1Splits Between Student and Expert Attitudes and Beliefs about PhysicsAmy LytleIntroductionMost in the field of physics education research believe that student attitudes and beliefshave a strong effect on how students will learn, and are interested to know how and why. Also,educators often have implicit or explicit goals of teaching students something about the nature ofscience and learning, and the role of physics in everyday life, among other things, which E. F.Redish calls the “hidden curriculum” [1]. In order to assess what effect student beliefs will haveon their performance, or evaluate the effectiveness of teaching goals which aim to shape studentbelief, researchers in the field have created surveys that may be implemented easily in theclassroom [2-4]. The Colorado Learning Attitudes about Science Survey (CLASS) wasdeveloped by researchers at the University of Colorado for the purpose of extending on previoussurveys which probe student attitudes and beliefs [5,6].With the Maryland Physics Expectations (MPEX) survey [2], Redish and fellowresearchers showed that, unless the elements of the hidden curriculum are attended to specificallyin the course, students will generally develop attitudes which move further away from those theirteachers hope to develop. They conjectured that students may need to develop attitudes andpractices different from the more sophisticated attitudes of experts in the field in order tosucceed. Traditionally taught courses may in fact produce a split between what students learn todo and what they actually believe. McCaskey, et al. [7] used the content assessment survey, theForce Concept Inventory (FCI), in an effort to probe the effects on assessment due to splitsbetween belief and understanding. To do this, they asked students to complete the survey twice,first according to what “they really believe,” and then according to how they thought a scientistwould answer. The students they tested indicate splits in belief on many of the questions. Theevidence of so many splits on student’s beliefs about the content of physics suggests there wouldalso be many splits in students’ attitudes about learning physics.In this paper, I examine the results of the CLASS in which students were asked to answerits questions in two different ways: according to what they believe, and according to how theythink a physicist would answer. Specifically, I look at discrepancies between the students’answers in these two perspectives, and how accurately students were able to predict expert2opinions. I discuss the significance of the discrepancies in light of others’ research on thedifferences between expert and novice attitudes and beliefs.DataThe CLASS survey was administered to a large-university, algebra-based introductoryphysics course, with an enrollment of approximately 175 students. The majority of the studentswere pre-med-related majors, many in the life sciences. The data consist of aggregate datacollected from the post- results of the CLASS survey given in this course, in which the studentswere asked to answer survey questions in two ways: first, according to their own beliefs, andsecond, according to how they believe a physicist might answer. Each response falls into one ofthe five Likert-scale responses (strongly agree, agree, neutral, disagree, strongly disagree).In addition, there are video recordings of interviews with five of the students from theclass, in which the interviewer goes through each question of the survey, asking students toanswer in these two different ways, and to clarify both their interpretation of the question andwhy there might be differences in the student’s own beliefs in comparison with the student’spredictions of a physicist’s beliefs. Among the students interviewed, there are three women andtwo men, most of whom are in the first or second year. All of the students were enrolled in thecourse in order to satisfy a requirement for a major other than physics. The interviewer askedeach student a similar set of questions, about his year, major, whether he enjoyed the lectures andlabs, how he studied, and how much time he typically spent per week working on the class. Theinterviewer then asked each student to go through each question on the survey, to answer in thetwo different ways described above, and to provide a brief explanation for how she came to herdecisions.The raw data are included as an appendix, including the CLASS questions, the expertresponses, and the numbers of each type of response by the students, for each of the two waysthey answered the questions.AnalysisFor clarity in analysis of the data, the students’ responses will be simplified as follows:there will be no distinction made between strongly (dis)agree and (dis)agree, leaving only threecategories of responses: overall agreement, overall disagreement, and neutral. The CLASS was3administered to a group of experts [5] in order to define the expert response for each question. Inorder to examine the splits between a student response and the expert response to each question,I have plotted only the questions for which the expert response was overall agree or overalldisagree, leaving out the three questions on which the expert response was neutral (# 7, 24, and29). These specific questions are examined in a different way below. This way, the studentresponses will either be favorable (indicating the same as the expert response), unfavorable(indicating the opposite of the expert response), or neutral.100806040200% favorable responses100806040200% unfavorable responsesWhat a physicistwould saya)100806040200% favorable responses100806040200% unfavorable responsesWhat students believeb)Fig. 1. A summary of student responses to the CLASS, indicating the percentage of favorableand unfavorable responses for each question. a) What students thought a physicist would say andb) what student themselves believe.Figure 1 shows a summary of students’ responses to the questions on the CLASS. Eachpoint represents one question of the survey, with the vertical axis showing the percentage offavorable responses to the question, and the horizontal axis showing the percentage ofunfavorable responses. The horizontal line represents where the points would lie if all of theresponses were either favorable or unfavorable. The