Educational Methodologies
Teaching Style Beliefs Among U.S. and Israeli Faculty Linda S. Behar-Horenstein, Ph.D.; Gail S. Mitchell, M.P.H., R.D.H.; Netta Notzer, Ph.D.; Randy Penfield, Ph.D.; Ilana Eli, D.M.D. Abstract: The purpose of this study was to determine if self-reported teaching style beliefs were different among faculty at a U.S. and an Israeli dental school. Teacher-centered practices refer to beliefs that the teacher holds the subject matter expertise and students are generally passive learners who must be told what to think. Student-centered practices refer to beliefs that students must learn how to construct their own understanding. Student-centered teaching is directed towards enabling students to think about complex issues. Twenty-seven of fifty-eight (47.37 percent) faculty at a dental school in the United States and thirty of thirty-four (88 percent) Israeli dental faculty teaching in basic science courses completed the Teaching Behavior Preferences Survey (TBPS). The TBPS is a thirty-item instrument that measures two domains of teaching styles—teacher-centered (TC) and student-centered (SC)—and four subdomains: methods of instruction (MI), classroom milieu (CM), use of questions (UQ), and use of assessment (UA). Findings revealed that there were no significant differences in student-centered and teacher-centered teaching practices and methods of instruction, classroom milieu, and use of questions. There was a significant difference between the U.S. and Israeli groups in their reported use of assessment. The U.S. faculty reported a greater preference for student-centered assessment practices than did the Israeli faculty. Dr. Behar-Horenstein is Professor, Department of Educational Administration and Policy, University of Florida; Ms. Mitchell is Director of Curriculum and Instruction, College of Dentistry, University of Florida; Dr. Notzer is Professor, Medical Education Unit, University of Tel Aviv, Tel Aviv, Israel; Dr. Penfield is Assistant Professor, University of Miami; and Dr. Eli is Professor, Medical Education Unit, University of Tel Aviv, Tel Aviv, Israel. Direct correspondence and requests for reprints to Dr. Linda S. Behar-Horenstein, University of Florida, Department of Educational Administration and Policy, P.O. Box 117046, Gainesville, FL 32611-7046; 352-392-2391, ext. 299 phone; 352-392-0038 fax;
[email protected]. Key words: teaching styles, teaching beliefs, assessment, survey research Submitted for publication 12/9/05; accepted 5/1/06
P
rofessional educators typically do not teach from a knowledge base of educational training and evidence-based teaching practices.1 However, the ways in which faculty have been taught has a significant influence on how they teach.2-5 Understanding the relationships between beliefs and teaching is important. For example, how dental students are educated about the treatment of oral disease is grounded in the faculty’s beliefs and their knowledge base of teaching.6 Although there have been studies about elementary and secondary educators’ preferences for teaching styles, little is known about the teaching preferences of faculty in higher education in general and in dental education in particular.7,8 The purpose of this study was to determine if self-reported teaching style beliefs were different among faculty at a U.S. and an Israeli dental school. Teaching styles can be viewed “as a broad dimension or personality type” that encompasses an educator’s August 2006 ■ Journal of Dental Education
approach, attitude, or beliefs towards teaching that can be observed over time.9 Teaching styles can be viewed as the relationship between teacher and students or as the tasks of teaching such as how the teacher carries out instruction or organizes learning or the classroom environment. The notion of stability or patterns is central to the definition of teaching styles.9 Teaching styles can be implicit and tacit or can be explicitly known by the teacher. The characteristics of one’s teaching style may be influenced by philosophy of how students should or can learn or by beliefs about how students should be taught. Teaching styles tend to become ingrained after many years of practice and can be difficult to change unless teachers receive intensive feedback or are influenced by powerful stimuli.9 The association between the faculty responses to the Teaching Behavior Preferences Survey (TBPS) was assessed. Teaching styles’ beliefs were measured 851
across two domains—teacher-centered (TC) and student-centered (SC)—and four subdomains: methods of instruction (MI), classroom milieu (CM), use of questioning (UQ), and use of assessment (UA). Actual teaching styles were at the discretion of the faculty. Teacher-centered beliefs are grounded by the conviction that the teacher is the subject matter expert, whereas students are passive learners who must learn what to think. Teaching in ways that are primarily responsive to students’ learning needs/ styles represents student-centered belief items. The TC and SC scores are independent scales. Instructional methods refer to the type of teaching provided in the classroom. Classroom milieu refers to the type of structure provided during classroom instruction. Assessment refers to use of assessment techniques during classroom instruction. Questioning refers to the type and use of questions asked during classroom instruction. Representative items for each of the four subscales are shown in Table 1. Prior to this study, the TBPS was pilot-tested on 108 educators including dental and dental hygiene professors, alternative education teachers, and graduate students in education. A reliability analysis revealed that the alpha coefficient for the teachercentered scale was .64. The alpha coefficient for the student-centered scale was .72.
Methods We selected the two schools for this study because of our affiliations. The School of Dental Medicine at Tel Aviv University accepts forty-two to fifty-five students per year. The school is state-
supported but it is fully autonomous. Dental studies in Israel are based on a six-year program. Students enter after they have graduated from high school and have completed their compulsory three-year military service. The first three years of dental school are devoted to basic science and preclinical studies. At the end of this period, students receive a Bachelor in Science degree (B.Sc.) in medical sciences and then continue their studies in the dental clinical phase. During years four through six, students study subjects specifically related to clinical dentistry and treat patients in the various dental disciplines. Upon graduation, students are conferred a D.M.D. degree and are qualified to participate in the state board examinations in dentistry. Once they pass the board exam, they receive a license to practice dentistry in Israel. The University of Florida’s College of Dentistry (UFCD) accepts approximately eighty to eightythree students per year. The University of Florida is a research-intensive institution and is partially state-supported. The dental school is accredited by the Commission on Dental Accreditation (CODA). Students enter a four-year program after receiving or while completing a baccalaureate degree. The first two years consist of basic science and preclinical studies. Students work with patients in intramural and extramural clinical settings during their last two years. They take the National Board Part I at the completion of their second year and Part II in the middle of their fourth year. After students complete their courses and degree requirements and have passed the board examinations, they are conferred the D.M.D. degree. Following graduation, they may take state or regional licensure examination(s) and/or apply to a residency program.
Table 1. Representative teacher-centered or student-centered items by subscale for the Teaching Behavior Preferences Survey
852
Subscale
TC/SC Items
Methods of Instruction
• My teaching is guided by instructional strategies. • My teaching is guided by instructional strategies that help learners make meaning. • I adjust my teaching techniques based on learners’ behaviors.
Classroom Milieu
• My learning environment is efficient and highly structured. • My learning environment encourages learners to work quietly and stay on task. • My learning environment is supportive and cooperative.
Assessment
• I use the same assessment techniques for each objective of the lesson. • I change assessment technique according to the lesson objectives. • My assessment techniques are not influenced by the material I teach.
Questioning
• I ask learners questions frequently to determine what they understand. • When I ask learners questions, their answers typically require them to provide justification. • I ask learners questions infrequently because I believe they can synthesize material.
Journal of Dental Education ■ Volume 70, Number 8
For this study, we identified clinical faculty personnel at both institutions as dental faculty who taught basic science courses in the areas of oral pathology and oral medicine. We matched basic science courses at each university. The UFCD faculty pool for the matched basic science courses was fifty-eight. The Israeli faculty pool for the matched basic science courses was thirty-four. During the spring of 2003, a total of fifty-seven faculty, which included twenty-seven (N=58, 47.37 percent) U.S. faculty and thirty (N=34, 88 percent) Israeli faculty teaching in basic science courses, completed the Teaching Behavior Preferences Survey (TBPS). The UFCD faculty was somewhat diverse. Fifty-three (92.4 percent) were Caucasians, while five (8.6 percent) were of other races or ethnicities. Fifty-two (89.7 percent) were male, and six (10.3 percent) were female. There were sixteen clinical and forty-two nonclinical faculty members. The ethnicity of the Israeli faculty is homogeneous, with all thirty-four being Israelis. Twenty-two (64.7 percent) were male, and twelve (35.3 percent) were female. There were twenty clinical and fourteen nonclinical faculty members. The U.S. dental faculty completed the TBPS via web-based administration. For the Israeli participants, the TBPS was initially translated into Hebrew by the medical education staff and then verified by translating it back into English. The questionnaire was circulated in person by the medical education staff and returned by mail.
Survey Research Methodology Survey research is an efficient way to collect data. However, surveys are limited if there are errors in coverage, sample, measurement, or nonresponse.10 To reduce the possibility of coverage error, we surveyed the matched basic science course faculty pool at each university. To reduce measurement error, we pilot-tested the survey to ensure that the items were unambiguous and that each item offered each participant the same response options. To reduce nonresponse error, we made three attempts to collect data from both the UFCD and Israeli faculty. Response bias, which occurs when responses do not accurately reflect the views of the sample, is a potential concern in surveys in general. Individuals’ responses, for example, may be overtly negative or positive. Because the frequency distribution of the responses was normal, we believe that response bias is not a threat to our findings. August 2006 ■ Journal of Dental Education
Agreeing-response bias also represents a potential threat to conducting survey research. The agreeing-response bias, or acquiescence—the tendency for respondents to agree with attitude statements—has produced a large research literature mainly within psychology.10 There are different interpretations of agreeing-response. Psychologists who emphasize its importance regard acquiescence as a personality trait.11 However, others have challenged this view, stating, for example, “The overwhelming weight of evidence now points to the fact that the agreement tendency is of little importance either as a measure of personality or as a source of systematic invalidity in measures of . . . sentiments.”12 Some researchers argue that agreeing-response bias is of some importance, but that it contributes much less to covariation between items than content.13 Others have characterized agreeing-response as a form of deference by low-status respondents.14 The phenomena of agreeing-response bias could “play a dominant role in interviewing educationally heterogeneous populations, yet disappears entirely when student samples and self-administered questionnaires are employed.”12,13 The method employed in this self-administered questionnaire and researchers’ concern about the agreeing-response bias phenomenon are no greater than in other instruments using this type of measure. Finally, we decided to use the commonly used Likert scale in our survey because it minimizes the participants’ overall response time and because it is an efficient method of collecting data. Additionally, we chose Likert because ultimately we were interested in developing a psychometrically valid scale.
Description of the Teaching Behavior Preferences Survey (TBPS) The TBPS assessed individuals’ beliefs about particular teaching styles. The TBPS consists of thirty randomly intermixed statements that describe teachercentered or student-centered styles. The items were derived from a review of the historical and emergent literature on teacher effectiveness.15-19 Respondents were asked to read statements such as “I use the same teaching techniques for every learner in my course” (#18). Using a 5-point Likert scale where 5=highly agree and 1=highly disagree, respondents indicated the number that most nearly represented their level of agreement. Scores on the instrument reveal the strength of an individual’s beliefs for styles within two 853
domains—teacher-centered and student-centered— and four subdomains (teaching styles): instructional methods, classroom milieu, questioning, and assessment techniques. Examples of the teacher-centered and the corresponding student-centered instructional practice for each subdomain are described below. A teacher-centered (TC) item in the instructional methods subdomain is “I teach idiosyncratically using the techniques that I enjoy.” The student-centered (SC) item is “I match my techniques to instructional objectives.” For the classroom milieu subdomain, a TC item is “My learning environment is efficient and highly structured.” The SC item is “My learning environment is supportive and cooperative.” For the use of questions subdomain, a TC item is “I ask learners questions infrequently because I believe that they can synthesize material.” The SC item is “I ask learners questions frequently to determine what they understand.” For the use of assessment subdomain, a TC item is “I use the same assessment techniques for each objective of the lesson.” The SC item is “I change assessment techniques according to lesson objectives.” Faculty were only asked to indicate their level of agreement with each of the thirty items. First, means and standard deviations for the teacher-centered and student-centered domains for the U.S. and Israeli faculty groups were calculated. Then, means and standard deviations for the teaching styles (methods of instruction, classroom milieu, assessment, and use of questions) were calculated. Independent samples t-tests comparing U.S. and Israeli mean scores were also computed to determine if there were significant differences. In social science, it is standard practice to treat Likert-scale data as interval, meaning that the psychological distance between the ratings of 5 (strongly agree) and 4 (agree) are assumed to be equal to the differences between ratings of 1 (highly disagree) and 2 (disagree). Most social scientists analyze Likert data using parametric techniques such as t-tests unless they have reason to believe that population data are not normally distributed. The frequency of the scores for UFCD and Israeli faculty samples was normally distributed. However, some authors have observed that using techniques on individual items rather than aggregated “data will frequently be skewed,” and often “do not capture the true limits of the attitude.”20 Since the data was aggregated, the analytical techniques support the use of the t-test. Further, the t-test is appropriately used when researchers are comparing two means to determine
854
the probability that differences reflect real differences between groups of subjects rather than a chance variation in data. Such methods of analysis are statistically powerful in the sense that they have a good chance of detecting relationships that exist in data.
Results The alpha coefficient among the U.S. faculty for the student-centered (SC) domain was .72; for the teacher-centered (TC) domain, it was .75. An item analysis showed that among the fifteen SC items, five had discrimination scores of less than .20. Of the TC domain, three items had a discrimination score of less than .20. Mean scores for the SC domain ranged from 3.21 to 4.17. Mean scores for the TC domain ranged from 2.00 to 4.52. The alpha coefficient for the SC and TC domains approached high reliability.14 The alpha coefficient among the Israeli faculty for the student-centered domain was .51; for the teachercentered domain it was .56. An item analysis showed that among the fifteen student-centered items, four items had a discrimination score of less than .20. In the teacher-centered domain, eight items had a discrimination score of less than .20. Mean scores on the student-centered items ranged from 2.42 to 4.46. Mean scores on the teacher-centered items ranged from 1.83 to 4.28. Given the large range among the mean scores for the items, it is not surprising that the reliability scores for Israeli SC and TC scales are lower than the U.S. group.21 Descriptive statistics for the United States and Israel, the total scores for the teacher-centered domain and student-centered domains, and the subdomains including the methods of instruction, classroom milieu, assessment, and use of questions subscales are shown in Table 2. The independent samples t-tests comparing the U.S. to the Israeli mean scale scores (Table 3) found that assessment (t= 2.961, p
