Innovations in Studio Physics at Rensselaer Karen Cummings and Jeffrey Marxa) Rensselaer Polytechnic Institute, Troy, NY 12180 Ronald Thornton and Dennis Kuhl Center for Science and Mathematics Teaching Tufts University, Medford, MA 02155 In 1993, Rensselaer introduced the first Studio Physics course. Two years later, the Force Concept Inventory (FCI) was used to measure the conceptual learning gain <g> in the course. This was found to be a disappointing 0.22, indicating that Studio Physics was no more effective at teaching basic Newtonian concepts than a traditional course. Our study verified that result, <g FCI,98> = 0.18 0.12 (s.d.), and thereby provides a baseline measurement of conceptual learning gains in Studio Physics I for Engineers. These low gains are especially disturbing because the studio classroom appears to be interactive and instructors strive to incorporate modern pedagogies. The goal of our investigation was to determine if incorporation of research-based activities into Studio Physics would have a significant effect on conceptual learning gains. To measure gains, we utilized the Force Concept Inventory and the Force and Motion Conceptual Evaluation (FMCE). In the process of pursuing this goal, we verified the effectiveness of Interactive Lecture Demonstrations (<gFCI> = 0.35 0.06(s.d.) and <gFMCE> = 0.45 0.03(s.d.)) and Cooperative Group Problem Solving (<gFCI> = 0.36 and <gFMCE> = 0.36), and examined the feasibility of using these techniques in the studio classroom. Further, we have assessed conceptual learning in the standard Studio Physics course (<g FCI,98> = 0.18 0.12(s.d.) and <gFMCE,98> = 0.21 0.05(s.d.)). In this paper, we will clarify the issues noted above. We will also discuss difficulties in implementing these techniques for first time users, and implications for the future directions of the Studio Physics courses at Rensselaer. 1 I. INTRODUCTION Discussion of Studio Physics at Rensselaer Introductory physics at Rensselaer Polytechnic Institute is taught in a "studio" format with nearly 1000 students per term enrolling in Physics I or Physics II .1 The defining characteristics of Studio Physics are integrated lecture/laboratory sessions, small classes of 30 to 45 students, extensive use of computers, collaborative group work, and a high level of faculty-student interaction. Each section of the course is led by a professor or experienced instructor, with help from one or two teaching assistants. The teaching assisants’ roles are to circulate throughout the classroom while the students are engaged in group work. There is currently no explicit training of teaching assistants. As a result, there is great variation in their effectiveness. Introductory Studio Physics is a calculus-based, two semester sequence equivalent to the standard physics for engineers and scientists. Classes meet twice a week for sessions lasting 110 minutes each. The studio model has reduced the number of contact hours – from 6.5 hours per week to less than 4 hours – without significantly reducing course content. An expectation of some independent learning on the part of the students has become the norm. The studio format was first introduced at Rensselaer in 1993. During the Fall 1995, Cooper used the Force Concept Inventory (FCI) to measure conceptual learning gains.2,3 The fractional gain in conceptual learning ,<g>, was found to be a disappointing 0.22. This indicated that the studio format was no more effective than a traditional course structure in regard to teaching for conceptual learning. The fractional gain ,<g>, is defined as follows: % Correct post-instruction % Correct pre-instruction 100 % Correct pre-instruction This expression is often referred to in the literature as the "g" or "Hake" factor, and is the ratio of the actual gain to the maximum possible gain.4 The low gain in student understanding in Studio Physics classes was puzzling because the 2 Discussion of Studio Physics at Rensselaer studio classroom appeared to be interactive. One noticeable difference between Studio Physics and other interactive curricula is that the activities used in the studio classroom are predominately "traditional" activities adapted to fit the studio environment and incorporate the use of computers. Standard laboratory exercises were simply modified to allow the data to be collected or analyzed via the computer, without making any real changes to the nature of the exercise. For example, a traditional laboratory activity which uses a spark timer or photogates to measure the acceleration of an object in free fall has been replaced with a video analysis activity in which students measure the acceleration of an object in free fall. In general, the activities used are not based on the findings of physics education research in that they do not attempt to directly address known student misconceptions and employ neither cognitive conflict5 nor bridging techniques6. Since the introduction of the studio format in 1993, a standard approach to instruction in these courses has evolved. Although the Physics I and Physics II courses are broken into a total of approximately 20 class sections, taught by 10 different instructors, all students enrolled in a given course follow the same syllabus, do the same homework assignments, and take common exams as a single group, both at finals and during the semester. A standard course design including daily lectures, in-class activities and solutions, homework assignments and solutions, and reading assignments is provided by a course supervisor for use by all instructors. The course supervisor is also responsible for exam development. The motivation for this approach is two-fold. First, it reduces redundancy in class preparation. Second, it provides for consistency in the material covered by the various instructors. Nearly all instructors adhere to this standard course design. Nevertheless, inherent to the Studio Physics model is a certain flexibility that enables motivated instructors to include diverse approaches. Methods of Inquiry The physics education community assiduously develops creative techniques for engaging students in their own learning process. Two of the authors7 were Studio Physics I instructors during the Spring semester of 1998, and merged two such techniques, Interactive Lecture Demonstrations (ILD) 8 and Cooperative Group Problem Solving (CGPS)9,10 , with the standard set of activities in the Studio classrooms at Rensselaer. These approaches are discussed in more detail below. Our intent was to ascertain the effectiveness of these techniques while concurrently establishing the feasibility of incorporating these methods into Studio Physics. As part of our study, we divided the Studio Physics I sections into two broad categories – experimental and standard. Standard classes were taught by instructors who delivered the "standard" studio instruction that was described above. Experimental classes were taught by instructors who modified the course design to include either Interactive Lecture Demonstrations, Cooperative Group Problem Solving or both. There were seven standard classes and five experimental classes. Two of the experimental sections were taught by one of the authors and the other three experimental sections were taught by another author.7 The seven standard classes were taught by four different instructors, one of whom was the course supervisor. There was no overlap between the instructors for experimental and standard sections. To offer some objective measure of the conceptual learning gains in the two groups, the authors administered two diagnostic exams, the Force Concept Inventory3 and the Force and Motion Conceptual Evaluation (FMCE)11, to every student both preand post-instruction. II. REVIEW OF CURRICULA As noted above, two techniques developed by the physics education community were incorporated into the standard studio model of instruction on an experimental basis. Interactive Lecture Demonstrations 3 Interactive Lecture Demonstrations (ILDs) are an instructional technique developed by R. K. Thornton and D.R. Sokoloff. 8 They were developed and refined based on the findings of education research, and exploit the real-time data acquisition and display powers of microcomputer based laboratory (MBL) tools12. Interactive Lecture Demonstrations were designed to create an active learning environment in large lecture classes, or any class where providing equipment for student use is an obstacle. The steps currently prescribed by Thornton and Sokoloff for use during all Interactive Lecture Demonstrations are as follows: 1. The instructor describes and performs the demonstration without the use of MBL tools. 2. Students record their names and individual prediction on a Prediction Sheet that is to be collected at the end of the demonstration period. 3. Students engage in small group discussion about these predictions. 4. Students make any desired changes to their predictions on the prediction sheet. 5. The instructor elicits common student predictions from the class. 6. The instructor performs the demonstration with MBL tools which allow the data to be displayed to the entire class. Attention is called to the most important features of the graphs, and how they relate to the physical situation. 7. Students record and discuss the results. 8. The instructor discusses analogous physical situations. Interactive Lecture Demonstrations have been developed for a wide range of topics. The Interactive Lecture Demonstration packet which was commercially available in 1998 was used in this study.13 The purchased package provided a teacher's guide, computer files for the Mac or PC, student Prediction Sheets and Result Sheets, and teacher's presentation notes for each of four Interactive Lecture Demonstration sequences. These sequences cover kinematics (through units on Human Motion and Motion with Carts) and dynamics (through units on Newton's First and Second Laws, and Newton's Third Law). Each sequence takes between forty and fifty minutes to perform. The commercially available package was implemented "right out of the box" at Rensselaer with little preparation by Cummings and Marx who were motivated, but had no prior experience in performing Interactive Lecture Demonstrations. Deviations from the prescribed implementation of the Interactive Lecture Demonstrations occurred and were a consequence of two overlapping areas of difficulty. The first area of difficulty was the instructors' inexperience performing the Interactive Lecture Demonstrations which was compounded by the use of new software and hardware. Consequently, they paid less attention than desired to following the pedagogical suggestions set forth in the teacher's notes. Especially notable were consistent failures to make analogies to similar situations and to have students discuss the result. Furthermore, students were routinely allowed too much time to make their predictions. This was primarily a result of the instructor’s desire to have every student complete the prediction before continuing. The additional prediction time was counterproductive because some students lost interest and moved on to other things, rather than staying focused on the demonstration. The second area of difficulty encountered was performing Interactive Lecture Demonstrations in physically small studio classrooms with a student population accustomed to interacting with one another and their instructors. An important hindrance was that the classrooms used have level floors, as opposed to the raised tiers of seats present in most lecture halls. Consequently, students in the back of the room had trouble seeing the demonstration. Additionally, since studentfaculty interaction is the norm, instructors answered questions on an individual basis that would have been more constructive if they would have been asked and answered publicly. Furthermore, students in the studio courses at Rensselaer have begun to expect interaction with their peers, and hence tended to share predictions too soon, or even to make initial 4 predictions as small groups rather than individually. As a result, some students never made a personal "intellectual commitment" to their predictions. It was also routinely difficult for instructors to elicit incorrect answers from the class. This made discussion of common misconceptions awkward. Cooperative Group Problem Solving Cooperative Group Problem Solving (CGPS) is a strategy for peer instruction in quantitative problem solving which was developed by P. Heller, et. al.9,10 This instructional approach involves the following: 1. Formation of well-functioning cooperative groups 2. Presentation, modeling and reinforced use of an explicit and effective problem solving strategy 3. Use of "context-rich" problems for group practice 4. Rewarding cooperative learning through grading In order to facilitate learning, instructors organize teams of three students in which high-, average-, and under-achievers are represented. (Students are not informed of this grouping procedure.) Further, students are informed about the roles and responsibilities taken on by members of well functioning cooperative groups and are encouraged to assign these roles to members of their group. In this study, a student's initial achievement level was based on the results of conceptual diagnostic tests (FMCE and FCI), and groups were rearranged every three to five weeks following each course exam. A key tenet of Cooperative Group Problem Solving is that students, as opposed to expert problem solvers, have not yet developed a generalized strategy for quantitative problem solving. Hence, students are explicitly taught such a strategy during the first days of the course. The strategy adopted for use in the experimental group was the same as that used at the University of Minnesota14 and involves a five-step process described as follows: 1. Comprehend the problem by drawing a picture and noting the given information. 2. Represent the problem in formal terms by translating the picture into a diagram or graph. 3. Decide on the general approach and write down equations that you believe appropriate. 4. Combine equations in such a way as to solve for the target quantity before substituting in the given information. Check to make sure that units are consistent. If units are consistent, substitute in the given information and solve for a numerical answer. 5. Check to make sure that your numerical answer is reasonable, that you have answered the correct question and that units are correct. The problem-solving strategy outlined above is modeled by the instructor during lectures and then practiced by students in their group. Deviations from the prescribed approach discussed above resulted predominately from testing and grading constraints within the Studio Physics course structure. As previously mentioned, students in all sections took common exams throughout the semester. The course supervisor frequently included material based on the standard activities on these exams. Hence, the investigators felt that it was imprudent to completely displace these activities. Instead they opted to spend one class period per week on Cooperative Group Problem Solving using recitation style context rich problems available from the University of Minnesota14, and one class period per week doing the same activity that students in the standard sections did. Further, context-rich problems were not included on exams, although part of the students’ class activity grade was based on their cooperative group problem solving work. The investigators encountered three other major difficulties. The first was that due to time constraints, the instructor in the Cooperative Group Problem Solving sections did not model the problem solving technique as often as 5 desired. The two other difficulties we encountered appear to be more general in nature. The first was student resistance to assignment of "roles" within the group. Several students expressed that they were very uncomfortable with this process. As a whole, the students could not be encouraged to adopt roles within the groups, and hence this aspect of the technique soon died out. The second general difficulty we encountered was that the problemsolving procedure outlined above is typically not relevant when solving textbook-style homework problems like those assigned in the Studio Physics course at Rensselaer. Students were required to use the problem-solving procedure on all their homework assignments and resented having to use this procedure if they could easily have solved the problem without it. III. ASSESSMENT AND EVALUATION As mentioned above, the authors administered two diagnostic examinations, the Force Concept Inventory and Force and Motion Conceptual Evaluation (FMCE). These two exams were labeled as "Part A" and "Part B" of a single exam packet, and both exams were given pre- and post-instruction. Preinstructional testing was done during the first class session. The authors allotted 60 minutes for the students to finish the two exams – 25 minutes for the Force Concept Inventory and 35 minutes for the Force and Motion Conceptual Evaluation. The time allotted for each exam was determined by dividing by the total number of questions (77) into the total exam time (60 minutes) and then multiplying by the number of questions on the particular exam in question (47 for the Force and Motion Conceptual Evaluation and 30 for the Force Concept Inventory). The latest version of the Force Concept Inventory was used. The version of Force and Motion Conceptual Evaluation used had 43 questions on force and motion topics and 4 questions on the conservation of energy. This yielded an allotted time which was less than that suggested by the exams' authors. However, most students finished both exams and all were given the same amount of time to work on the exams during pre- and post-testing periods. One of the authors was present for every administration of the exam. Post-instructional testing was done after all of the relevant material had been covered in the course. This was approximately two-thirds of the way through the semester, or about ten weeks after pre-instructional testing. In the spring 1998 semester, studio classes were divided into the categories discussed below based on the nature of the instruction they received. The division of students into these categories was essentially a random assignment. Students chose to enroll in a particular section of the course based on scheduling issues and before any information as to which professor would be assigned to teach the class became available. Division of class sections into these categories was based on the section instructor’s willingness to experiment with new methods and materials. groups as an in-class activity. They were then given the last three Interactive Lecture Demonstration sequences. A Cooperative Group Problem Solving model was implemented in sections 4 and 9. Sections 6 and 8 were given context-rich problems as extra, in-class activities on three occasions throughout the semester. Table I summarizes the breakdown of the experimental groups. Table I: Instructional techniques used in experimental sections. Section A. Standard Group Seven class sections comprised the Standard Group. These students were taught the standard studio course. In this model, the first 30 minutes of class was devoted to answering questions and working problems on the board. Then next 1020 minutes were devoted to a brief lecture, complete with derivations and examples. The remainder of class time was used by the students to work on an in-class assignment based on the day's material. The scope of these assignments ranged from pen and paper exercises to spreadsheet exercises to computerbased laboratories. For the most part, students were able to complete the in-class assignments before the class hour was through. Some instructors found activities to occupy the students' time; others simply let students leave early. 4 6 Instructional Techniques Used Incomplete Full ILD ILD CGPS Sequence Sequence X X X 7 8 X X 9 11 X X Table I: Instructional techniques used. Sections referred to as “experimental” are those in which either Interactive Lecture Demonstrations (ILDs), Cooperative Group Problems Solving (CGPS) or both techniques were used. B. Experimental Group Five sections were taught by either Cummings or Marx and comprised the Experimental Group. This group of students had an instructional experience which was predominately the same as that of the standard group. However, these groups were also exposed to Interactive Lecture Demonstrations, Cooperative Group Problem Solving or both. Students in sections 4 and 11 were given all four Interactive Lecture Demonstration sequences. Students in sections 6 and 8 did a simplified version of the "human motion" Interactive Lecture Demonstrations in small 6 Primarily, experimental activities were done in place of the work performed by the standard group. On the occasions that time allowed, they were done in addition to that work. Hence, we estimate that the inclusion of these activities resulted in an increase in instruction time of about 1% for Interactive Lecture Demonstrations and about 5% for Cooperative Group Problem Solving. The additional time came about because the experimental sections did not leave class early, while the standard sections occasionally did. The topics covered by the experimental and standard groups were identical; the two groups remained synchronized throughout the semester and took common exams. 0.6 C. Intermediate Group 0.4 IV. RESULTS AND DISCUSSION Figure 1 provides an overview of the results of this investigation. We have assessed conceptual learning in the standard Studio Physics course during the Spring, 1998 semester to be <g FCI> = 0.18 0.12(s.d.) and <gFMCE> = 0.21 0.05(s.d.). In the studio sections in which Interactive Lecture demonstrations were preformed we found <gFCI> = 0.35 0.06(s.d.) and <gFMCE> = 0.45 0.03(s.d.). In studio sections in which Cooperative Group Problem Solving was used we found <gFCI> = 0.36 and <gFMCE> = 0.36. The fractional gains discussed here and represented by the height of each bar in Figure 1 is the <g> factor discussed in the introduction. In this analysis, we considered only students for which there were both pre- and post-test scores (matched samples). 7 Gain <g> Inclusion of a group of students who could act as an indicator of instructor influence, separating the effects of the instructional techniques and curricular materials from the influence of the instructor (an instructor control group), was not part of this study's design. Nevertheless, such a group serendipitously formed. Section 7 began the semester as a standard section, taught by a professor other than one of the authors. However, due to low enrollment in this class and in section 8 (which ran concurrently in another room), section 7 was merged with section 8 approximately three weeks into the semester. From this point on, these students were taught by one of the authors. Aside from having missed the first three sequences of Interactive Lecture Demonstrations(out of 4), they had an identical educational experience as the students who began the semester in section 8. Hence, we do not consider section 7 to be part of the Standard or Experimental Group, but rather they are a weak control for the influence of the authors on the outcome. We will refer to section 7 as the Intermediate Group. FMCE 0.5 FCI 0.3 0.2 0.1 0.0 Standard Intermediate ILDs CGPS CGPS+ILDs Figure 1: Fractional gain on the Force and Motion Conceptual Evaluation (FMCE) and the Force Concept Inventory (FCI) for groups of students having had various instructional experiences. The fractional gain for a group of students was calculated using the average of post-test scores for the group and the average of pre-test scores for the group. (This is referred to as "calculating the gain on averages".) For two reasons, we chose to calculate average gain in this manner, rather than to average the individual student gains (referred to as “calculating the average of the gains”). First, it allowed us to keep students who achieved a 100% correct score on the pretest in the study. Individual gains can not be calculated for such students, and so they cannot be included in the investigation if one chooses to calculate the average of individual gains. Second, calculating the average in this way reduces the skewing which occurs when students who pre-test with quite high scores then post-test with somewhat lower scores. The error bars shown in Figure 1 represent the standard deviation of the averages of class sections, with each section weighted equally. Every question on the Force Concept Inventory was considered, and equally weighted, in calculation of pre-test and post-test scores on this exam. In contrast, several questions on the Force and Motion Conceptual Evaluation were disregarded in score calculations. The disregarded questions are those which have been found to be important for identifying students’ conceptual models. Hence, they remain a part of the assessment. However, these questions are not appropriate for inclusion in an overall measurement of the level of student understanding. Additionally, several groups of closely related questions on the Force and Motion Conceptual Evaluation are considered as a unit when calculating a total score on the exam. 11,15 This method of calculating a single number score on the Force and Motion Conceptual Evaluation was done on the advice of the exam’s developers and is discussed in detail in ref. 15. Average fractional gain for each section on the Force and Motion Conceptual Evaluation and Force Concept Inventory is shown in Figures 2a and 2b respectively. Standard sections are on the left-hand side while experimental sections are on the right-hand side; between them is the intermediate section (section 7). Standard and experimental group averages for the Force and Motion Conceptual Evaluation and the Force Concept Inventory are indicated by a thin, horizontal line spanning their respective groups. On both exams the average gain for the experimental group was approximately twice that of the standard group. Moreover, the lowest-scoring experimental section was at least one standard deviation away 0.6 0.5 0.4 <g> 0.3 0.2 0.1 0.0 -0.1 1 2 3 5 10 12 13 7 6 8 11 4 9 Section Figure 2A: Fractional gain on the FMCE by class section number. Standard sections are on the left; experimental sections are on the right. The average gain with standard deviation is indicated for each of the two categories. 0.6 0.5 <g> 0.4 0.3 0.2 0.1 0.0 -0.1 1 2 3 5 10 12 13 7 6 8 11 4 9 Section Figure 2B: Fractional gains on the FCI by class section number. Standard sections are on the left; experimental sections are on the right. The average gain with standard deviation is indicated for each of the two categories. 8 Table II: Section-by-section data SECTION 1 2 3 Standard 5 10 12 13 Intermediate 7 6 8 Experimental 11 4 CGPS ILD GROUP 9 N 27 17 27 28 30 23 32 13 39 22 40 30 FMCE Pre Post ave ave 37.3 55.1 25.4 44.1 25.7 39.9 43.2 51.7 31.8 43.0 43.6 55.8 40.9 54.9 34.8 54.3 37.3 63.7 33.0 65.1 29.6 60.7 32.2 67.6 32 40.3 61.6 Pre ave 49.6 43.0 42.1 61.0 42.0 57.5 53.7 56.7 53.3 52.2 46.3 50.1 FCI Post ave 63.9 51.7 54.8 59.0 56.6 63.9 66.2 66.7 69.8 66.4 68.2 66.8 <g> ave 0.28 0.15 0.22 -0.05 0.25 0.15 0.27 0.23 0.35 0.30 0.41 0.33 0.36 33 49.9 68.0 0.36 <g> ave 0.28 0.25 0.19 0.15 0.17 0.22 0.24 0.30 0.42 0.48 0.44 0.52 N 28 18 28 29 31 24 33 13 40 23 41 28 Table II: Section-by-section data for the Force and Motion Conceptual Evaluation (FMCE) and the Force Concept Inventory (FCI). P o st-T est S co re (% ) 100 80 60 40 20 Standard Sections 0 0 20 40 60 80 P re -T e s t S c o re (% ) N= Increased 20% or more Increased from 10% to 20% No change, to increased by 10% Decreased up to 10% 100 184 33.7% 18.5% 31.0% 14.7% 97.8% Figure 3A: Post- versus pre-test score on the Force and Motion Conceptual Evaluation (FMCE) for students in standard sections. The size of the bubble indicates the number of students represented by the point. The lines shown are lines of constant gain. The lowest of the four lines shown corresponds to a gain of –0.20, the line which passes through the origin corresponds to a gain of zero, and the highest line corresponds to a gain of + 0.40. The associated table indicates the percentage of students who increased their exam score by the percentage shown. 10 Post-Test Score (%) 100 80 60 40 20 Experimental Sections 0 0 20 40 60 80 Pre-Test Score (% ) N= Increased 20% or more Increased from 10% to 20% No change, to increased by 10% Decreased up to 10% 100 163 63.8% 18.4% 14.7% 2.5% 99.4% Figure 3B: Post- versus pre-test score on the Force and Motion Conceptual Evaluation (FMCE) for students in experimental sections. The size of the bubble indicates the number of students represented by the point. The lines shown are lines of constant gain. The lowest of the four lines shown corresponds to a gain of –0.20, the line which passes through the origin corresponds to a gain of zero, and the highest line corresponds to a gain of + 0.40. The associated table indicates the percentage of students who increased their exam score by the percentage shown. from the average of the standard group. Table II contains the section-by-section data. It is interesting to note results for the intermediate group and compare them to those for Section 8. Despite the fact that these two groups had the same instructor, and were students in the same class for most of the semester, the intermediate group had conceptual 11 learning gains which were more in line with the standard studio sections than with Section 8 or other experimental sections. Recall that Section 8 had a complete sequence of Interactive Lecture Demonstrations, while the intermediate group did not. Figure 3 shows scatter plots of students' posttest score versus their pre-test score on the Force and Motion Conceptual Evaluation. Figure 3a shows this result for the standard sections and Figure 3b shows this result for the experimental sections. The size of the bubble indicates how many students fell on the same coordinate. The smallest bubble indicates that one student had that set of scores, the next largest bubble means there were two students on that coordinate, and so on. Since there was a smaller number of students in the experimental group, the tables below the graphs indicate the percentage of students that raised (or lowered) their grade by the amount indicated from pre- to post-test. The diagonal lines shown are lines of constant gain. It is apparent from these graphs that students in the experimental group did better in terms of absolute gains on the Force and Motion Conceptual Evaluation. A strikingly similar trend was seen for the Force Concept Inventory results. Furthermore, when we plotted these data for either exam, we found that there were fewer experimental students around or below the zero gain line. There were also many more students in the upper left-hand region (low pre-test scores and gains of over 50%) in the experimental group than in the standard group. Upon viewing the data in Figure 3, we note that weaker students (i.e. students with low pretest scores) benefited from being in the experimental sections. What about the strongest students (i.e. those with high pre-test scores)? Figure 4 is a graph of the average <g> for students divided into groups based on whether their pre-test scores were in the upper, middle or lower third of the entire pool for their group (experimental or standard). Figure 4a represents this result on the Force and Motion Conceptual Evaluation and Figure 4b is for the Force Concept Inventory. These figures clearly show that all students, whether they pre-tested high, middle or low, benefited from these experimental teaching techniques. This result was consistent on both exams. The correlation coefficient between pretest score and gain, for the entire group of students (experimental and standard taken together, N = 347), is -0.06 for the Force Concept Inventory and + 0.16 for the Force and Motion Conceptual Evaluation. Somewhat stronger correlations seem to exist for subsets of the population. V. CONCLUSION Overall this study implies that the standard studio format used for introductory physics instruction at Rensselaer is no more successful at teaching fundamental concepts of Newtonian physics than traditional instruction. The average <g> on the Force Concept Inventory reported here for standard studio sections falls within the range of earlier reported values for traditionally taught courses.4 This result is disappointing in light of the fact that Rensselaer has expended the effort and resources necessary to break-up large (500+ student) classes into small (35-45 student) sections. Rensselaer has introduced group work and computer use as components of in-class instruction. Furthermore, lecture time has been reduced. In general, the Studio Physics classrooms appear to be interactive and students seem to be engaged in their own learning. Nevertheless, use of the studio format alone does not produce improvement in conceptual learning scores as compared to those measured on average in a traditionally structured course. The implication of this study, that ostensibly interactive classrooms do not necessarily result in above average levels of conceptual learning, verifies the work of others. For example, Redish, Saul and Steinberg found that even lectures “with much student interaction and discussion” had very little impact on student learning.16 After lengthy investigations, Kraus reported: In many of our efforts to improve student understanding of important concepts, we have been able to create an environment in which students are mentally engaged during lecture. While we have found this to be a necessary condition for an instructional intervention to be successful, it has not proved sufficient. Of equal importance is the nature of the specific questions and situations that students are asked to think about and discuss.17 Force and Motion Conceptual Evaluation Gain <g> 0.8 0.7 Experimental Sections 0.6 Standard Sections 0.5 0.4 0.3 0.2 0.1 0.0 Bottom Third Middle Third Top Third Figure 4A: Average <g> on the Force and Motion Conceptual Evaluation (FMCE) for students divided into groups based on whether their pre-test scores were in the upper, middle or lower third of the entire pool for their group (experimental or standard). Force Concept Inventory Gain <g> 0.8 0.7 Experimental Sections 0.6 Standard Sections 0.5 0.4 0.3 0.2 0.1 0.0 Bottom Third Middle Third Top Third Figure 4B: Average <g> on the Force Concept Inventory (FCI) for students divided into groups based on whether their pre-test scores were in the upper, middle or lower third of the entire pool for their group (experimental or standard). 13 However, introduction of research-based techniques and activities does have clear beneficial effects. Interactive Lecture Demonstrations generated significant gains in conceptual understanding with remarkably little instructional time. Cooperative Group Problem Solving resulted in similar conceptual learning gains and seemed to also provide a mechanism which fostered improved quantitative problemsolving skills. Students in Cooperative Group Problem Solving sections not only had significant gains on the Force and Motion Conceptual Evaluation and Force Concept Inventory but also performed better on the problem solving section of the last course exam. Nevertheless, implementing Cooperative Group Problem Solving required a semester-long commitment on the part of the instructor. As a result of our investigations, we are optimistic about the future of the Studio Physics program at Rensselaer. The entire infrastructure necessary for true interactivity in the classroom is in place; we feel we need only to adopt researched-based student activities. a) Current address: Department of Physics, University of Oregon, Eugene, Oregon. 1. J. Wilson, "The CUPLE Physics Studio," Phys. Teach., 32 (12), 518-522, (1994). 2. M. A. Cooper, An Evaluation of the Implementation of an Integrated Learning System for Introductory College Physics, Ph.D. thesis, Rutgers, The State University of NJ, 1993. 3. D. Hestenes, M. Wells, and G. Swackhamer, "Force concept inventory," Phys. Teach. 30(3), 141-158 (1992). 4. R. Hake, "Interactive-engagement versus traditional methods: A six-thousand-student survey of mechanics test data for introductory physics courses," Am. J. Phys. 66(1) 64-74 (1998) 5. P. W. Hewson and M. G. A'Beckett-Hewson, "The role of conceptual conflict in conceptual change and the design of science instruction," Instr. Sci. 13, 1-13 (1984). 6. J. Clement, "Using bridging analogies and anchoring intuitions to deal with students' preconceptions in physics," J. Res. Sci. Teach. 30(10), 1241-1257 (1993). 7. K. Cummings and J. Marx were instructors in Studio Physics I for Engineers during the Spring, 1998 semester, and experimented with the use of research based activities in the Studio classroom. Cummings taught sections #4, 9 and 11. Marx taught sections #6 and 8 as well as a weak control group, section #7. 8. See for example, D. Sokoloff and R. Thornton, "Using Interactive Lecture Demonstrations to Create an Active Learning Environment," Phys. Teach. 35(10), 340-347 (1997). 9. P. Heller, R. Keith and S. Anderson, "Teaching Problem Solving through Cooperative Grouping-Part 1:Group versus individual problem solving" Am. J. Phys. 60(7) 627-636 (1992). 10. P. Heller and M. Hollabaugh, "Teaching Problem Solving through Cooperative Grouping-Part 2: Designing problems and structuring groups,”Am. J. Phys. 60(7) 637-644 (1992). 11. R. Thornton and D. Sokoloff, "Assessing student learning of Newton's Laws: The Force and Motion Conceptual Evaluation and the Evaluation of Active Learning Laboratory and Lecture Curricula," Am. J. Phys. 66 (4), 338-352 (1998). 12. R.K. Thornton, D.R. Sokoloff, "Learning motion concepts using real-time microcomputer-based laboratory tools", Am. J. Phys. 58 (9), 858-867 (1990). 13. Mechanics Interactive Lecture Demonstration Package (ILD), Vernier Software, 8565 S.W. BeavertonHillsdale Hwy., Portland, Oregon, 97225-2429, 503-297-5317. 14. “Instructor’s Handbook,” identified as “TA Orientation, School of Physics and Astronomy, Fall 1997. See also http:\\www.physics.umn.edu/groups/physed/ 15. K. Cummings, D. Kuhl, J. Marx and R. Thornton, "Comparing the Force Concept Inventory and the Force and Motion Conceptual Evaluation”, to be submitted, Am. J. Phys. (1999). 16. E. F. Redish, J.M. Saul and R.N. Steinberg, “On the effectiveness of active-engagement microcomputerbased laboratories,” Am. J. Phys. 65, 45-54 (1997). 17. Pamela Ann Kraus, Promoting Active Learning in Lecture-Based Courses: Demonstrations, Tutorials, and Interactive Tutorial Lectures. Ph.D. dissertation, University of Washington, 1997, University Microfilms, UMI Number 9736313. 15