Jos Beishuizen

Studying textbooks: Effects of learning styles, study task, and instruction

Abstract This study was aimed at clarifying the influence of instructional support (focusing on metacognitive or cognitive levels of task accomplishment) and task constraints (exam preparation or searching for a particular text unit) on the way students with different learning styles (deep processing versus surface processing) completed a study task in a hypertext reading environment. We expected deep processing students to benefit from metacognitive support and surface processing students to learn from cognitive support. In Experiment 1 these expectations were partly confirmed. Strategic hints (metacognitive level) had little effect, apparently because the students did not possess any domain knowledge at the outset. An advance organizer (cognitive level) proved harmful to deep processing students but beneficial to surface processing students. Experiment 2 showed that students who combined self-regulation with deep processing and students who combined external regulation with surface processing outperformed students with complimentary combinations of regulation style and processing style. It is argued that selection skills may underlie both regulation and processing styles for studying expository texts.

Study strategies in a computer assisted study environment

Abstract A Computer Assisted Study Environment (CASE) was developed as a tool for diagnosing study problems, to be used together with other sources of information, such as learning style questionnaires and clinical interviews. Forty-one students were observed during a 1 h period of studying a text book chapter in CASE. The stages of orientation, planning, and execution were clearly separated by dividing the 1 h study session into three periods. Students could spend an unlimited part of the hour on orientation (first period). Then, not included within the hour, a plan for the task had to be made (second period), after which the remaining time could be devoted to execution of the plan or to any other form of study (third period). We administered a learning style questionnaire, measured reading speed and pretested the students on prior knowledge of the content of the study task. These data were correlated with product and process indicators collected in CASE in order to find out whether various sources of information about learning styles and study strategies provided converging evidence about potential causes of study problems. Process and product indicators of study strategies in CASE revealed differences between deep and surface learning students in orientation and planning activities. However, their actual study behaviour did not vary according to learning style. So, the differences between surface and deep learning become more apparent when we look at study activities before and after actual reading and processing information. As far as learning outcomes are concerned, students with a deep learning style obtained better results than students with a surface learning style, even for factual knowledge. Deep learning students knew more about the subject of the diagnostic study task, and developed a higher reading speed. Both student characteristics significantly determined learning outcomes.

Content effects in self-directed inductive learning

Abstract A study is presented into the effect of familiar and abstract task content on self-directed inductive learning. Two groups of learners performed a self-directed inductive learning task in which they conducted experiments to discover the model describing the relations between five independent variables and a dependent variable. One group performed a task of familiar content, the other performed an isomorphic abstract task. During learning, indicators of the inductive learning process were gathered by means of think-aloud protocols and asking standardized questions to the learners. As expected, learning outcome was higher in the group performing the concrete task. It appeared that the number of hypotheses, research plans and inferences stated were predictive for successful learning outcome, especially in the abstract task. The content effect was explained in terms of a narrowing down of the number of possible hypotheses to be considered in the concrete task, compared to the abstract task.

Search and Study Strategies in Hypertext.

Learning styles are usually assessed by asking students what they do while studying. In our study, we tried to relate what students said they do to what they actually did during two different study tasks. Three questions were addressed. First, we wanted to know if students with different learning styles use different study strategies. Second, we were interested in the role of regulation in strategy use. Finally, we wanted to know whether strategy use was consistent over two different reading tasks (a search task and an exam-preparation task). To measure learning styles we used the Inventory of Learning Styles developed by Vermunt and Van Rijswijk (1987). Based on the cognitive processing scales, students were identified as deep processors or surface processors. By means of a computer-supported reading environment, their actual study behavior on two different reading tasks was recorded. The results show that in a study task demanding careful searching for information, learning styles and directly observed study strategies are clearly related. Deep processors use a global point of view to navigate through the text. Surface processors more often take decisions at a local level. However, these effects were not found during a traditional exam-preparation task. As expected, surface processors were more consistent in their strategy use on both tasks than deep processors.

Asking questions during self-directed inductive learning: effects on learning outcome and learning processes

Asking learners standardized questions during performance of a self-directed inductive learning task might be a useful way to complement think aloud protocol data. However, asking questions might also scaffold the learning process and thus influence the exact processes one wants to study. In the study described in this paper two groups of learners performed a computerized self-directed inductive learning task in which they conducted experiments to discover the relations between five independent variables and one dependent variable. In one condition, the learners thought aloud, in the other the learners were asked additional standardized questions pertaining to specific reasoning steps during learning. Measures of learning outcome and learning processes were collected. It appeared that the questions did not influence learning outcome. With respect to learning processes no differences were found, except that learners in the no questioning condition more often repeated experiments. It was concluded that the questions do not seem to threaten the validity of research findings.

Using Hypertext for Studying and Information Search.

A text in hypertext format is a database of text units without any pre-defined order. Concepts maps, text links, and other navigation tools enable the user to explore the database. This study explored the conditions for making hypertext a better study environment than a traditional linear text. Two tasks were investigated: an open exam preparation task and a closed search task. Within the context of the latter task the influence of the students learning style was traced. For open study tasks, like preparing for an examination, hypertext does not provide clear advantages over a linear text with a table of contents and an index. The first experiment showed that in the hypertext condition more time was spent on actually studying text units explaining important topics. However, there was no increase in text comprehension as compared with a linear text condition. The potential advantages of hypertext may be better utilized in closed search tasks, in which an answer to a particular question has to be found. Because hypertext puts heavier cognitive demands on the student, the quality of the learning style of the student is crucial to success. In the second experiment, we found that both deep processing students and surface processing students were able to find a requested text unit in a hypertext unit in a hypertext database, provided that their regulation style matches their processing style. That is, deep processing students should act on the basis of internal control, whereas surface processing should seek external guidance. The availability of local navigation facilities (like text links or facilities for full text search) contributes to the usability of hypertext, in particular for those users who prefer a surface processing style. However, because surface processing students are vulnerable to losing track, they need external guidance to support their search attempts.

Domain expertise and knowledge acquisition from “nonlinear” expository text

Abstract This article examines some important theoretical issues with regard to how nonlinear text structures of the kind embodied in hypertext might affect learning. Empirical results showing that well-structured nonlinear texts are as effective as conventional linear ones for learning factual information and that they can enhance inferential reasoning are presented and discussed. It is suggested that using a graphic device called a “conceptual map” to represent the relationships between text concepts can help the reader form a coherent view of the textual information and constrain nonlinear reading to useful juxtapositions of text units.

No Correlation between Inferencing Causal Relations and Text Comprehension

Abstract In this study, the causal network model of Trabasso and Van den Broek was used to create short narrative texts and to analyse inferencing processes induced from verbal protocols collected during reading and thinking aloud tasks. Inferential skills were taught of inferencing by offering 10- and 11-year-old students a training programme in which both local and global coherence were topics of explanation and practice. The programme enhanced inferential skills but did not increase performance on a Dutch standardised text comprehension test. It was argued that the test format (answering multiple choice questions about an available text) may have been responsible for this lack of transfer. Therefore, a second experiment was conducted in which correlations between performance on reading and thinking aloud tests and comprehension tests were analysed with both different texts and identical texts. Again, no correlations showed up between reading and thinking aloud performance and comprehension scores on the closed comprehension test with the text at hand. Several tentative explanations were offered. The reading and thinking aloud test may have drawn substantially on verbal fluency. Apart from that, the testing format of the text comprehension test was seriously questioned.

Adaptive tutoring of arithmetic skills

Based on the assumptions that (a) addition and subtraction are rule-based cognitive skills, and (b) that the acquisition of cognitive skills is a result of a developmental process in which stages can be distinguished, a “genetic model” was proposed in which emergence of addition and subtraction skills in the domain of the natural numbers 20–100 is described in six stages. The stages were chosen on theoretical and empirical grounds (several studies on addition and subtraction skills conducted at the State University of Leiden). In a longitudinal experiment, 16 second grade elementary school students were followed during 11 months. A diagnostic computer program was used to collect data on their addition and subtraction performance. The strategies appeared to be employed fairly consistently within a stage of development. The proposed genetic model accounted for most of the performance levels displayed by the students. The conclusion seems warranted that instruction should be adapted to the student’s actual stage of development.