Education-line Home Page

Epistemological beliefs in learning to teach: Resolving conceptual and empirical issues

Paper presented at the European Conference on Educational Research Ljubljana, Slovenia. September 17-20, 1998.

Bob Elliott

School of Professional Studies
Faculty of Education,
Queensland University of Technology

K.W. Chan

Hong Kong Institute of Education


The ways in which students learn and comprehend complex subject matter has been the subject of educational research for considerable time. Of particular note has been the line of research which investigates the relationship between learning and the beliefs that students hold about the nature of knowledge and the process of its acquisition. Grouped together, these sets of beliefs are referred to as epistemological beliefs. This line of research about the influences on learning can be traced to early studies conducted by Perry (1968) with Harvard undergraduates. In his study, Perry(1968) suggested that students go through stages of epistemological development in a sequential and linear manner. That is, he proposed that in the early stages students regard knowledge in absolute terms- as either right or wrong- and believe that it is transmitted through authority. As they reach the later stages of their development, Perry concluded that they realize that there are multiple possibilities ways for viewing knowledge and there are times when one must make a strong , yet tentative commitment to some ideas.

Based on Perry's scheme, Ryan (1984) developed a dualism scale and classified students as either highly "dualist" or highly "relativist" and attempted to link epistemological beliefs to academic performance. More recently, investigation of the relationship between Perry's conception of epistemological beliefs and metacomprehension has also been attempted by Glenberg and Epstein (1987). Conflicting and inconsistent results were reported in these studies. One explanation of this was possibly due to the assumption that personal epistemology is unidmensional and develops in a fixed progression of stages. Schommer, (1990, 1994a, 1994b) has argued that a counter position is that epistemological beliefs should be conceived as a multidimensional system of more or less independent beliefs. By a multidimensional system she means that there is a set of beliefs to consider as constituting personal epistemology and by more or less independent beliefs she means that individual beliefs within the system may develop at different rates or may be inconsistent with each other. For example:

_an individual may believe that knowledge is highly complex and involves an intricate network of ideas. Yet that person may also believe that knowledge is completely certain, that is, never changing. (Schommer &Walker, 1997:175)

Based on these assumptions, Schommer(1990) hypothesized five epistemological dimensions, initially conceived as continua, viz. Simple Knowledge, Certain Knowledge, Omniscient Authority, Quick Learning and Innate or Fixed Ability. Simple Knowledge means "Knowledge is simple rather than complex". Omniscient Authority implies "Knowledge is handed down by authority rather than from reason". Certain Knowledge indicates "Knowledge is certain rather than tentative". Innate Ability means "The ability to learn is innate rather than acquired". Quick Learning is "Learning is quick or not at all". These beliefs were initially conceived as continua as shown in Table 1.

Table 1 Schommer's (1990) Hypothesized Epistemological Dimensions

  1. Source of knowledge: From knowledge is handed down by omniscient authority to knowledge is reasoned out through objective and subjective means.(Omniscient Authority)
  2. Certainty of knowledge: From knowledge is absolute to knowledge is constantly evolving. (Certain Knowledge)
  3. Organization of knowledge: From knowledge is compartmentalized to knowledge is highly integrated and interwoven.(Simple Knowledge)
  4. Control of learning: From ability to learn is genetically predetermined to ability to learn is acquired through experience.(Fixed Ability)
  5. Speed of learning: From learning is quick or not-at-all to learning is a gradual process.(Quick Learning)

Based on five hypothesized dimensions, an Epistemological Belief questionnaire consisting of 63 items was developed by Schommer and tried with American university/college students in a number of studies. (Schommer, 1990, 1992, 1993a, 1993b). The 63 items were grouped into 12 conceptual subscales/subsets which were used as variables in factor analysis. Her studies reported consistent findings that four factors or dimensions (out of the five proposed dimensions) were extracted in exploratory factor analysis using Principal Axis Factoring Method followed by varimax rotation. These four factors or dimensions found from the empirical work are Simple Knowledge, Certain Knowledge, Quick Learning and Innate Ability. The hypothesized Omniscient Authority factor was not found in her studies as the subscales for omniscient authority did not load with substantial loading values.

Schommer's work has initiated other researchers' investigations in the area of the dimensions of epistemological beliefs and the relation to various aspects of learning. Research studies by Schommer and other researchers (e.g. Dunkle, Schraw & Bendixen, 1993; Dweck & Bempechat, 1993; Dweck & Leggett, 1988; McDevitt, 1990; Schommer, Crouse & Rhodes, 1992; Spiro, Coulson, Feltivoch, & Anderson, 1988) have indicated that epistemological beliefs are related to students' persistence, active inquiry, integration of information, and coping with complex and ill-structured domains. (see review by Schommer, 1994a, 1994b).

Jehng, Johnson and Anderson (1993) examined American university students' epistemological beliefs as a function of their educational level and field of study using a questionnaire adapted from Schommer. They reported five factors /dimensions, viz. Certain Knowledge, Omniscient Authority, Innate Ability, Quick Learning and Orderly Process in a confirmatory factor analysis study. Schommer interpreted Jehng et al's finding of Orderly Process as resembling her conceptualization of Simple Knowledge. However, such a comparison should be made with caution because it is based on the meaning of the statements representing the orderly process and what Schommer defined as Simple Knowledge. Notwithstanding this point, the Jehng et al study suggested that students' beliefs about learning are a product of the activity, the culture, and the context in which they are cultivated.

Another study by Qian and Alvermann (1995) examined the relationships between epistemological beliefs and learned helplessness in secondary school students' learning science concepts. This study revealed only three factors for the epistemological belief structure. These were extracted from exploratory factor analysis. The 3 factors isolated by these researchers were : Simple-Certain Knowledge , Quick Learning and Innate Ability where the first factor is a combination of Simple Knowledge and Certain Knowledge. In this study, items related to Omniscient Authority were first removed before developing the revised Schommer epistemological beliefs questionnaire.

While there has been a general interest in the relationship between belief structures and learning, in the field of teacher development there has also been a similar interest. Specifically, there has been a change of research interest in teacher education from studying teachers' classroom behavior and teaching skills to examining teachers' thinking and beliefs. Such a shift is based on the assumption that a need for increased understanding of teachers' decision making and practice in the classroom (Armour-Thomas, 1989; Clandinnin & Connelly, 1987; Clark & Peterson, 1986; Fenstermacher, 1978). Although it is a difficult task to study teachers' beliefs, it is likely to be a rewarding area of research if concepts of specific beliefs are carefully operationalized, appropriate methodology chosen, and design thoughtfully constructed (Pajares, 1992). Pintrich (1990) has suggested that beliefs will ultimately prove the most valuable psychological construct to teacher education. In fact, it has been recognized that many of the obstacles towards educational reform have their basis in the existing beliefs of teachers because such beliefs may determine the disposition of teachers towards a particular change.

Research in teacher education has indicated that student teachers bring with them prior knowledge and ideas (forming their initial beliefs) which filter knowledge acquisition in teacher education programs offered in colleges and universities ( e.g. Calderhead & Robson, 1991; Hollingsworth, 1989; Tilleman, 1995). Subsequently, to understand more closely how teachers can be developed and how new changes in educational practice can be introduced, it is important to understand teachers' belief structures.

Of all the beliefs held by teachers, the beliefs about the nature of knowledge and learning, known collectively as epistemological beliefs, appear to be those which may influence teachers' choice and decisions in the classroom. It is likely that the methods a teacher uses, how the teacher manages the class, what to focus in learning and so on are all likely to be influenced by the beliefs the teachers holds about knowledge and its acquisition. It is likely to be helpful to teacher educators in constructing teacher education programs and in interacting with students if they have a knowledge of their students' epistemological beliefs.

Like a considerable amount of educational research the work on epistemological beliefs has been undertaken in America. Results from this research are spoken about as if they are culturally neutral. For this reason it is important to test the validity of such research in particular cultural contexts before it is considered in those contexts. It is important to determine if there are differences in research findings from one context to another and whether the assumptions behind much of the research in one context holds true in another.

In summary, the study of epistemological beliefs is a developing area of research which is rewarding but difficult due to the complexities involved. There is still much that is unknown and needs to be uncovered such as the nature of beliefs and the influence of cultural context on epistemological beliefs. Schommer has contributed to the area by her challenge of the unidimensional nature of beliefs and by initiating other researchers' investigations in this area . However, there is considerably more work to be undertaken of both an empirical nature and a conceptual nature. For example, there is a need to validate the conceptual subscales proposed by Schommer through empirical analysis. Further, since Schommer's research is based on her 12 hypothetical subscales as variables there is a need to investigate how these 12 subscales are defined and measured because there is no research which evidences the existence of these 12 subscales and the applicability of the questionnaire to different cultures.

Objectives of Study

Student teachers' epistemological beliefs probably influence their conception about learning, and consequently, their preference for a certain way of teaching in terms of approaches and classroom management. Therefore, a fuller understanding of epistemological beliefs held by teacher education students would be of much value to teacher educators. Further, epistemological beliefs are likely to be shaped by the socio-cultural context in which they are developed and much of the research is based in an American context with little known about Chinese context. Accordingly, one of the objectives of this study is to determine whether Hong Kong student teachers' epistemological beliefs are represented by the same four dimensions as reported by Schommer in her American sample. In particular, it is important, given the different cultural contexts, to investigate whether the dimension of Omniscient Authority is one of the belief dimensions held by the Hong Kong teacher education students given that the term "authority" is mentioned so frequently in the Chinese Society/Culture.

Moreover, this study intends to examine whether Schommer's hypothesis that the dimensions of epistemological beliefs are relatively independent also holds true for a different culture. In Qian and Alvermann's (1995) study, Certain Knowledge was found to merge with Simple Knowledge Factor, so this study also aims at finding whether such complex factors or merging factor structure of epistemological beliefs also exist in the Asian culture.


The final year students of the full-time two year Certificate of Education Course (CE course) of the Hong Kong Institute of Education were chosen as the sample for this study. They comprised of 88 male and 234 female students taking various subject combination as their electives or fields of study in teaching. The age range was from 20 to 26. The CE course at present is the only full-time course to prepare preservice nongraduate teachers. The course is designed for "A" level matriculants and hence the students in the CE course are equivalent to undergraduate students in universities and colleges despite it being a subdegree program. Like other institutions for teacher education, there are more females than males in the program and all of them have taken some foundation courses in professional studies including education psychology, educational theories, learning and instruction and sociology of education as well as their subject disciplines.

Schommer's 63 items questionnaire was used as the instrument for data gathering. This instrument was provided in both Chinese and English language. Demographic data including their age, sex and electives taken were also collected for further research studies.


Exploratory factor analysis using principal axis factoring and varimax rotation was used to analyze the data. The 63 items were grouped into the 12 subscales based on Schommer's conceptual system. The 12 subscales/subsets were used as variables for factor analysis and data analysis was based on the varimax (orthogonal) rotation structure as oblimin rotation yielded a similar factor structure. Eigen value of 1.00 was taken as cut-off point and three factors were extracted accounting for 46.5% (cum pct) of the total variance. The fourth factor had a eigen value of .98953. If this fourth factor was taken together, then the cum pct. of variance would be 54.7. The situation of one factor with an eigen value marginally less than one was quite similar to one of the results reported by Schommer (1993). She chose an eigen value of .98 as the cutoff to produce a four factor structure. On examining the scree plot in this analysis there was a sharp break indicating a 3 factor solution. Therefore a factor structure of 3 was used to describe and explain the epistemological dimensions held by Hong Kong student teacher sample in this study. A factor structure as shown in Table 2 was obtained by displaying the factor loading of the three factor solution in descending order. Loading of 0.40 and above are shown in bold and italic type.

Table 2

Loading for three factors with eigen values greater than 1.00 as cut-off

(PAF: Varimax Rotated factor Matrix) in descending order:

Factor 1 Factor 2 Factor 3 TR>
Can't learn how to learn .68765 .21368 -.36457
Success is unrelated to hard work .60500 .09739 .02729
Learn first time .41194 .11692 .16768
Avoid integration .35140 .25955 .12982
Concentrated effort is a waste of time .31609.02157 .01389
Don't criticize authority .10863 .80418 .07867
Knowledge is certain .08740 .42856 .08864
Learning is quick .23372.40020.22753
Ability to learn is innate .15885 .08965 .61359
Avoid ambiguities .01214 .06204 .47888
Seek single answers -.29334 .19230 .35124
Depend on authority .01569 .23453 .31349

In Table 2, the subscale for Omniscient Authority (Don't criticize authority) loads with a very high value on the same factor as the subscale (Knowledge is certain) for Certainty Knowledge. Other factors were also complex combinations with two or more of the Schommer subscales being represented. Such findings mean it is difficult to interpret the belief structure.

Interpretation of the generated factor structure (as shown in Table 2) began with examining the loading of variables on the factors. Factors were given descriptive titles on the basis of high-loading subscales of items. Factor 1 has a prominent loading greater than 0.6 of two subscales representing "Innate Ability" ("Can't learn how to learn" and "Success is unrelated to hard work") and one subscale ("Learn first time") representing "Quick Learning". As the two subscales for "Innate Ability" loaded on Factor 1 with substantially high loading compared with that for "Quick Learning", so Factor 1 was named as "Innate Ability". Factor 2 has a loading of 3 subscales representing 3 different dimensions, "Omniscient Authority", "Certain Knowledge" and "Quick Learning". The most obvious one comes from the subscale "Don't criticize authority" representing "Omniscient Authority" with a loading value of 0.8, and the other two had a loading of 0.4 to 0.43. In view of the fact that the subscale called "Don't criticize authority" had much a higher loading value compared with the other two, Factor 2 was labeled "Omniscient Authority". The association of the subscales for authority and certain knowledge on the same factor could lead one to name factor 2 as "Authority-Certain" which means knowledge coming from authority is certain. Factor 3 had two subscales with loading values greater than 0.4. A subscale "Ability to learn is innate (0.61) representing "Innate Ability" and a subscale "Avoid ambiguities" representing "Certain Knowledge" (0.48). The two were of complete different nature. The merging of the subscales made it difficult to name the complex factor. As two other subscales for "Innate Ability" had loaded appreciably on Factor 1, the loading of the other Schommer subscale for "Innate Ability" on Factor 3 was out of position and inconsistent. Similar results were reported in Schommer's previous studies and she simply commented that this finding was inconsistent with her hypothesis. Having reconsidered the whole factor structure, which was complex because of the merging factors obtained, it was decided to label the factors in this sample as : Factor 1 (Innate Ability), Factor 2 (Omniscient Authority) and Factor 3 (Certain Knowledge).

The identified factor structure was problematic in that a number of subscales belonging to different dimensions based on Schommer's hypothesized conceptual framework loaded on the same factors, making it difficult to interpret and label factors if the original assumption of independent dimensions as proposed by Schommer was held. Some subscales tended to combine together to form complex factors such as "Authority-Certain" in Factor 2.

What was obvious in this study was that the Omniscient Authority Factor, which was not reported to load significantly in Schommer's study had a very high loading of .80 on Factor 2 ("Don't criticize authority"). This might be because of the different sociocultural contexts of the two studies. "Authority" is usually respected in Asian cultures and there is a tendency not to query or criticize the viewpoints of authority figures. The traditional child rearing process in Asian countries, such as China, highlight respect for elders and authority figures. This contrasts with many Western countries, such as North America, where democracy and liberty are emphasized to greater extents.

The other factors extracted in this study were similar to Schommer's findings viz. "Innate Ability" and "Certain Knowledge". The tendency of merging "Innate Ability" and "Quick Learning" in Factor 1 might result in a new labeling to factor 1 as "Ability-Quick". That may mean that quick learning is associated with innate ability. Again this might be due to cultural factors in that Hong Kong students may believe that innate ability determines not only whether one can acquire knowledge or not , it also determines the speed of acquisition of knowledge. Therefore if one is bright, one is born that way and can acquire knowledge quickly even at the first time of reading.

The above three factor structure gives an overall picture of the epistemological belief system held by the Hong Kong teachers education students in the sample of study. The finding suggests they believe strongly in the knowledge handed down by authority, and hence do not criticize what the experts or those in authority say, and believe these kind of knowledge are unambiguous and certain. This pattern was similar to that noted by Perry's who found university undergraduates at the early stage of development had a strong belief in authority and certain knowledge. The Hong Kong teacher education students in this study came from Form 7 graduates (high school leavers for university entrance) who were going to complete two more years of study in the teaching education institute, hence they were not far off from high school leavers and might exhibit similar patterns or trends of belief system. This was further enhanced by the traditional Chinese culture in authority belief which made them to believe and follow the suggestions and practice of experts and authorities. In addition, they also believe strongly in "Innate Ability/Fixed Ability" in learning, and quick learning which was similar to Schommer's findings of college and high school students. The major differences in finding of this study from that of Schommer fell in the following respects.

(1) Instead of relatively independent dimensions, the factors or dimensions underlying epistemological beliefs of Hong Kong teacher education students were related and merged into complex factors.

(2) The factor or dimension: " Omniscient Authority " is obvious and carries a high contribution to the beliefs system of the Hong Kong sample

(3) The factor "Quick Learning" did not load significantly compared with others (with loading values greater than 0.6). Very often subscales for "Quick Learning" came together with other subscales associated with "Simple Knowledge", "Certain Knowledge' and "Omniscient Authority".

Furthermore, even when the eigen value of .98 as cutoff was taken to generate four factors, the structure was not simplified and clarified. Loading of subscales of different dimensions with high loading values in the same factor continued. That is, the issue of merging of factors was not resolved by considering a four factor structure and did not generate a clearer picture.

The fact that the Schommer results could not be replicated in the Chinese context is not surprising given that other studies also point to the difficulties of adopting North American instruments in other social and cultural contexts. For example Mori (1997) had obtained different results from Schommer in a Japanese context and Arredondo & Rucinski (1966) point out that Schommer's instrument may not be applicable in a Chilean context.

As a way forward to resolve these conceptual and empirical issues to investigate Epistemological Beliefs in Hong Kong, the original data was re-examined by taking each of the 63 items as variables for factor analysis. The 63 items did not cluster or load together in the manner proposed by Schommer with her 12 conceptual subscales.

Accordingly, three iterations of item redevelopment were undertaken beginning with the original 63 items. Each iteration consisted of:

Reviewing each of the items for cultural relevance by discussing with students, lecturers within the Hong Kong Institute of Education;

Focusing particularly on those items with were marginally excluded from the previous exploratory factor analysis- i.e. factor loading between .3 and .4;

Reducing the number of items in the instrument (students reported that the length of the instrument was problematic);

Writing further items to tap identified dimensions when the number of items in subscale was small.

Finally, a confirmatory factor analysis was undertaken to validate the scale that was developed using this iterative procedure.

As a final result, a four subscale instrument has been developed in which the four subscales are:

Belief in authority/expert knowledge

Belief in the certainty of knowledge

Belief that learning requires significant effort

Belief that ability to learn is innate.


Attempts to replicate the Schommer findings in a Chinese context yielded merging or complex factors of epistemological beliefs with Omniscient Authority as one of the prominent factors. Such a finding might imply the importance of cultural factors in shaping epistemological beliefs. Such a finding also highlights the difficulties of using instruments developed in one culture with a different setting.

Accordingly, an instrument relevant to the Hong Kong context has been developed to map students' epistemological beliefs. This instrument is currently being used to identify the relationships between students' beliefs in the nature of knowledge and knowing and personal theories about teaching and learning.


Armour-Thomas, E. (1989). The application of teacher cognition in the classroom: A new teaching competency. Journal of Research in Development in Education, 22(3), 29-37.

Claderhead, J. & Robson, M. (1991). Images of teaching: Student teachers' early

conceptions of classroom practice. Teaching & Teacher Education, 7, 1-8

Clandinin, D. J. , & Connelly, F. M. (1987). Teachers' personal knowledge: 'What counts as personal' in studies of the personal. Journal of Curriculum Studies, 19(6), 487-500.

Clark, C.M., & Peterson, P. L. (1986). Teachers' thought process. In M.C.

Wittrock(Ed.) Handbook of Research on Teaching, 3rd ed., New York: Macmillan, 255-296.]

Dunkle M. F., Schraw G. J. , & Bendixen, L. (1993). The relationship between

epistemological beliefs, causal attribution, and reflective judgment. Paper presented at the Annual Meeting of the American Educational Research Association, Atlanta.

Dweck, C. S. & Bempechat, J. (1983). Children's theories of intelligence:

Consequences for learning. In : S. G. Paris, and G. M. Olson, and H. W.

Stevenson (Eds.) Learning and Motivation in the Classroom, Erlbaum, Hillsdale, N. J. 239-256.

Dweck, C. S. & Leggett, E. L. (1988). Asocial-cognitive approach to motivation and

personality. Psychological Review, 95, 256-273.

Fenstermacher, G. D. (1978). A philosophical consideration of recent research on teacher effectiveness. In Shulman, L. (Ed.) Review of Research in Education (Vol. 6, 157-185). Itasca, IL: Peacock.

Glenberg, A.M., & Epstein, W. (1987). Inexpert calibration of comprehension. Memory & Cognition, 10, 597-602.

Hollingsworth, S. (1989). Prior beliefs and cognitive change in learning to teach. American Educational Research Journal, 26, 160-189.

Jehng, J., Johnson, S.D. & Anderson, R.C. (1993). Schooling and students' epistemological beliefs about learning. Contemporary Educational

Psychology,18, 23-35.

McDevitt, T. M. (1990). Mothers' and children's beliefs about learning. Child Study

Journal, 20, 105-128.

Mori, Y. Epistemological beliefs and language learning beliefs: What do language learners believe about their learning. Eric Document. ED406657.

Pajares, M.F. (1992). Teachers' beliefs and educational research: Cleaning up a

messy construct. Review of Educational Research, 62(3), 307-332.

Perry, W.G., Jr. (1968). Patterns of development in thought and values of students in a liberal arts college: A validation of a scheme. Cambridge, MA: Bureau of Study: Counsel, Harvard University. (Eric Document reproduction Service

No.ED 024315).

Pintrich, P.R. (1990). Implications of psychological research on student learning and college teaching for teacher education. In W.R. Houston(Ed.), Handbook of Research on Teacher Education (826-857). New York: Macmillan.

Qian , G., & Alvermann, D. (1995). Role of epistemological beliefs and learned helplessness in secondary school students' learning science concepts from text. Journal of Educational Psychology, 87(2), 282-292.

Ryan, M.P. (1984). Monitoring text comprehension: Individual differences in epistemological standards. Journal of Educational Psychology, 76, 1226-


Schommer, M. (1990). Effects of beliefs about the nature of knowledge on comprehension. Journal of Educational Psychology, 82(3), 498-504.

Schommer, M. (1993a). Epistemological development and academic performance among secondary students. Journal of Educational Psychology, 85(3), 406- 411.

Schommer, M. (1993b). Comparisons of beliefs about the nature of knowledge and learning among postsecondary students. Research in Higher Education, 34(3), 355-370.

Schommer, M. (1994a). Synthesizing epistemological belief of research: tentative understandings and provocative confusions. Educational Psychology Review, 6(4), 293-319.

Schommer, M. (1994b). An emerging conceptualization of epistemological beliefs

and their role in learning. In R. Garner & P. A. Alexander (Eds.), Beliefs

About Text and Instruction with Text. (25-40). Hillsdale, N. J.: Lawrence

Erlbaum Associates.

Schommer, M., Crouse, A., & Rhodes, N. (1992). Epistemological beliefs and mathematical text comprehension: Believing it is simple does not make it so. Journal of Educational Psychology, 84, 435-443.

Spiro, R. J., Coulson, R. L., Feltovich, P. J., & Anderson, D. K. (1988). Cognitive

flexibility theory: Advanced knowledge acquisition in ill-structured domains. In V. Patel & G. Groen (Eds.) Tenth annual conference of the Cognitive Science Society (pp.375-383). Hillsdale, N. J.: Erlbaum.

Tillema, H. H. (1995). Changing the professional knowledge and beliefs of teachers: A training study. Learning and Instruction, 5, 291-318.

This document was added to the Education-line database 09 November 1998