studied at their own pace a computer-based tutorial on introductory spreadsheet use, including selecting cells, entering data, and constructing formulae to act upon data. Information related to the use of these functions was presented on-screen, under the student's control, by a series of text overlay boxes. In addition to this self-paced initial time on instruction, on completion of each overlay box sequence students completed an additional acquisition phase where they were instructed to study or imagine the complete worked example.

No significant difference was found between groups on time on initial instructions (i.e., time spent reading the materials). On the test questions, students who imagined worked examples for spreadsheet operations achieved higher test scores (d = 0.46) and spent less time on questions (d = 0.64) than those who studied the materials. However, the imagination group also spent longer imagining the worked examples than the study group on a one-tailed t-test (d = 0.48), raising the possibility that the increased time spent imagining was the source of their augmented performance.

In Experiment 2, time on acquisition (i.e., imagine versus study time) was controlled for by equalising acquisition time at 30 seconds studying or imagining per worked example (following self-paced initial instruction from the text overlay boxes). Using spreadsheet computer-based tutorial materials similar, apart from minor modifications, to Experiment 1, Cooper et al. (2001) replicated the findings of the first study. Again, no significant difference was found between groups on time on initial instruction, excluding this variable as an explanation for any group differences. The mental practice group again achieved a higher average test performance score (d = 1.04) and spent less time on test questions (d = 1.02) than the study group.

The results obtained in the first and second experiments were reversed in the third experiment, again using a computer-based spreadsheet tutorial. This study was conducted earlier in the school year, meaning that students had not completed the algebra component of the Grade 7 mathematics syllabus. In addition, teachers at that school stated that '... in their view, the students of this school had a lower skill level in mathematics and knowledge of computers than students at the schools used in the previous experiments' (Cooper et al., 2001, p. 74). These inter-school differences were expected to reverse the mental practice effect, as students would lack requisite formula construction schemas possessed by students in the previous studies. Time spent studying was predicted to be more productive than time spent imagining, because studying would provide more opportunity than imagining for schema construction activities. Again, no significant difference was found between groups on initial instruction time. As predicted, instructions to study produced higher average test performance than imagine instructions (d = 0.79), as well as lower average times to test solution (d = 0.64).

In the fourth experiment, Cooper et al. (2001) provided further evidence for the importance of prior ability/knowledge levels to mental practice efficacy by controlling knowledge and ability as an independent variable. Students from both high-ability and low-ability graded mathematics classes were randomly allocated to imagine or study conditions for a computer-based spreadsheet tutorial. Time on initial instruction, left as self-paced in the previous studies, was fixed across students in this study. Cooper et al. found a main effect favouring the high-ability group on test scores and times, but also found an interaction between level of knowledge and type of instruction. For high ability students, imagine instructions were associated with higher test performance than study instructions (d = 0.63), while the reverse was true for the low ability students (d = 0.66).

In their fifth and final experiment, Cooper et al. (2001) provided additional empirical evidence for a schema acquisition/schema automation account of mental practice benefits, arguing that prior knowledge level, not ability per se, is the principal determinant of such benefits. Two groups of novices, matched on ability, were compared over a two-phase, self-paced learning sequence for coordinate geometry materials presented via a computer-based tutorial. After the first phase of initial instructions, one group imagined worked examples, while the other studied the same materials. After the second phase, students either studied or imagined a second set of worked examples, with the instructions received in the first phase reversed for each group. Cooper et al. hypothesised that study instructions after the first phase would support schema acquisition processes, with imagine instructions after the second phase further supporting schema automation processes. By comparison, the alternative sequencing (imagine then study) was expected to support neither initial construction nor later automation.

Time spent on self-paced instruction was virtually identical for the two conditions, discounting this variable out as an explanation for any subsequent group differences. Ceiling effects prevented analysis of percentage of test questions correct; however, students in the study-imagine condition spent almost half the time answering test questions compared with those in the imagine-study condition (d = 1.17). Cooper et al. (2001) argued this significant difference could best be explained with reference to differential support between the conditions for the construction and automation of relevant schemas, since students were matched on abilities.

Further studies by Ginns et al. (2003) support Cooper et al.'s (2001) conclusions regarding study and mental practice sequencing. Experiment 1 investigated the use of mental practice in learning complex materials. Novices in the domain of HTML, a quasi-programming language designating layout of screen elements on Internet pages, studied a self-paced computer-based tutorial on introductory HTML, including set periods of imagination or study of HTML worked examples. Attempting to understand and learn both the specifics of HTML (such as the various 'tags') and the more general structure of HTML (particularly nesting tags within each other) would be expected to impose a high cognitive load on novices. Under these conditions of almost no relevant prior knowledge it was argued schema construction activities (supported by instructions to study) are more appropriate than schema automation activities (supported by instructions to imagine). The study group outperformed those in the imagine group on a composite HTML coding test score using a one-tailed t-test (d = 0.83), supporting the hypothesis that mental practice would be a counter-productive exercise for novices.

Experiment 2 investigated the use of mental practice in high school geometry, a domain where mitigating factors of prior knowledge and material complexity could be controlled more explicitly than was possible with HTML. Prior knowledge was controlled by (a) using paired sampling according to a grade-wide mathematics test, and (b) by running the experiment at a point in the school year when necessary prior knowledge had been obtained from class activities, but the specific geometry rules had not been learned. Complexity of the materials was controlled by designing the worked examples to have simple two-step solution paths, considerably simpler than the nested problem spaces typical of HTML. After studying two geometry rules in isolation then in a combined worked example, Grade 7 students who imagined further worked examples paired with practice questions solved similar test questions at speeds comparable to those in the study condition, but solved significantly more transfer questions (d = 1.14) and solved them more quickly (d = 1.08) than students who studied. This was in accordance with predictions that the mental practice and study conditions should not differ in effect on similar problem performance (as both conditions were expected to support initial schema construction), but should differ on transfer test performance because of greater support for automation provided by imagining. However, because of ceiling effects on the similar test questions, it was not possible to conduct a statistical analysis of mental practice's differential effect on similar versus transfer test performance, and the schema automation hypothesis remains a plausible but unproven explanation of the results of Experiment 2.

Experiment 3 replicated Cooper et al.'s (2001) fifth study, again contrasting study then imagination of worked examples to an imagination then study sequence. Prior knowledge was controlled in a similar fashion to Experiment 2, with paired sampling of students and by conducting the experiment at a time in the school year when students could be expected to have acquired prerequisite knowledge. Students in the imagine-study group were expected to learn less in the first stage because imagine instructions were not expected to support schema construction as much as study instructions. In the second phase, the instructions for each group were reversed, so that students who had initially studied worked examples imagined further examples in the second phase and vice versa. This reversal was predicted to favour the study-imagine group, because their initial advantage in schema construction would be compounded by imagine instructions supporting schema automation. Students in the imagine-study group, in contrast, were expected to neither construct (in phase 1) nor automate schemas (in phase 2) to a level comparable with the other condition. Ceiling effects were found on test scores, precluding analysis of this variable. However, a small yet significant difference in total time to solution on test questions, favouring the study-imagine group (d = 0.24), provided some support for study (supporting initial schema construction) followed by imagination (supporting schema automation) of worked examples.

While the above imagination effects were all found with high school students, Leahy and Sweller (2004) have replicated these results with adult learners. Twenty-four primary school teachers were randomly assigned to a study condition and an imagination condition. The learning task was interpreting contour maps, a topic with which participants had some prior knowledge, as map reading is part of the primary school curriculum in the state where the participants worked. Students either studied or imagined the steps involved in calculating gradient ratios using paper-based worked examples, with the time for studying or imagining equivalent for both conditions. In the test phase, learners in the imagination condition significantly outperformed those in the study conditions (d = 0.90). Leahy and Sweller (2004) thus extended the findings of Cooper et al. (2001) and Ginns et al. (2003) by demonstrating mental practice of worked examples is effective for adult learners. But might other effects associated with cognitive load moderate these learning gains?

The importance of cognitive load

The results discussed in the section above have been framed within Cognitive Load Theory (for reviews see Sweller, 1999, 2003; Sweller et al., 1998). The various cognitive load theory effects largely concern the design of materials to facilitate schema construction by novices, with full or partial automation being a possible secondary benefit of redesigns to reduce extraneous cognitive load. If the cognitive load theory account of mental practice described above is accurate, predictable interactions should be obtained when mental practice is investigated in tandem with experimental instructional materials designed to reflect one of the cognitive load effects.

Leahy and Sweller (2004) investigated interactions between the imagination and split attention effects. The split attention effect occurs when learners are presented with instructional materials requiring them to mentally integrate disparate sources of information (e.g., text and associated referents in a mathematical diagram) which are unintelligible in isolation. Under such conditions, a learner's working memory resources must be directed towards first mentally integrating the disparate sources of information, meaning these resources are unavailable for schema construction. Alternatively, if related sources of information are integrated (e.g., by positioning text and related graphical elements closely together), split attention is minimised, and working memory resources can be applied directly to schema construction (Sweller, 1999). Leahy and Sweller (2004) theorised that learners given split-source materials would find imagining more difficult than if given integrated materials, reducing test performance. Likewise, students in the study/integrated or study/split-source conditions were hypothesised to have fewer opportunities to construct and automate schemas than those in the imagine/integrated condition. Using a 2 (imagine vs. study) x 2 (integrated vs. split-source) factorial design, they tested Grade 4 students' learning of line graph worked examples representing temperature variations. They found a significant interaction between the two experimental factors, with students in the imagine/integrated condition outperforming students in the other three conditions on test scores (d = 1.29 for imagine/integrated condition vs. average of other conditions).

The experiments described above indicate mental practice of worked examples can lead to substantial learning gains compared to conventional studying. However, this general effect is moderated by knowledge level. Students who lack specific schematic knowledge should study until they have incorporated the to-be-learned information into a unified schema, capable of being retrieved from long-term memory, held active in working memory, and imagined. Moreover, the format in which to-be-learned information is presented may have important consequences for the ease with which learners first understand, then imagine the instructions.

Future directions

There is still much to learn about the boundary conditions of imagination of worked examples. Fundamental questions remain about evidence for a schema automation explanation of the imagination effects described above. The schema automation explanation of the imagination and study-imagination effects described above provided to date rests on a theory-driven expected pattern of results between similar and transfer tests. It draws on Kotovsky, Hayes and Simon's (1985) argument that rule automation primarily supports transfer test performance. Automated rules are held to lighten the considerable cognitive load on working memory typical during problem-solving of transfer problems by 'weak' problem-solving methods such as means-ends analysis (Sweller, 1988), increasing chances of a solution being found. However, Cooper et al. (2001) did not find a clear distinction between performance on similar versus transfer tests across their complex of experiments. Ginns et al. (2003) also obtained ceiling effects on similar test scores and times in Experiment 2, and again on similar question scores and times in Experiment 3. Future studies require materials that allow sufficient variance on both similar and transfer tests if valid tests of a 'construction then automation' (i.e., study then imagine) sequence theory can be made. A schema automation explanation would also benefit from process variable evidence, such as performance on secondary measures. Modification of dual-task methods such as those used by Chandler and Sweller (1996) or Brunken, Steinbacher, Plaas and Leutner (2002) may provide such evidence.

Another important issue for the use of mental practice concerns the present unavailability of a suitable mental practice readiness metric. While such a metric should help develop theories regarding mental practice, as discussed above, there is also a need for a practical metric (or failing that, a set of heuristics) to identify readiness that educators can use in day-to-day teaching; otherwise, educators must rely on their intuition to judge students' readiness levels for a given domain. Such a metric or set of heuristics should help teachers identify if a student has unified several related elements of information into a single schema capable of being held in toto in working memory, meaning that s/he is thus capable of imagining the schema. Kalyuga and Sweller (2004) have developed a method for rapidly assessing whether a student possesses a schema for a given problem, based on measuring the extent to which a student can correctly state the first step in solving a given class of problems. They found performance on such tests was highly correlated with more traditional tests of knowledge in algebra and geometry. In subsequent experiments, they demonstrated that when the rapid schema assessment method was used to assign learners to instructional conditions consonant with their knowledge levels, these students outperformed students who viewed instructional materials that were either dissonant (Experiment 3) or random (Experiment 4) with respect to their knowledge levels. These results suggest that the rapid schema assessment method might also be used by educators to determine whether or not a student would benefit from switching from studying to imagination. A high score on a rapid schema assessment test would indicate the student has constructed a schema capable of being held in working memory, and that further study at this point would be redundant. Mental practice, in contrast, should act to automate the schema. Further research is needed to test this possible application of the rapid schema assessment method.

Finally, there is a pressing need for research that replicates the above experimental results in more naturalistic settings. However, the replication of the worked example effect originally generated using one-on-one experimental testing (Cooper & Sweller, 1987; Sweller & Cooper, 1985), in realistic classroom settings (Carroll, 1994; Ward & Sweller, 1990; Zhu & Simon, 1987) gives reason to believe that the imagination effect (i.e., mental practice of worked examples) should be similarly robust in one-to-many educational settings.

Conclusion

This review has discussed an instructional intervention, mental practice, whose effectiveness has been experimentally demonstrated across a wide range of domains. Meta-analysis of such experiments has indicated its effectiveness is moderated by several factors, in particular the extent to which the task involves cognitive components (Driskell et al., 1994). More recent research (Cooper et al., 2001; Ginns et al., 2003; Leahy & Sweller, 2004) has built on these core findings, demonstrating that imagination of worked examples is instructionally effective and efficient compared to study, provided students have appropriate prior knowledge levels, and the worked examples are designed appropriately so as to minimise extraneous cognitive load.

While these conclusions are based on a relatively small number of studies, the strength of the effects, their generation using both paper-based and computer-based learning materials across several domains, and their basis in clearly delineated theory should give instructional designers confidence in applying these results to their own contexts. Typically, these effects are generated through periods of mental practice and linked practice questions of less than ten minutes, over and above general instruction. The availability of a number of review publications on the effective design of both worked examples (e.g., Atkinson et al., 2000; Sweller, 1999, 2003; Sweller et al., 1998) and multimedia learning resources (e.g., Mayer, 2001), emphasising considerations of cognitive architecture, should be of particular assistance in successfully designing either paper-based or multimedia learning resources incorporating mental practice. The strength of these findings should likewise be of interest to educational researchers, in motivating further research to determine the generalisability of these effects across different educational domains; the identification of additional moderating variables besides prior knowledge; and easily administered methods for assessing mental practice readiness.

Keywords

imagination

cognitive processes

cognitive load theory

visualisation

meta analysis

cognitive skills

References

Anderson, M. P. (1981). Assessment of imaginal processes: Approaches and issues. In T. M. Merluzzi, C. R. Glass, & M. Genest (Eds.), Cognitive assessment. New York: Guilford Press.

Atkinson, R. C. (1975). Mnemotechnics in second-language learning. American Psychologist, 30, 821-828.

Atkinson, R. K., Derry, S. J., Renkl, A., & Wortham, D. (2000). Learning from examples: Instructional principles from the worked examples research. Review of Educational Research, 70, 181-214.

Brunken, R., Steinbacher, S., Plaas, J.L., & Leutner, D. (2002). Assessment of cognitive load in multimedia learning using dual-task methodology. Experimental Psychology, 49, 109-119.

Carney, R. N. & Levin, J. R. (2000). Mnemonic instruction, with a focus on transfer. Journal of Educational Psychology, 92, 783-790.

Carroll, W. (1994). Using worked examples as an instructional support in the algebra class-room. Journal of Educational Psychology, 86, 360-367.

Chandler, P, & Sweller, J. (1996). Cognitive load while learning to use a computer program. Applied Cognitive Psychology, 10, 151-170.

Chi, M., Bassok, M., Lewis, M., Reimann, P., & Glaser, R. (1989). Self-explanations: How students study and use examples in learning to solve problems. Cognitive Science, 13, 145-182.

Cooper, G. & Sweller, J. (1987). Effects of schema acquisition and rule automation in mathematical problem-solving transfer. Journal of Educational Psychology, 79, 347-362.

Cooper, G., Tindall-Ford, S., Chandler, P., & Sweller, J. (2001). Learning by imagining. Journal of Experimental Psychology: Applied, 7, 68-82.

Corbin, C. B. (1967).The effect of covert rehearsal on development of a complex motor skill. Journal of General Psychology, 76, 143-150.

Corbin, C. B. (1972). Mental practice. In W. P. Morgan (Ed.), Ergogenic aids and muscular performance. San Diego, CA: Academic Press.

Driskell, J. E., Copper, C., &Moran, A. (1994). Does mental practice enhance performance? Journal of Applied Psychology, 79, 481-492.

Dunbar, R. I. M. (2000). Causal reasoning, mental rehearsal, and the evolution of primate cognition. In C. Heyes, & L. Huber (Eds.), The evolution of cognition: Vienna series in theoretical biology (pp. 205-219). Cambridge, MA: The MIT Press.

Ericsson, K.A., Krampe, R.T., & Tesch-Romer, C. (1993). The role of deliberate practice in the acquisition of expert performance. Psychological Review, 100, 363-406.

Ginns, P., Chandler, P., & Sweller,J. (2003).When imagining information is effective. Contemporary Educational Psychology, 28, 229-251.

Glass, G.V. (1976). Primary, secondary, and meta-analysis of research. Educational Researcher, 5, 3-8.

Kalyuga, S. & Sweller, J. (2004). Measuring knowledge to optimize cognitive load factors during instruction. Journal of Educational Psychology, 96, 558-568.

Kotovsky, K., Hayes, J. R., & Simon, H.A. (1985).Why are some problems hard? Evidence from Tower of Hanoi. Cognitive Psychology, 17, 248-294.

Leahy, W. & Sweller, J. (2004). Cognitive load and the imagination effect. Applied Cognitive Psychology, 18, 857-875.

Mayer, R. E. (2001). Multimedia learning. Cambridge: Cambridge University Press.

Murphy, S. M. & Jowdy, D. P. (1992). Imagery and mental practice. In T. S. Horn (Ed.), Advances in sport psychology. Champaign, IL: Human Kinetics.

Perry, H. H. (1939). The relative efficacy of actual and imaginary practice in five selected tasks. Archives of Psychology, 34, 5-75.

Peynircioglu, Z. F. (1995). Covert rehearsal of tones. Journal of Experimental Psychology: Learning, Memory, & Cognition, 21, 185-192.

Rawlings, E. I. & Rawlings, I. L. (1974). Rotary pursuit tracking following mental rehearsal as a function of voluntary control of visual imagery. Perceptual and Motor Skills, 38, 302-309.

Renkl, A. (1997). Learning from worked-out examples: A study on individual differences. Cognitive Science, 21, 1-29.

Richardson, A. (1967). Mental practice: A review and discussion (Part 1). Research Quarterly, 38, 95-107.

Sackett, R. S. (1934). The influences of symbolic rehearsal and retention of a maze habit. Journal of General Psychology, 10, 376-395.

Sackett, R. S. (1935). The relationship between symbolic rehearsal and retention of a maze habit. Journal of General Psychology, 13, 113-128.

Suinn, R., M. (1997). Mental practice in sport psychology:Where have we been, where do we go? Clinical Psychology: Science and Practice, 4, 189-207.

Sweller, J. (1988). Cognitive load during problem solving: Effects on learning. Cognitive Science, 12, 257-285.

Sweller, J. (1999). Instructional design in technical areas. Camberwell,Vic.:ACER Press.

Sweller, J. (2003). Evolution of human cognitive architecture. In B. Ross (Ed.), The psychology of learning and motivation (pp. 215-266). San Diego, CA: Academic Press.

Sweller, J. & Cooper, G. (1985).The use of worked examples as a substitute for problem solving in learning algebra. Cognition and Instruction, 2, 59-89.

Sweller, J, van Merrienboer, J., & Paas, F. (1998). Cognitive architecture and instructional design. Educational Psychology Review, 10, 251-296.

Ward, M. &Sweller, J. (1990). Structuring effective worked examples. Cognition and Instruction, 7, 1-39.

Wong, R. M. F., Lawson, M. J., & Keeves, J. (2002). The effects of self-explanation training on students' problem-solving in high-school mathematics. Learning & Instruction, 12, 233-262.

Yates, F.A. (1966). The art of memory. Chicago: University of Chicago Press.

Zhu, X. & Simon, H. (1987). Learning mathematics from examples and by doing. Cognition and Instruction, 4, 137-166.

Dr Paul Ginns is at the Institute of Teaching and Learning, University of Sydney, NSW 2006.

Email: p.ginns@itl.usyd.edu.au