Review of Munyofu et al.’s
the Effect of Different Chunking Strategies
in Complementing Animated Instruction
Faculty of Education, University of Wollongong, Australia
Munyofu, M, Swain, W J, Ausman, B D, Lin, H, Kidwai, K & Dwyer, F 2007, ‘The effect of different chunking strategies in complementing animated instruction’ Learning, Media and Technology, vol.32, no.4, 407-419.
Munyofu et al.’s (2007) “the Effect of Different Chunking Strategies in Complementing Animated Instruction” aimed to examine the instructional effects of different chunking strategies used to complement animated instruction in terms of facilitating achievement of higher order learning objectives.
They first reviewed the chunking theory and the use of animation, summarising and comparing different researchers’ findings. Then they drew the common grounds of the two fields respectively. It is believed that information is processed by human as chunks of similar equivalent data but not in single strands or discrete entities, so it is helpful to combine elements to form a smaller number of groups when presented with a large set of elements (Simon 1974). There are different explanations to such chunking strategies. Whatever theory or model, all research studies have concluded that chunking is an important tool in learning because it helps to reduce information overload. Chunking of text significantly enhances short term memory (STM) and the process of transfer from STM to long term memory (LTM), helps learners to transform information from general to specific and to understand relationships between given items information, and facilitates learners to progress more efficiently towards higher levels of learning (Miller 1956; Newell & Simon 1972; Chase & Simon 1973; Cooper 1998; Gobet, F. 1998; Gobet, F. 1998; Gobet, F. & Simon 1998; Gobet, F. & Simon 1998; Lane, Gobet & Cheng 2000; Carter, Hardy & Hardy 2001; Cowan 2001). On the other hand, animation is believed to assist attention-gaining, presentation and practice in instruction, and contribute to computer-based instruction by providing strategies that facilitate interaction between students and the content (Reiber 1990). Based on these findings, Munyofu et al. (2007) hypothesised that the use of simple and complex chunks of text and static visuals used to complement animated instruction would differentially affect performance on tests measuring different educational objectives, and stated two research questions: (1) how effective animation might be for different types of learning outcomes and (2) how different types of instructional strategies used to complement animated instruction might improve the effectiveness of animated instruction.
To explore the hypothesis and answer the research questions, Munyofu et al. (2007) organised a small-scale comparative study involving eighty-five students in three treatment groups: animated program instruction, simple visual-text (static images and verbal explanation) chunked animated program instruction and the animated complex visual-text chunked program instruction. The difference between simple and complex chunked instructions was the content. Simple chunks only dealt with one content area while the complex chunks explained two or more related content areas. The students interacted with their respective web-based instructional treatments and completed four criterion measures.
The results indicated that significant differences in achievement were found to exist in facilitating higher order learning objectives when chunking strategies were specifically designed and positioned to complement the animated instruction. The results also indicated that complex chunking is more effective in reducing the cognitive load present in an animated instructional environment, and that students need prerequisite knowledge before being able to profit from animated instruction designed to facilitate higher order learning outcomes (Munyofu et al. 2007).
The main contribution of the study is to combine the two domains of previous research—chunking theory and the use of animation in instructional material design, and to explore what would happen if the two theories are applied together to different extent. Multimedia and interface designers would benefit from the study in terms of selection of chunking strategies for users’ better understanding and more effective learning. However, the reason why different chunking strategies lead to different cognitive results was not explored enough, which was a weakness of the study.
In this review, the research methodology and the results of Munyofu et al.’s (2007) study will be analysed and discussed. To provide developers and designers of interactive learning environment and effective interfaces, this review will also highlight the implications of the study to provide guidelines for relevant practice.
Pilot Studies and Treatment Design
The main method of Munyofu et al.’s (2007) study was to measure and compare the learning outcomes of three groups of students who received different forms of web-based instruction about the physiology and functions of the human heart, in which different chunking strategies were used to deliver the same knowledge. Prior to formally providing them with the treatments, three pilot studies were conducted to (1) identify where static visualisation was not being effective in facilitating achievement in the instructional materials, (2) locate areas in the instruction when animation might be incorporated into the instruction to reduce difficulty, and (3) identify areas in the instruction where chunks might be embedded into the instruction to facilitate the effectiveness of the animated sequences.
As the three treatments were supposed to provide the students with different forms of instruction, Munyofu et al. (2007) needed to ensure that the animations with different chunking strategies should be applied properly in the instruction materials. They conducted the three pilot studies to prepare for the final version of the instructional materials that were to be studied by the participants. This procedure could make the research procedure more scientific and study result more trustable.
However, although a brief description of each pilot study was given, the results were drawn based on an unclear procedure, and the conclusions did not specifically match the purposes of the pilot studies addressed by Munyofu et al. (2007). The first pilot study measured the programmed instruction’s difficulty. Those test items with 0.60 difficulty index were considered to be the items where the animation could be used effectively to increase the effectiveness of the instruction and learner achievement, but the participants of the measure and the procedure and method of measure have not been provided and detailed. This leads readers to doubt whether the pilot studies’ participants were the same as the treatment receivers in the later procedure, which should influence the final measure results. Furthermore, while 24 areas with a 0.60 difficulty index were addressed and 10 animations were to be placed on those frames that address those problematic areas in the content, it was still unknown how the 10 animations could cover the 24 areas. And “area” in the discussion lacked definition. In the second pilot study where computer animation was created to complement the static visualisation in the 24 areas, the results indicate that animation did not make any significant difference on learners’ performance and the difficulty of the items were not reduced. It seems that such conclusion did not match the second purpose of the pilots—“locate areas in the instruction when animation might be incorporated into the instruction to reduce difficulty” because they failed to do so. Based on the finding of Pilot 2, Munyofu et al. (2007) in Pilot 3 worked upon the animated program instruction (Treatment 1) developed in Pilot 2 to developed two additional treatments—animated program instruction with simple visual/text chunked information (Treatment 2) and animated program instruction with complex visual/text chunked information (Treatment 3). Then they hypothesised that Treatment 2 and 3 would result in better learning outcome than Treatment 1 did.
As Treatment 1 did not help improve learners’ performance, Treatment 2 and 3 could be expected to make a significant difference. However, this arrangement generates another question—could Treatment 2 and 3 make a much better result if Treatment 1 did help improve learners’ performance significantly comparing to non-animated treatment? Such a combination of treatments for comparison with animated program instruction taking effect included was not identified nor discussed by Munyofu et al. (2007). Furthermore, as Treatment 2 and 3 were built upon Treatment 1, those learners who received Treatment 2 and 3 were actually given the exposure to additional information that the learners receiving Treatment 1 did not have. It is doubtful whether the difference that Treatment 2 and 3 made should be attributed to the additional information per se or the nature of the chunking strategies that has a special impact on cognition. These two weaknesses of the treatment design would lead to low reliability of the result and conclusion of the whole study.
The instructional modules in the treatments were created following Instructional Consistency-Congruency Matrix (Dwyer 1994). That would help ensure the learning activities in the modules were directly focused on dependent measures. The criterion tests, each consisting of 20 test items, were used in previous research (Dwyer 1978). They were designed to measure different types of learning objectives of the modules, e.g., facts, concepts, rules/principles and comprehension type of information. The total score of the measure was added up from the results of Drawing test, Identification test, Terminology test and Comprehension test. The latter two tests focused on higher order thinking. Kuder-Richardson Formula 20 (KR-20) was applied to measure the internal consistency reliability of the four tests and the Total test. The coefficients were: 0.83 for the Terminology test, 0.81 Identification test, 0.83 Drawing test, 0.77 Comprehension test and 0.92 for the Total test (Dwyer 1978, p. 45). This indicates that the examination that Munyofu et al. (2007) adopted is likely to correlate with alternate forms (a desirable characteristic) as the values of the coefficients were high (Cortina 1993).
The four tests measured the learners from different aspects. For example, the first two tests were for basic knowledge, while the last two were for higher order thinking. This combination of tests could help determine what level of learning had been promoted by the animated program instructions with different chunking strategies. However, the concern from treatment design would lead the measure to be problematic. Because Treatment 2 and 3 were designed to promote higher order thinking by emphasizing specific information on the animation and particularly Treatment 3 “focused more on relationships in the content area rather than single bits of information in a linear presentation” (Munyofu 2007, p. 414), it would be “unfair” for the learners receiving Treatment 1 to participate in the Terminology and Comprehensive tests. It is a common sense that those who receiving more education in an area would perform better in the area than those who do not receive the education. It is evident that Munyofu (2007) would like to match Treatment 2 and 3 with the Terminology and Comprehension tests. This pre-existed preference might have had negative impact on the methodology of the study.
Main Findings and Discussion
It turned out that significant differences existed in the learners’ scores in the Terminology and Comprehension tests among treatments. On the other hand, there was no significant difference in achievement found to exist on the other tests and the Total test. Munyofu et al. (2007) specifically pointed out that Treatment 3 (programmed instruction with complex visual animated text) was significantly more effective than the Treatment 1 on the Terminology and Comprehension tests at the 0.05 level. This result has met the expectation of Munyofu et al. (2007). As a result, they claimed that the use of chunking strategy apparently reduced the cognitive load requirements and enabled the students to successfully interact with and process the information being provided by the animation, and furthermore, when related pieces of information are given in chunks (complex chunking), students are better able to make the necessary connections among the pieces and thereby facilitate higher order learning objectives, but the facility is not evident when single content is given in a simple chunking format.
The significant difference in the learners’ scores in the Terminology and Comprehension tests among treatments does indicate the additional chunks of information lead to the learners’ better performance on higher order learning. However, could this mean that the additional chunks of information in Treatment 2 and 3 really result in the increase of effectiveness of higher order learning? In Munyofu et al. (2007), the two subjects for the comparison are unclear. From the paper, it seems that the two subjects are “programmed instruction with animation (Treatment 1)” and programmed instruction with animated text (Treatment 2 and 3)”. Actually, what Munyofu et al. (2007) were comparing were “nothing” and “chunk of text”. The result of the comparison can only indicate that “chunk of text” is better than “nothing”, but can not guarantee “chunk of text” a more effective approach to promote higher order learning than other strategy. For example, if two people feel hungry and only one of them eats a piece of bread and then feels not that hungry, we can not say “bread” is a kind of more effective food. When Munyofu et al. (2007) mentioned “more effective”, the object of “effective” was unclear. In the claim “Treatment 3…was significantly more effective than the Treatment 1 on the Terminology and Comprehension tests at the 0.05 level”, “more effective” may refer to the scores in the two tests rather than learning per se. Munyofu et al. (2007) also claimed that chunking strategy reduced the cognitive load requirements and enabled the students to successfully interact with and process the information being provided by the animation. However, from the treatments and the results of the measure, there was no evidence indicating the status of the learners’ cognitive load and the behaviour of interacting with and processing the information.
Response to Research Questions and Purpose
Back to the two research questions and purpose, it seems that Munyofu et al. (2007) did not adequately response to them in the result and discussion sections. While the two research questions focus on the effectiveness of animation and instructional strategies of animated instruction, Munyofu et al.’s (2007) discussion on the results was limited on the comparison of Treatment 1 and 3 in terms of their contribution to learners’ higher order learning, but lacked addressing the impact of animation per se on the learning outcome. On the other hand, although the purpose of the study was stated as to examine the instructional effects of different chunking strategies used to complement animated instruction in terms of facilitating achievement of higher order learning objectives, Munyofu et al.’s (2007) discussion on the results did not pay much attention to the difference in instructional effects between Treatment 2 and 3 which actually had different chunking strategies (simple and complex). Although the result did indicate that significant differences existed in students scores in the Terminology and Comprehension tests among treatments, Munyofu et al. (2007) did not specifically compare Treatment 2 and 3. Their strong opinion on the difference between Treatment 1 and 3 did not help achieve the goal of the study.
Although there are weaknesses in Munyofu et al.’s (2007) study, educational multimedia and interactive interface designers and developers can still benefit from the results from which some guidelines can be drawn:
(1) The effects of purely using animation in programmed instruction are limited. From the Pilot Study 2, although animation was created to complement the static visualisation, it did not make any significant difference on learners’ performance. So designers may need to avoid using pure animation.
(2) Animation with relevant chunked information (visuals or text) attached in programmed instruction may help learners achieve higher order learning objectives. As there was significant difference in the learners’ scores in the Terminology and Comprehension tests among treatment, if educators want to specifically promote learners’ understanding of rules, relations and principles, they can apply this approach on instructional material design.
(3) To maximise the benefit of using chunking strategies in animated instruction, the relevant chunked information can be positioned alongside one animation on the same frame. Those learners who received Treatment 3 had the highest scores in the measures. As a result, educational designers can firstly identify the information within the same areas and then try to chunk the information and put them into one relevant animation. This specifically helps learners’ with their higher order learning.
This review analyses the methodology and the result of Munyofu et al.’s (2007) study on the effect of different chunking strategies for animated instruction and highlights the implication for educational multimedia and interactive interface designers and developers. It identifies the main contribution of the study as to combine the two domains of previous research—chunking theory and the use of animation in instructional material design, and to explore the effects of the combination. It is useful to conduct the three pilot studies in the development of the materials so that a valid experimental environment could be created, but the procedure of the pilot studies was not well addressed which generated doubts about the participants, the method for determining the difficulty of the test items, the definition of “area” in treatments development, etc. The concern about the learners’ exposure to additional information in the Treatment 2 and 3 extends from the research design phase to the result and discussion. This weakness has significantly influenced the reliability of the study and led the claims that Munyofu et al. (2007) made to be contrived and farfetched. Finally, although the research questions and purpose were mentioned in the discussion, they were not adequately addressed and discussed. The most important step to achieve the goal of the study was not processed, which is to compare the results of Treatment 2 and 3 because they were the two treatments that had different chunking strategies. The implication for educational designers is drawn from the result of Munyofu et al.’s (2007) study. The guidelines in it are believed to be useful. Further research on the application of the chunking strategies exampled in Treatment 2 and 3 is necessary and encouraged.
Carter, T, Hardy, C A & Hardy, J C 2001, ‘Latin vocabulary acquisition: an experiment using information processing technique of chunking and imagery’ Journal of Instructional Psychology, vol.28, no.4, 225-229.
Chase, W G & Simon, H A 1973, ‘The mind’s eye in chess’, in W G Chase, Visual information processing, Academic Press, New York, 215-281.
Cortina, J M 1993, ‘What is coefficient alpha? An examination of theory and applications’ Journal of Applied Psychology, vol.78, no.1, 98-104.
Cowan, N 2001, ‘The magical number 4 in short-term memory: a reconsideration of mental storage capacity’ Behavioral and Brain Sciences, vol.24, no.1, 87-114.
Dwyer, F 1978, Strategies for improving visualized instruction, Learning Services, State College, PA.
Dwyer, F 1994, ‘One dimension of visual: a paradigm and its implementation’, in D Moore & F Dwyer, Visual literacy: a spectrum of visual learning, Educational Technology Publications, Englewood Cliffs, NJ, 383-401.
Gobet, F 1998, ‘Expert memory: a comparison of four theories’ Cognition, vol.66, no.2, 115–152.
Gobet, F 1998, ‘Memory for the meaningless: how chunks help’. inProceedingsofTwentieth Meeting of the Cognitive Science Society, Mahwah, NJ.
Gobet, F & Simon, H A 1998, ‘Memory’ Expert chess memory: revisiting the chunking hypothesis, vol.6, 225-255.
Gobet, F & Simon, H A 1998, ‘Pattern recognition makes search possible: comments on Holding’ Psychological Research, vol.61, no.3, 204–209.
Munyofu, M, Swain, W J, Ausman, B D, Lin, H, Kidwai, K & Dwyer, F 2007, ‘The effect of different chunking strategies in complementing animated instruction’ Learning, Media and Technology, vol.32, no.4, 407-419.
Newell, A & Simon, H A 1972, Human problem solving, Prentice Hall, Englewood Cliffs, NJ.
Reiber, L P 1990, ‘Animation in computer based instruction’ ETR&D, vol.38, no.1, 77-86.
Simon, H A 1974, ‘ How big is a chunk?’ Science, vol.183, 482-488.