The way information has been retained and transferred into meaningful output has baffled the minds of Cognitive Psychologists for decades. It is in an effort to discover how this becomes possible that various Cognitive Psychologists have developed hypotheses and presented models. Estes, (1975) posited that some learning theories support hypotheses that are based on instruction that leads to learning outcomes that he regard as a two element paradigm. He however, supports a three element paradigm which involves instructions, memory structure, and learning outcome. If there is not a three structure paradigm how might one account for the retention and the processing that must take place if there is no memory structure? But the structure of the memory system is still a source of controversy.
According to the modal model there is not only a memory structure but there are different kinds of memory. But most contemporary researchers assume that there are three types of memory; a sensory memory structure or register, a short term store, and a long term store. There is also support for a two structure model. This is regarded as a two storage system and this is where the emphasis lies. Support for a storage system was highlighted from (Mulner, 1959) research. Her research supports the hypothesis that if the hippocampus was removed it would be difficult for new learning to take place. Eichenbaum, (2000) states that the Hippocampus is seen as critically involved in the rapid encoding of events as associations among stimulus elements and context, in the encoding of episodes as events, and in linking episodes by common features into relational networks that support flexible inferential memory expression. Mulner, (1959) further posited that although items of learning could be held in short term memory, there is no evidence that they were transferred to Long term memory. Atkinson and Shiffrin (1968) supported (Mulner 1959) findings. Zechmeister and Nyberg (1982) posited that information enters the memory system through a sensory register that records information impinging on the sense organs.
The concept of working memory has been introduced as a part of the short term memory (Bradderly and Hitch, 1974; Hastie and Carlston ( 1980). Short term memory takes information as concepts from the sensory register and maintains activated knowledge drawn from long term memory. Long term memory is comprised of semantic long term memory and episodic long term memory. Semantic long term memory stores structural information. This is information that is not dependent upon a particular time or place. Episodic long term memory stores contextually dependent information. That is information about specific events or episodes. Klatzsky (1980) purports that Episodic Long term memory is constantly changing. This is so because as (Conway, Cohen, and Stanhope, 1991; Semb, Ellis, and Aroujo, 1993) stated, although some of what is learned is lost, the amount is not significantly great. Bahrick, (1984); Bahrick, Bahrick, and Wittingler (1975), Bahrick and Hall (1991); Conway, Cohen, and Stanhope (1991) reported retention intervals as long as fifty years. From their research they discovered that persons retained a substantial amount of the Spanish, Algebra, and psychology that they were taught in school. Research carried out by Cane and Willey (1939) and Hovland (1940) supported the hypothesis that persons who are given multiple opportunities for learning had better retention.
But if one is merely interested in assessing what students have learned over a period of time, the focus will be merely on assessing remembering. However, meaningful learning supersedes mere remembering. Bransford, Brown and Cocking (1999); Lambert and Mc Combs (1998) stated that meaningful learning is recognized as an important educational goal. For meaningful learning to take place instruction must go beyond the simple presentation of factual knowledge and that assessment task should require students not just to recall or recognize but they should be able to construct meaning from what is learned. Hence, students should be able to understand what is learned, apply knowledge, analyze, evaluate and use knowledge to create. If the objective of the teacher is to assess the degree to which students have learned some subject matter content and retained it over some period of time, the focus would be on just one class of cognitive process, namely, those associated with remembering.
Mayer (2001) posited that two of the most important educational goals are to promote retention and to promote transfer (which, when it occurs, indicates meaningful learning). Retention is the ability to remember material at some later date in much the same way it was presented during the instruction. Transfer is the ability to use what is learned to solve new problems, answer new questions, or facilitate learning new subject matter (Mayer and Wittrock 1996). In other words retention requires students to remember what is learned, where as transfer requires not only retention but also the application of knowledge to old and new situations (Bradford, Brown, and Cockling, 1999; Detterman and Sternberg, 1993; Heskell, 2001; Mayer, 1995; McKeogh, Lambert, and Marini, 1995; Phye, 1997). Remembering is therefore the sole ingredient of retention. On the other hand transfer involves remembering, understanding, applying, analyzing, evaluating, and creating.
If the retention of information is the focus then the main focus of the cognitive process is remembering. However, if the focus is transfer it shifts to the other five cognitive processes; understanding, applying, analyzing, evaluating, and creating. Mayer, (2001) stated that students understand when they can associate old knowledge with new ones. In other words if students are able to make connections as they formulate new concepts ideas, and create new schematic formulations, they have demonstrated that they have learned.
At the application stage (Mayer, 2001) students are able to use what is learned to execute procedures. In other words they are able to carry out tasks based on that knowledge. For example if instruction is based on how to bake a cake, the student should be able to bake the cake. Therefore the bass for application are remembering and understanding.
As (Mayer, 2001) continues to review Bloom’s Taxonomy, he states that to analyze involves breaking ideas, concepts, and schemas into their component parts and demonstrate how the parts are related to each other and to the whole structure. The bases for analysis are; remembering, understanding and applying.
In order to evaluate students must be able to remember, understand, apply, and analyze. Evaluation is the ability to make judgment that is based on a given criteria. Mayer (2001) states that the standards may either be quantitative of qualitative. Evaluation is further described as judgment about internal consistency and critique which is external consistency. At this level of transfer students should be able to detect inconsistencies between an operation and some external criteria.
The final stage is that at which students are able to synthesize aspects of what is learned to produce a whole, a concept or a schema or something that is tangible such as a machine or a work of art. It is that which enables one to develop hypotheses with a view to solving problems. Therefore in order to create, one must be able to remember, understand, apply, analyze, and evaluate.
Chandler and Sewell (1991); Mayer (2001, 2005); Mayer and Moreno (2003); Sweller (1999, 2005) posited that meaningful learning requires learners to engage in appropriate cognitive processing during learning. These cognitive processes include attending to relevant information, mentally organizing the selected information into a coherent structure, and integrating the incoming information with existing knowledge. This is regarded as the triarchic theory of cognitive load and it postulates three kinds of cognitive processing during learning.
The first is extraneous processing in which the learner engages in cognitive processing that is not related to the instructional goal or in some instances there are no instructional goals. It is just learning for learning sake. The other is essential or intrinsic processing in which the learner mentally represents the material and which is determined by the inherent complexity of the material. The third is the generative or germaine processing of material such as organizing and integrating the selected material with the desire to understand the lesson.
Generative processing is similar to transfer of learning that produces the ability to create. According to (Mayer ,2005; Mayer and Moreso, 2003; Sweller, 2005) line texts can be converted into a graphic organizer through selected relevant text and organized into a coherent structure. When the scaffolding of graphic organizers is provided, learners are less likely to waste precious cognitive capacity on extraneous processing which thereby reduces cognitive load and frees up capacity for essential and generative processing.
De Jong (2005); Kirsner, Sweller, and Clark (2006); Klahr and Nigane, (2004); Lillard, 2005); Mayer, (2003, 2004) forwarded that activity theory is based on the idea that deep learning occurs when students are encouraged to engage in productive learning activities. Constructing a graphic organizer can be considered a productive learning activity because the learner must engage in an activity that is related to the instructional objective – selecting relevant ideas from the text and organizing them in a coherent structure. Activity theory purports that learner generated graphic organizers do. However three experimental researches carried out by Stull and Mayer (2007) proved the opposite. Below is the full text of experiment 1. The summaries of the other two experiments along with the summary of experiment 1 are included in the appendix.
Experiment 1 (Highest Complexity)
The purpose of Experiment 1 was to test whether students better understand a scientific passage when they are asked to generate graphic organizers (following pretraining in how to generate hierarchies, lists, flowcharts, and matrices) in spaces in the margin or when the passage contains author-provided graphic organizers. In Experiment 1, participants read a 1,133-word passage about a topic in biology that was augmented by 27 author-provided graphic organizers (author-provided group), participants constructed their own graphic organizers from scratch (learner-generated group), or participants did not receive or construct graphic organizers (control group). Our primary focus is on comparing the author-provided group and the learner-generated group on measures of understanding.
Participants and design. The participants were 156 college students recruited from the psychology participant pool at the University of California, Santa Barbara. The study was based on a
between-subjects design, with three levels of graphic organizer use (author provided, learner generated, and control) as the single factor. Fifty-one students served in the author-provided group, 51 students served in the learner-generated group, and 54 students served in the control group. The mean age was 19.4 years (SD =1.5), the percentage of men was 29.5%, and the mean SAT score was 1184.5 (SD =161.4).
Materials. The paper materials consisted of a participant questionnaire, two pretraining documents (author-provided and learner generated versions), three reading passages (i.e., author-provided, learner-generated, and control versions), six short-answer test sheets (one retention and five transfer questions), and eight sentence-completion (all retention questions) test sheets, each printed on an 8.5 x 11 in. (21.25 x 27.5 cm) sheet of paper. The participant questionnaire solicited basic demographic information, including the participant’s age, sex, and SAT scores.
The full version of the pretraining document was developed for the author-provided and learner-generated group, and the control version of the pretraining document was developed for the control group. The full version of the pretraining document consisted of a two-page document printed on facing pages. The left page described and illustrated four types of graphic organizer (concept list,
concept hierarchy, concept flowchart, and compare-and-contrast matrix). The right page contained a four-paragraph reading passage laid out in a two-column design. The left column contained the biology text, and the right column contained each of the four types of graphic organizer, horizontally aligned with the matching type description on the left page. The reading material was extracted from a popular college-level general biology textbook, then edited slightly to meet the desired page format as well as to remove external references, but without altering the book like style or the author’s voice in the source material. The passage described three biologically important polysaccharide molecules-starch, glycogen, and cellulose. The control version of the pretraining document contained the identical biology text from the left column of the right page but not the left page describing and illustrating the graphic organizers or the integrated graphic organizers from the right column of the right page. These modifications were made without alteration to the text layout, so the right column was empty.
The control version of the reading passage consisted of six pages containing 1,133 words organized into 12 paragraphs, with three figures containing four black-and-white photographs. The three figures with four photographs were required to augment the written descriptions and to maintain the book like character of the material. The page layout matched the two-column design used in the pretraining document. The material was extracted from the same textbook used for the pretraining to maintain a consistent voice and character between the readings. The material described eight reproductive barriers between species (temporal, habitat, gametic, behavioral, mechanical, hybrid inviability, hybrid sterility,
and hybrid breakdown) and was divided into two barrier groups (prezygotic and postzygotic). The author-provided version used the identical text, figures, and illustrations but included 27 graphic organizers, each placed in the margin near the corresponding text. One concept hierarchy graphic organizer augmented the introductory paragraph. Each of the eight reproductive barriers was
described by a single paragraph and augmented with three graphic organizers (one hierarchy, one list, and one flowchart). A hierarchy graphic organizer and a matrix augmented the conclusion. The learner-generated version was identical to the author-provided version except that all graphic organizers were removed, which left space for learners to construct their own graphic organizers. The version used by the control group was identical to that used by the learner-generated group. Example pages of the three versions are shown in Figure 1.
The six short-answer test sheets and the eight sentence -completion sheets each had a question printed at the top of the page, and at the bottom of each sheet were printed the following instructions: “Please keep working until you are asked to stop. Do not go back to any previous questions.” The eight sentence completion questions (eight retention questions) are presented at the top of the Appendix, and the six short-answer test questions (one retention and five transfer questions) are presented at the bottom of the Appendix.
Procedure. Participants were tested in groups of 1 to 5 and randomly assigned to one of the three graphic organizer treatment groups. Each participant was seated in an individual cubicle. First, participants were asked to read and sign an informed consent form, followed by a participant questionnaire to be completed at their own rate. Then they were given oral instructions to carefully read the pretraining document (with control participants receiving the control version and all other participants receiving the full version). In the author-provided group, participants were instructed to compare the descriptions and illustrations of the four types of graphic organizer with the illustrated example on the facing page. In the learner-generated group, participants were instructed to compare the descriptions and illustrations of the four types of graphic organizers with the illustrated example but also told that they would be asked to construct their own graphic organizers. The training lasted approximately 5 min and was intended to familiarize learners with each of four types of graphic organizers-list, hierarchy, flowchart, and matrix-by providing definitions and examples. In the control group, participants were only asked to read the passage but were not provided with or informed about graphic organizers. Participants were asked to stop reading at the end of 5 min, which proved to be more than adequate for the task. Next, participants were given further oral instructions that described the reading assignment, which they could complete at their own pace. These instructions asked the participants to read the material carefully and to be aware that questions about the reading would follow. In the learner-generated group, participants were told that they could construct their own graphic organizers in the margins of the page as they read the material and were informed that this might help them understand the material. Participants in the author-provided group and the control group were not instructed to generate graphic organizers during reading and did not generate any graphic organizers. The time to complete the reading was recorded for each individual. When all participants were finished reading the material, the stack of six short-answer test sheets was passed out. Participants were given oral instructions to work only on the top sheet, to keep working until they were asked to stop, and not to turn to the next sheet until asked to do so. Participants were carefully monitored for compliance. At the end of 3 min, participants were asked to immediately stop working on the current question, turn that sheet face down onto a finished stack, and begin the next sheet. After the last short-answer question, participants were given oral instructions that described the sentence-completion questions, which followed the same procedure as the short-answer questions except that participants were given 1 min to answer each question. The short-answer and sentence-completion questions were presented in the order listed in the Appendix. After the last sentence-completion question, all material was collected, and the participants were debriefed and excused.
Results and Discussion
Scoring. We computed the retention score for each participant by tallying the score for the first short-answer question (worth a maximum of 4 points) and the score for each of the eight sentence completion questions (worth a maximum of 16 points). On the short-answer retention question, participants received 1 point for mentioning each of four concepts: (a) prezygotic barrier with (b) before fertilization and (c) postzygotic barrier with (d) after fertilization. On each sentence-completion question (worth a maximum of 2 points each), the participant received 1 point for writing the correctly spelled term for the appropriate reproductive barrier and 1 point for the correct prefix for the barrier subgroup- prezygotic or postzygotic. The correct answers to the eight sentence-completion questions listed in the Appendix are (a) temporal and pre, (b) habitat and pre, (c) gametic and pre, (d)
behavioral and pre, (e) mechanical and pre, (f) hybrid viability and post, (g) hybrid sterility and post, and (h) hybrid breakdown and post. Partial terms (e.g., hybrid or sterility instead of hybrid sterility),
parallel concepts (e.g., time instead of temporal or geographic instead of habitat), and incorrect spellings (e.g., pro instead of pre) were not acceptable answers. Partial credit was awarded if participants provided only one of the two correct terms. Each participant could earn a maximum of 16 points on the eight sentence-completion questions and 4 points on the short-answer retention question, for a total possible of 20 points on the retention score.
We computed the transfer score for each participant by tallying the individual scores on each of the five short-answer transfer questions-short-answer questions 2 through 6 are listed in the Appendix. We scored each question by counting the unique concepts presented in the reading that were used appropriately by the participant to address each question. Acceptable concepts included the 10 specific reproductive barrier concepts: (a) prezygotic, (b) postzygotic, (c) temporal, (d) habitat, (e) gametic, (f) behavioral,(g) mechanical, (h) hybrid inviability, (i) hybrid sterility, and (j)hybrid breakdown. In addition, two general concepts were also counted: (a) crossing organisms to test whether reproduction was possible or recognizing that two species might have crossed to form a hybrid, and (b) mentioning that reproductive barriers maybe relevant to the explanation. Participants were allowed to describe the concepts with partial terms and parallel concepts, and misspelled terms were not counted as wrong. One point was awarded for each of the 12 concepts, for a maximum of 12 points per question. A second person scored all material. The interrater reliability measure was significantly correlated between these two scores (r =.826, p < .001). Discrepancies in the scores between these two scorers were individually evaluated in a blind, third review, which was used to determine the final score.
Data analysis. Data were analyzed with one-way analyses of variance comparing the performance of the three treatment groups on each of the dependent measures-transfer score, retention score, and study time. Our major focus was on comparing the author-provided and learner-generated groups, so for each dependent measure we conducted planned contrasts on the mean scores of these groups and computed the corresponding effect size on the basis of Cohen’s d (Cohen, 1988).2 Table 2 lists the mean and standard deviation of each of the three treatment groups on each of the three dependent measures.
Do readers who generate their own graphic organizers while reading a scientific passage learn better than readers who are given author-provided graphic organizers? The top left portion of Table 2 summarizes the mean transfer scores of the three groups in Experiment 1. There was not a significant effect of treatment on transfer scores, F(2, 153) = 1.32, MSE = 10.15, and the author- provided group did not differ significantly from the learner -generated group, t(153) = 1.30, d = 0.24. There is no evidence that constructing graphic organizers or even studying author- provided graphic organizers results in deeper learning.
The top middle portion of Table 2 summarizes the mean retention scores of the three groups in Experiment 1. There was not a significant effect of treatment on retention scores, F(2, 153) = 0.210, MSE = 21.38, and the author-provided group did not differ significantly from the learner-generated group, t(153) =0.56, d = 0.11. There is no evidence that constructing graphic organizers or even studying author-provided graphic organizers results in better memory for the presented material.
The top right portion of Table 2 summarizes the mean study times of the three groups in Experiment 1. There was a significant difference among the groups in mean study time, F(2, 153) = 82.86, MSE = 9.99, p < .001, and the author-provided group required significantly less study time than did the learner- generated group, t(153) =8.97, p < .001, d = 1.51. Although constructing graphic organizers did not result in better retention or transfer performance, it did require considerably more study time.
Although the main focus of this research was on comparing the test performance of the author-provided and learner-generated groups, the types and number of graphic organizers produced by the learner-generated group were also examined. The author-provided group received 27 graphic organizers
containing 506 words, whereas the learner-generate group produced a mean of 5.1 graphic organizers containing a mean of 84.2 words. The mean number of graphic organizers produced fell from 2.0 on page 1 to 0.5 on page 5; the mean number of words produced fell from 34.8 on page 1 to 10.0 on page 5. Although the number of graphic organizers produced in the learner-generated group was lower than that given to the author-provided group, all but 2 of the 51 participants in the learner-generated group attempted to construct graphic organizers. Exclusion of these 2 participants from the analysis did not alter the statistical results.
These results are contrary to the prediction that graphic organizers facilitate learning. Furthermore, there is no evidence that generating graphic organizers resulted in better learning than simply
viewing them on the page, although there is evidence that more study time was required when students generated their own graphic organizers. The open-ended nature of the learner-generated treatment might have been too demanding and confusing for the learners. Although a majority of participants in the learner-generated group attempted to construct graphic organizers, these graphic organizers varied greatly in form and quality. Participants might have been overwhelmed by the requirement to both select and implement appropriate graphic organizers, both of which might have contributed to extraneous cognitive load. For participants in the author-provided group, the margins of the pages were
densely crowded with graphic organizers, which were potentially confusing to interpret as participants attempted to compare the concepts in the text with the appropriate graphic organizer. This might also have contributed to additional extraneous cognitive load. To address these issues, we reduced the complexity of the treatment in Experiments 2 and 3 by offering fewer graphic organizers to both groups and partially completed graphic organizer templates to the learner-generated group.
Knowledge of how memory works is important to teachers and Cognitive Psychologists as they seek to discover ways and means to enhance learning. But it is possible that the brain can become so inundated with ideas that much of what comes to it simply decay.
Bahrick, (1979) stated that much of what is learned in classrooms is lost soon after final examination. Higbee (1977) posited that people forget what they learned in school (usually within a short time after an examination). Never-Benjamin (1990) forwarded that if this is the case it is very serious. Neisser (1982) expressed that there is a difficulty in finding studies that support retention of academic instruction. But Nesser (1982) might not have been searching wide enough. In fact the literature that is available is replete with the suggestion that much work has been carried out. Wert, (19370 suggested that studies in the area of zoology, biology, and psychology, found retention from a few months to three years. It has also been put forward that (keller, 1968) personalized system of instruction and (Blooms, 1968)learning for mastery often include a measure of retention.
Studies by (Gaskey and Gates, 1985; kulik, Kulik, and Bangert- Drowns, 1990) posited that students in all conditions retained much of what was taught. Conway, Cohen, and Stanhope, (1991); Semb, Ellis, and Aranjo (1993) stated that although forgetting does occur, the amount loss is not as great as expected by popular belief. Farr’s ( 1987) opinion is that the most important variable in long term memory retention is the degree of original learning. Evidence from laboratory studies shows that increasing the number of learning trials enhances retention. Research has also proven that retention often depends on the instructional strategy that is used. A comparison of studies by ( Austin and gilbert, 1973 ; Breland and Smith 1974, Cooper and Greiner, 1971; Corey and Mc Michael, 1974; Glasnapp et al. 1978, Lu, M., 1976; Lu, P. 1976; Schwartz, 1981; Semb et al., 1993; Sharples et al., 1976) advanced that course objectives, content , length, and tests were the same for all the groups in the studies. The only difference was the instructional delivery strategy and mastery criteria. All the studies show that mastery conditions produce superior academic performance at the end of the course.
Mckenzie and White (1982) observed high levels of retention for students actively involved in learning. In their study three groups of students learned geographical facts and skills. One group was given a learning program which includes pictures, slides, worded examples, sample test items, indications of relevance of information to subsequent application, and transfer of verbal proportions to maps, diagrams and slides. The remaining two groups were given learning program and field excursion. Treatment groups were formed from eight and ninth grade classes from two different schools. The classes were not ability tracked, and class assignments to treatment groups were random. Students in the excursion classes were assigned to either a traditional excursion or a processing excursion. For the traditional excursion students were given an explanatory field guide designed to reinforce the learning program