Effect of Students' Seat Location on Programming Course Achievement

9 downloads 62921 Views 124KB Size Report
Traditional, off-line computer science courses, such as a programming ... bility of receiving good grades, whereas those who prefer the back have a higher.
Effect of Students’ Seat Location on Programming Course Achievement Motoki Miura† and Taro Sugihara‡ † Department of Basic Sciences, Faculty of Engineering, Kyushu Institute of Technology ‡ School of Knowledge Science, Japan Advanced Institute of Science and Technology, [email protected], [email protected]

Abstract. Classrooms for computer-related courses are intrinsically larger than regular lecture rooms because of the existence of installed PCs as well as network and outlet facilities. Some of the previous studies have investigated the effect of students’ seat location on their performance, particularly for large class sizes. In this study we examined the effect of seat location in several computer-related courses. We analysed attendance record with seat location, delay caused by tardiness and student score with factors of distance and direction. Two-way ANOVA test revealed the main effect of ‘distance’ in a large classroom. We also confirmed a weak correlation between the score and the delay. Key words: computer science education, seat position, learning space condition, classroom, seat layout

1

Introduction

Traditional, off-line computer science courses, such as a programming course, are often held in a special classroom which provides personal computers (PCs) for learners. Typically, the PCs are utilized to enhance both knowledge and skill through the learners’ experience and play significant roles in the learning process. Nevertheless, in terms of learning space, the placement of PCs restricts the layout and size of the classroom. Even though the latest PC has reduced much in size, network and outlet facilities still occupy the learning space. For this reason, special classroom with PCs are larger than conventional classrooms. Moreover, limitations of the classroom facility and teaching resources often make it difficult to increase the number of courses by dividing them into smaller classes. We considered that the large-sized classroom may decrease the learning effect, because the distance between the teacher and students increases, making it difficult for the student to pay attention to the teacher. In this paper, we report our assessment of the learning effects regarding students’ seat location in large-sized, computer-equipped classrooms.

2

2

Motoki Miura and Taro Sugihara

Related Work

Research activity on enhancing learning of computer science discipline by ICT (information & communication technology) is inherently high because of the affinity. Various computer-aided learning materials and computing augmented learning management systems (CALMSs) have been proposed and evaluated through motivation for better teaching and learning experience [1]. However, there is little research and few surveys that consider both learning space and computer science education. Lynch et al. applied studio-based learning environment for IT related courses [2]. The “studio” is a place where students constantly interact within a group, with their peers and mentors. The capacity of the room is relatively small, and they do not mention the influence of seat location on learning. Pomales-Garcia et al. examined seat arrangement for preventing cheating in exam in terms of human factors [3]. They proposed several non-traditional seat arrangements and concluded that “concentric rectangles” and “look-away” arrangements are better alternatives to the traditional classroom seating. We investigated the effect of seat location during CS lectures. Several studies have been conducted regarding seat location [4–6]. Perkins et al. reported that the initial seat location in their physics course significantly affected student attendance, performance and attitudes [6]. Benedict et al. conducted a study of economics courses within large lecture rooms and reported that students who prefer to sit towards the front of the room have a higher probability of receiving good grades, whereas those who prefer the back have a higher probability of receiving poor grades (D/F) [5]. On the other hand, Buckalew et al. studied nine psychology classes involving over 200 students and reported that the seating position was unrelated to student performance [4]. These studies indicate the possibility of a correlation between seat location and student performance, but the conditions are not fully determined. In addition, there is no investigation for covering computer-aided courses. We analyzed our results to accumulate the instances and contribute to further studies on student seat location.

3 3.1

Method Classrooms

Figure 1 (C-2G) and Figure 2 (C-2B) show the two PC-equipped lecture rooms utilized for this assessment. The width of C-2G was 15.6 m and the depth was 14.7 m. The width of C-2B was 18.1 m and the depth was 11.0 m. The number of PCs for students installed in C-2G and C-2B were 100 and 82, respectively. Two PCs were arranged on each desk, and students sat facing the PCs. Thus, two students (e.g. Nos. 1 and 13) faced each other sitting on opposite sides of a desk. Therefore, they had to turn approximately 90 degrees to face the lecturer and screens in the front of the classroom because PCs were held on individual desks. The Desktop PCs and their monitors were installed such that they could

Effect of Students’ Seat Location on Programming Course Achievement

3

not be moved around. Because of the low ceiling in C-2B, four small projection screens were installed.

Fig. 1. Layout of Room C-2G

3.2

Lecture courses

Table 1 provides a list of lecture courses for this study. We collected attendance records with seat locations from October 2009 to February 2011. The lecture courses are categorized into two types: (1) courses in computer literacy and presentations for freshman and (2) courses in fundamental C programming and Fortran programming with basic numeric analysis for sophomores. In the former type of course, students organized in groups of 4–5 members and collaborated on reviewing projects based on engineering themes and issues. 50% of the score was marked for group reports and presentations and rest was marked individually for personal reports. In the latter course, 60% of the score was marked by mid-term and final exams and the remaining 40% was marked by daily assignments. All programming courses were marked individually. The all-lecture courses were mandatory for graduation. Note that the majority of the students were in the engineering programmes but not all were interested in programming and computer science. The lecture-based courses were conducted by the first author. No TAs (teaching assistants) were employed. Basically, we

4

Motoki Miura and Taro Sugihara

Fig. 2. Layout of Room C-2B

Table 1. Lecture courses attended with student seat location ID 7 8 9 11 12 13 17 18 19

Type Grp Grp Indiv Indiv Indiv Indiv Indiv Grp Grp

Content Term Room Num Max Computer Literacy Oct’09 C-2G 82 95 & Presentation –Feb’10 C-2G 66 96 Fortran Programming (2009/2nd) C-2G 82 100 C Programming Apr’10 C-2B 54 100 C Programming –Jul’10 C-2B 82 100 C Programming (2010/1st) C-2G 67 100 Fortran Programming Oct’10 C-2G 80 100 Computer Literacy –Feb’11 C-2G 72 92 & Presentation (2010/2nd) C-2G 64 93

Avg 88.9 88.4 78.5 84.8 84.3 84.3 74.8 83.8 84.4

SD 4.04 11.9 13.1 15.5 18.5 21.1 22.8 9.54 9.26

did not designate specific seats for the students, and they freely chose their seats for each lecture. Therefore, seat selection may be affected by the students’ individual volitions and motivations. The lecture room, number of students and statistics regarding their final score are displayed in Table 1. Note that the standard deviation for the programming course scores was larger than that for the computer literacy course scores. Because the score of the computer literacy course was evaluated through presentations, the standard deviations decreased. 3.3

Obtaining Seat Locations

At the beginning of each lecture, students were asked to submit an attendance record by a prepared Linux command on each console. The command invoked a Java Web Start programme which sent an IP address and a student ID to an attendance server. The IP address for each student PC was assigned by a DHCP server which was configured to assign fixed and constant IP addresses to each

Effect of Students’ Seat Location on Programming Course Achievement

5

PC. Thus, the teacher could obtain seat locations of all students who attended the classes as well as the time when each attendance was submitted. Incidentally, the students could browse their attendance record by accessing the attendance server. Most of the seat locations in the attendance record were valid; however, some exceptional data were eliminated by the following cases. First, we removed records of students who retook the course. Second, we removed data for withdrawing students. Third, outlying attendance data that were reported before 30 minutes or after 60 minutes of the lecture start time were eliminated. Finally, attendance data at mid-term and final exams were removed because we strictly specified the seat positions in these exams.

4

Assessment & Result

We set following two hypotheses. 1. Distance from lecturer’s podium to student’s seat affects their grade. Grades of students sitting in front of classroom are higher than those of students sitting in the rear because of their decreased motivation to pay attention to lectures. 2. The facing direction of students affects their grades. Grades of students who sit with their back to the lecturer are lower than grades of those who face the lecturer. It also affects the degree of attention paid to the lecture by the students. To verify the hypotheses, we analysed the attendance records in the following steps: (1) prepare a set of course-type, user, room, seat, seat-group, final score of the user, delay (in seconds) of attendance record since lecture-start time data for all 7,489 attendance items, (2) calculate count, average and standard deviation (SD) for each course-type, room and seat-group. The course-type means the programming/computer literacy & presentation described in 3.2. Since the student seats were not fixed, we only adjusted the final score of the student to the position for each of the attendance records. Table 2 and 3 show the analyzed data in programming courses. Each box represents the seat-group, corresponding to the box of Figure 1 (C-2G) and Figure 2 (C-2B). Note that the boundary of the box is not apparent in the classroom desks, and students were not aware of these boundaries. Numbers in bold font show a maximum and those in underlined bold font indicate a minimum. From Table 2, it is clear that many students preferred the seats in the rear of the classroom; however, from Table 3, this trend was not observed in C-2B room. One of the reasons is that the entrance of C-2G room was located at the back of the room, whereas that of C-2B was at the front (see Figure 1 and 2). Another possibility is the depth of room C-2B (11.0 m) which was lesser than that of C-2G (14.7 m).

6

Motoki Miura and Taro Sugihara Table 2. Summary of programming course at room C-2G [seat group ID] num of attendance avg of score SD of score avg of behind time(s) SD of behind time(s) [seat group ID] num of attendance avg of score SD of score avg of behind time(s) SD of behind time(s)

[17] 194 82.71 12.87 141 850 [7] 175 87.08 10.59 285 821

[18] [15] [16] [14] [13] [12] [11] 197 190 175 141 166 160 175 81.13 82.11 81.42 82.87 83.89 81.82 85.14 9.89 9.49 8.18 11.56 9.39 11.27 12.86 519 234 142 378 195 313 113 838 702 661 763 761 859 745 [8] [5] [6] [4] [3] [2] [1] 129 133 91 65 135 52 176 82.52 79.74 82.48 76.78 86.86 81.9 83.74 13.75 14.37 11.64 15.13 13.98 13.33 15.68 265 164 123 527 170 612 178 724 601 694 988 720 1059 877

Table 3. Summary of programming course at room C-2B [seat group ID] num of attendance avg of score SD of score avg of behind time(s) SD of behind time(s) [seat group ID] num of attendance avg of score SD of score avg of behind time(s) SD of behind time(s)

[19] 76 91.49 11.75 42.1 415 [9] 75 86.72 14.77 140 508

[20] [17] [18] 68 74 69 93.47 90.39 83.12 8.09 11.83 11.76 138 -136 117 524 338 430 [10] [7] [8] 35 68 55 89.17 86.38 83.73 9.68 13.4 13.33 242 -96.12 184 469 281 440

[15] 59 86.42 10.81 -6.68 402 [5] 68 87.18 10.75 101 427

[16] [14] [13] [12] [11] 53 58 80 45 74 78.34 84.05 86.4 90.02 89.19 10.3 9.65 8.58 6.23 8.81 298 77.8 53.1 -11.9 -67.2 627 407 495 474 412 [6] [4] [3] [2] [1] 58 77 76 57 76 83.6 89.82 85.76 86.35 91.41 14.83 8.86 11.78 9.17 5.76 112 -49.5 -16.6 -59.0 39.4 458 214 491 299 388

Seat-groups with minimum attendance were [2] in Table 2 and [10] in Table 3. The reason is not clear, but students may dislike the areas because the facing direction is not consistent with the direction of the lecturer and projection screens. In terms of average score, we analysed the data by two-way ANOVA (two distances from front end × two directions of seats). Main effect of “distance” was statistically significant in the C-2G room (F [1, 6184] = 4.49, p = .034). Thus, our first hypothesis was supported; however, in terms of C-2B room, no effects were significant. The average score of seat-group [4] of the room C-2G was low. This phenomenon was caused by students’ anxiety for academic credit. Some students who recognized their insufficient comprehension sat near the podium; however, similar tendencies were not observed from the data in room C-2B. We calculated the average of delay in attendance reporting (tardiness) for each seat group. The ‘delay’ refers to the offset time of attendance submission from the beginning of the lecture in seconds. From the data, some characteristics of the students’ seating can be represented. Two-way ANOVA revealed that the

Effect of Students’ Seat Location on Programming Course Achievement

7

main effect of direction was statistically significant across distances in the C-2G room (F [1, 6184] = 8.38, p = .004). No other effects were confirmed. Regarding the C-2B room, similar tendencies were observed (F [1, 1297] = 2.18, p = .140) even though they were not significant. We considered that the tardy students had to sit in the remaining disadvantaged seats. Pearson’s correlation of the student’s final score with the delay in the programming courses was -0.33 (N = 344). Thus a weak negative correlation between delay and final score in programming course was confirmed. The result is reasonable because the amount of delay is inversely proportional to the amount of explanation and instruction that the student could accept. We also assessed the course-type of computer literacy and presentation. Preferential treatment to the back-seats was similar to the programming course, but no significant difference in scores among the seat-groups was observed. We consider that this similarity is caused by the following factors: (1) the amount of explanation was shorter than that in the programming course, (2) group work caused a uniform seat position effect and (3) originally, the standard deviation of the computer literacy course score was limited more than that of the programming course (described in 3.2). Therefore, we only assessed the data of the programming course in the following parts. Grouping by seat factors To browse and assess the data with two factors, facing direction and distance, which are posed in our hypotheses, we re-grouped by seat positions. Table 4 shows results of the re-grouping for groups on the left side of Table 2 when they were bundled with groups on the right side. The odd-numbered seat groups have advantage in facing direction over even-numbered seat groups in C-2G. Table 5 shows results of further bundling of the seat groups, with each seat group separated into the front and back. Table 6 and 7 also represent results of similar bundling to C-2B room. Numbers in bold font show a maximum and those in underlined bold font represent a minimum. The results in C-2G room indicate more variances than those of C-2B room. The phenomenon may be caused by differences in the depth of the room and the visibility of the projection screen.

5

Conclusion and Future Work

Computer-equipped classrooms are intrinsically larger than regular classrooms because of the existence of installed PCs and network/outlet facilities. Previous studies investigated the influence of students’ seat location on their performance, especially for large class sizes. In this paper, we examined the effect of seat location in several computer-related courses. We summarized the stored attendance records and students’ final scores, analyzing the scores in terms of both distance and facing direction. We observed that the data represented the student’s seating preference, and some interesting tendencies regarding seating positions and delay in registering attendance. Because two-way ANOVA revealed that the main

8

Motoki Miura and Taro Sugihara

Table 4. Summary of programming course at room C-2G: Groups on the left of ‘Table 2’ were bundled to those in the right [seat group IDs] [14,16] [13,15] [12,18] [11,17] num of attendance 316 356 357 369 avg of score 82.06 82.94 81.44 83.86 SD of score 9.86 9.48 10.54 12.92 [seat group IDs] [4,6] [3,5] [2,8] [1,7] num of attendance 156 268 181 351 avg of score 80.11 83.32 82.34 85.4 SD of score 13.5 14.62 13.63 13.49 Table 5. Summary of programming course at room C-2G: Re-grouping by seat factors of distance and direction [seat group IDs] [12,14,16,18(back)] [11,13,15,17(back)] num of attendance 360 372 avg of score 81.52 82.55 SD of score 10.61 11.62 [seat group IDs] [12,14,16,18(front)] [11,13,15,17(front)] num of attendance 313 353 avg of score 81.98 84.31 SD of score 9.77 11.03 [seat group IDs] [2,4,6,8(back)] [1,3,5,7(back)] num of attendance 173 341 avg of score 83.51 84 SD of score 10.04 13.7 [seat group IDs] [2,4,6,8(front)] [1,3,5,7(front)] num of attendance 164 278 avg of score 78.99 85.12 SD of score 16.26 14.39

effect of “distance” is significant in C-2G room, we verified our first hypotheses regarding distance. Our result suggested that room factors such as depth, size and visibility of projection screens can influence student achievements. Our findings can be applied to consider a better layout design of computer-equipped class-rooms and equalize learning conditions and opportunities. However, the “free-seat” condition did not control effects of student levels and motivations; we will consider these effects in our future work as we continue to analyze attendance records and build models that contribute to enhanced learning and education.

Acknowledgment We would like to thank all students attended to the lecture courses. Our research is partly supported by a grant-in-aid for Scientific Research (22650204).

Effect of Students’ Seat Location on Programming Course Achievement

9

Table 6. Summary of programming course at room C-2B: Groups on the left of ‘Table 3’ were bundled to groups on the right [seat group IDs] [15,16] [14,18] [13,17] [12,20] [11,19] num of attendance 112 127 154 113 150 avg of score 82.6 83.54 88.32 92.1 90.35 SD of score 11.31 10.86 10.46 7.6 10.47 [seat group IDs] [5,6] [4,8] [3,7] [2,10] [1,9] num of attendance 126 132 144 92 151 avg of score 85.53 87.28 86.06 87.42 89.08 SD of score 12.91 11.35 12.58 9.46 11.42 Table 7. Summary of programming course at room C-2B: Re-grouping by seat factors of distance and direction. We omitted the max/min expression because the average scores were almost the same. [seat group IDs] [12,14,18,20(back)] [11,13,15,16,17,19(back)] num of attendance 117 214 avg of score 87.97 87.58 SD of score 9.96 10.8 [seat group IDs] [12,14,18,20(front)] [11,13,15,16,17,19(front)] num of attendance 123 202 avg of score 87.19 87.44 SD of score 10.75 11.5 [seat group IDs] [2,4,8,10(back)] [1,3,5,6,7,9(back)] num of attendance 106 227 avg of score 86.98 86.1 SD of score 11.61 12.81 [seat group IDs] [2,4,8,10(front)] [1,3,5,6,7,9(front)] num of attendance 118 194 avg of score 87.66 88.02 SD of score 9.62 11.78

References 1. Guido R¨ oßling, Mike Joy, Andr´es Moreno, Atanas Radenski, Lauri Malmi, Andreas Kerren, Thomas Naps, Rockford J. Ross, Michael Clancy, Ari Korhonen, Rainer ´ Oechsle, and J. Angel Vel´ azquez Iturbide. Enhancing learning management systems to better support computer science education. SIGCSE Bull., 40:142–166, November 2008. 2. Kathy Lynch, Angela Carbone, David Arnott, and Peter Jamieson. A studio-based approach to teaching information technology. In Proceedings of the Seventh world conference on computers in education conference on Computers in education: Australian topics - Volume 8, CRPIT ’02, pages 75–79, Darlinghurst, Australia, Australia, 2002. Australian Computer Society, Inc. 3. Cristina Pomales-Garcia, Hector J. Carlo, Tanya M. Ramos-Ortiz, Iza M. FigueroaSantiago, and Sharlyn Garcia-Ortiz. Non-traditional exam seat arrangements. Computers & Industrial Engineering, 57(1):188 – 195, 2009. Collaborative e-Work Networks in Industrial Engineering.

10

Motoki Miura and Taro Sugihara

4. L. W. Buckalew, Jerry D. Daly, and K. E. Coffield. Relationship of initial class attendance and seating location to academic performance in psychology classes. Bulletin of the Psychonomic Society, 24(1):63–64, January 1986. 5. Mary Ellen Benedict and John Hoag. Seating location in large lectures: Are seating preferences or location related to course performance? The Journal of Economic Education, 35(3):215–231, 2004. 6. Katherine K. Perkins and Carl E. Wieman. The surprising impact of seat location on student performance. The Physics Teacher, 43:30–33, January 2005.