Paul W. O. Hoskin
Department of Geological Sciences

Central Washington University, Ellensburg, WA 98926, U.S.A.
E-mail: hoskin@geology.cwu.edu

A classroom-based experiment indicates that direct instruction in the definition and functioning of metacognition can affect increased student achievement in geoscience education. The cognitive framework for the experiment was learning style preferences as defined in the Felder-Silverman model. This helped students understand their learning preferences. Classroom-based discussion and set home-work activities helped students operate executive control over their learning preferences and cognitive processes. This was done in a reflective manner and was adapted on a task-to-task basis in order to maximize learning and achievement. Information on student learning style preferences as well as instructor teaching style preferences is the first-step in creating metacognitive control of learning practices and making a habit of maximizing learning experiences. The interpretation of the results of this case study is that instruction and exercise in metacognition can affect increased student achievement in geoscience education by up to four grade increments.

Education in geoscience is typically a mixture of geology and elements of chemistry, physics and biology. The approach to learning that a student develops in these latter sciences may not be the most effective in a discipline like geophysics or mineralogy, where there is synthesis between "packages" of knowledge from different sciences. As well as learning-style differences between disciplines, a student’s learning may be enhanced or hindered depending upon the teaching-style of individual instructors (Felder, 1995). Data to support such an assertion are rare, yet an increasing awareness of the learning preferences of students and teaching-style of college-level instructors (e.g., Srogi & Baloche, 1997) indicates a growing acceptance amongst instructors of the influence of their teaching-style upon student achievement levels. Instructors can help students help themselves by teaching geoscience students to continuously review their learning progress so that they can adjust, when needed, their learning process to continuously maximize learning and achievement. This process requires the student to have knowledge of their learning-style preferences and how to exercise effective executive control over their learning (cognitive) processes. This paper defines metacognition and its relationship to theories of learning and learning-style preferences. A classroom-based experiment in instruction in metacognition which is interpreted to have resulted in increased student achievement is described.

Description: "Cognition" refers to any process which allows one to know and be aware and encompasses processes by which we acquire, store, interpret, understand, and use information in the external environment as well as information stored internally (Rebok, 1987). These processes are different from those in our emotional, volitional and physical domains and occur in what is termed the "cognitive domain." Formal descriptions of the cognitive domain familiar to K–12 teachers include those such as Bloom’s Taxonomy (Bloom & Krathwohl, 1956), Gardner’s model of multiple intelligences (Gardner, 1983) and de Bono’s six thinking hats (de Bono, 1992). These descriptions have in common different observable cognitive actions at different levels. Bloom’s lowest level is simple recall of knowledge with higher levels including synthesis and evaluation. Cognition has to do with knowledge and thinking.

In the most simple terms "metacognition" is thinking about ones own thoughts—it is being aware of ones own cognitive processes, strengths and preferences (learning style), and being able to exercise control over these. Metacognitive thoughts are deliberate, planful, intentional, goal-directed, and future-oriented mental behaviours that can be used to accomplish cognitive tasks (Flavell, 1971; 1976). A distinction is made between cognitive processes (cognition) and executive functioning and control (metacognition); this distinction clarifies the two basic elements of learning, the "tools" and the "know how". Simple thoughts such as "I will learn better when reading this textbook chapter if I take notes or read aloud," illustrate metacognitive thinking and action.

Vygostky suggested that private speech is a precursor to self-regulatory behaviour. Private speech, defined as externalized thought, was found to commonly be correcting or reinforcing in nature in academically advanced children (Nelson & Narens, 1994). The apparent role of metacognition in cognitive and intelligence development is supported by Piaget’s theory of cognitive development. The bridge between Piaget’s Concrete operational period (7–11 years) and Formal operational period (11–15 years) is development of the ability to "metathink," that is, to think about thinking itself rather than about objects of thinking (Flavell, 1963). Piaget’s constructs of "assimilation" and "accommodation" include the concept of a self-regulating intelligence. This suggestion of metacognition as a bridge between developmental stages has been demonstrated in young children where once a certain level of cognition had been reached, metacognitive thinking emerged allowing the children to develop cognitive skills to a higher level (Crowley et al. 1997).

In information-processing theories of learning the storage and use of knowledge is likened to that of computers. Computers retrieve and use knowledge via algorithms (codes). Young learners have been shown to have small, simple algorithms whilst older learner’s algorithms include more information and loops (Simon, 1979; Anzai & Simon, 1979). These more sophisticated algorithms have the capacity for self-modification or self-regulation. Metacognition applied in these systems adds more loops and pathways to the algorithm, increasing problem-solving capacity and goal-reaching potential.

Metacognition has a role in behaviouristic learning theories also. According to Bandura’s social learning theory one can learn to symbolically learn behaviour through cognitive organization, rather than through experience (Bandura, 1977). Moreover, observation of self and self-regulation can be used to help reproduce desirable behaviour and outcomes through an iterative process (Zimmerman, 1983).

The traditional method of information transmission in college-level science courses is through the lecture. Students learn passively in lectures. Little or no thought is given by most students on how they might improve their learning in this environment. Likewise, little attempt is made by many instructors to minimize learning obstacles for students in their classes. Learning obstacles for students are some times derived from the teaching-style of the instructor. This leads to student underachievement in a particular course relative to their potential. The first-step in minimizing such obstacles is for instructors to understand their own learning preferences as these will be reflected in teaching style. Understanding of the learning style preferences of their students is also necessary for instructors who wish to see maximum achievement and success for students under their instruction. An understanding of teaching and learning preferences allows both the instructor and student to compensate for strong preferences that might otherwise simply form a learning obstacle. Compensation will usually comprise new, specific and intentional teaching strategies and learning practices, respectively (Terry, 2002). This compensation is metacognitive. Compensation is necessary as previous studies have shown that student achievement in math and science courses can be skewed toward those who learn sequentially (Drysdale et al. 2001; Ross et al. 2001) although these students form only a subset of all math and science majors.

Generalizations about teaching and learning preferences of geoscience instructors and students are given below. Descriptions are made within the definitions of learning style preferences of the Felder-Silverman model (Table 1; Felder & Silverman, 1988; Felder, 1993). Such generalizations are not correct for everyone, yet are based upon the author’s observations from four western countries, literature descriptions, and specific feedback within the context of the Felder-Silverman model from the author’s colleagues.

Students within the first two years of college study tend to exhibit a strong preference for active and visual learning. At this level, students are still strongly influenced by their previous high-school environment where teachers better understand and cater for strong active and visual learning preferences. These students prefer to see and do, and physically interact with new knowledge. More senior students have learnt to adapt, to some degree, and accommodate the differences between their own learning preferences and the style of lecture delivery. In most cases, this does not represent metacognition but survival adaptation. Potentially, the most successful students are those that have learning preferences most similar to that of their professors, either naturally or by preference adaptation.

The learning preference of an instructor will be reflected in how they structure, organize, and present their lectures. This comment from a sophomore student in 2001, "I have a teacher this semester who [only] speaks during his lecture, he does not use any visual aids…", indicates that the teacher has a strong preference as a verbal learner (in contrast to most of his junior students). This seems to be the learning preference for many academics and is no surprise given the strong verbal environment in colleges and universities in which they themselves have learnt and spent most of their lives. An active research career and adherence to scientific method tends to train people to exhibit preferences for reflective and sequential learning. These learning preferences of instructors may manifest in lectures as boring verbal monologues that pose a learning barrier to more active, global and visual students.

Hypothesis and methodology. The above generalities of learning style preferences were tested in a sophomore (second-year of a B.S. degree) mineralogy class at the University of Notre Dame, Indiana. The class was of typical size for this college and course, comprising four students. Given the

Table 1 Learning style preferences of the Felder-Silverman model* and descriptions.

Preference Description

* Felder and Silverman (1988); Felder (1993). Note: the model contains four continuums: active—reflective, sensing—intuitive, visual—verbal, sequential—global.

Limited student numbers, the investigation presented here is a case study and conclusions based on data gathered here may not be directly extrapolated to another context. Moreover, in-context (i.e., at the same university, in the same class in a subsequent semester) follow-up and repeat studies which would have provided a broader database were precluded as the author moved to another university. Despite such limitations, case studies do provide valid results within their context that are often supported by broader investigations (Bassey, 1999; Klein, 2000; Rodgers & Jensen, 2001).

Students were asked to complete an on-line questionnaire to assess their preferences within the Felder-Silverman model. Additionally, they were asked to provide more detailed information by way of a written questionnaire. This first-step in an experiment to test the hypothesis that direct instruction in metacognition leads to increased achievement (Carr et al. 1994; Scruggs, 1985) was performed at the half-way point of the course. Prior to this point in time no instruction in learning style preferences or metacognition was given. To ensure objective and unbiased grade assignment, all tests (formative and summative assessments), assignments, and examinations in the course were graded by a single person using grade keys which were written at the same time the assessment piece itself was written.

Students completed the on-line and written questionnaires in their own time and brought these with them to the next lecture. The lecture was preceded with a discussion of learning style preferences, the student’s personal results, and strategies for compensating for strong learning preferences (Felder & Soloman, 2000). Students were required to reflect on their results and the class discussion and write a half-page summary including personal strategies that they would take to improve their learning in mineralogy. Students were encouraged to continuously reflect and make changes in their cognitive functioning if required for greater success, on a task-to-task basis.

The results from the Felder-Silverman questionnaire were returned to the instructor. In addition to students making changes to their learning strategies, the instructor made changes to the teaching style to better accommodate the preferences of the students where these differed significantly from the instructor. This happened not only in the lecture, but the type of set home-work activity was varied to cover more than just textbook readings and problems.

Students exhibited a range of learning style preferences (Figure 1). A strong preference for a particular learning style is indicated by the location of a marker-box close to the end of the continuum in Figure 1. Two students (A and B) and the instructor had only moderate to no-preference positions, while two students (C and D) showed strong preferences for visual and active learning, respectively. Such strong preferences can isolate a student from the learning environment if the teaching style does not cater for that preference. Neither of these strong preferences were observed for the instructor. The top students in the class had no strong preference for any learning style and are potentially more able to cope in any learning environment. These students also have over-all

Figure 1 Comparison of instructor and student learning style preferences in the Felder-Silverman model. Final course grades for each student are given on the right-side.

Learning style patterns that do not differ significantly from that of the instructor, where in contrast, the lowest performing student (student D) has a negatively-sloped pattern that contrasts with the positively-sloped pattern of the instructor.

Prior to any formal knowledge of their learning style preferences and prior to direct instruction in metacognition, students indicated that they had made changes to their learning styles since arriving at college. One student stated: "College has caused me to change my study practices. I need to do more repetition, spend longer hours, do more independent work, and begin studying longer in advance [of an examination]." Another stated that "…due to poor lecture notes [in other courses] and weekly quizzes, I’ve started taking my own notes from the book, and have a weekly group discussion with a small group of friends." These statements represent proto-metacognitive action where conscious decisions have been made to expand the "tool-kit" of learning (more hours, more repetition, more notes), but not the "know-how" which represents true executive control of cognition.

In response to instruction in metacognition students adapted their learning practices where they deemed it was necessary. This was monitored by informal discussions in laboratories. Some students began drawing flow-diagrams and mind-maps that linked together packages of knowledge by common threads and provided a global and visual overview not always obvious from daily and "isolated" lectures. Other students started using CD-ROM libraries and texts and illustrating their lecture notes as they wrote them in an attempt to link visual information with verbal information. These students came to anticipate that in the verbal-dominated lecture they would need to link aural information with visual information if they were to retain or learn new material. They learnt that this was not necessary for every lecture in all their courses, but important at different times of the day and in some lectures.

Prior to instruction in metacognition, students were achieving well with the following grades being reported to administration and students at the mid-semester break: Student A: A–; Student B: B; Student C: C–; Student D: C+. Following instruction in the definition and operation of metacognition and after personal exercise of metacognition, students finished the course with these grades: Student A: A; Student B: B+; Student C: B; Student D: B–. This represents increased achievement for all students from one to four grade increments (e.g., four grade increments is equivalent to a grade increase from C– to B). It is not possible to fully eliminate factors such as increased student familiarity with the style of instruction and often-used methods of assessment, individually held beliefs and attitudes, etc., which have been shown to affect increased student achievement in some cases (e.g., House, 2000). To minimize and gauge, inasmuch as possible, the affect of one of these factors on student achievement, the instructor more often varied instructional methods and used strategies to cater for student learning preferences, and the success of case study participants in other courses where no instruction in metacognition was given was monitored and no systematic increase was found. Also, students who participated in this case study provided informal feedback to the instructor over time indicating they felt that instruction in metacognition and use of self-designed exercises to control and maximize cognitive processes was responsible for their increased success from mid-term to final course grades.

Metacognitive function as a wilful and purposeful control within the cognitive domain, can be exercised by students to maximize their learning and achievement. The results of a case study experiment in a sophomore mineralogy class indicates that direct instruction in metacognition can affect increased student achievement in geoscience education by one to four grade increments. Students took control of their means of learning, changing and adapting where needed. Continuous reflection on learning strategies, their success, and possible future adaptations was an important feature of the metacognitive process. In the context of this experiment, using learning style preferences as a cognitive framework was useful for students and allowed for instructor reflection on teaching style preferences. Direct instruction in metacognition has been shown in this environment to be an empowering tool for students who are required to achieve in a teaching/learning environment that may not always cater to their learning style preferences.

Merryn Dunmill, Christchurch College of Education, New Zealand, is thanked for drawing the author’s attention to metacognition and its role in student instruction. Professors Rodney H. Grapes and John C. Tipper, University of Freiburg, are thanked for sharing their views on the content of this paper. Discussions with education and geoscience faculty at the Ohio University, Athens, and Western Michigan University, Kalamazoo, improved a draft manuscript.