Learning JAVA Programming in Teams Huey-Wen Chou National Central University Taiwan hwc928@gmail.com Abstract In educational literature, close relationships among instructional design, cognitive load, and learning performance have been found. Because of the differences in personal characteristics between learners, instructors have to adjust instructional design accordingly in order to promote learners’ learning performance. This study employed JAVA programming language as learning subject to investigate whether different grouping types and prior knowledge level make differences in learners’ learning performance by a field experiment. The study followed a 2 × 2 quasi-experimental design with two independent variables “prior knowledge” (high and low) and “grouping types” (homogeneous and heterogeneous). The dependent variable was “learning performance” which includes cognitive and affective domains. A six-week field experiment including 134 undergraduates majoring in information management was conducted. The results are as following: 1. Subjects who have different prior knowledge level have significant difference in learning performance. 2. Female learners scored significant higher in learning enjoyment. According to the results above, implications for future research and directions were provided at the end. Keywords: prior knowledge, grouping types, learning performance, learning enjoyment, JAVA 1 1. Introduction The pedagogical practice of group projects helps students to learn how to corporate with others (Gallavan & Juliano, 2007). Team-based project requires some degree of cognitive collaboration (Schellens & Valcke, 2005). Cognitive collaboration within group members, people can not only mutually support others (Van Boxtel, Van der Linden, & Kanselaar, 2000) but also learn the knowledge from group members and therefore results in an effective learning as well as accomplishes more complicated tasks (Laughlin, Bonner, & Miner, 2002). Therefore, it’s important to foster students to have positive attitude toward teamwork. In this regard, this study investigates how team-based learning process affects individual learning performance as well as learner’s attitude toward teamwork. In the field of information technology discipline, programming language is one of the important training targets. Successful mastery of different programming languages requires the abilities of logical thinking and problem solving, as well as the knowledge of programming command. Each person has his or her own approach in understanding and analyzing the problem and evolving a solution plan. The way that a student writes computer commands, algorithms and scripts reflects his (her) unique cognitive process. These cognitive activities will be manifest through each person’s computer code writing (Choi, Deek, & Im, 2009). Among these programming languages, JAVA is one of the most widely adopted languages in writing the Internet based business computer applications. This study therein employs JAVA programming language as the learning subject to investigate whether learners of different prior knowledge levels when grouped in homogeneous or heterogeneous learning environment will have different learning performances. Gagné (1992) indicated that there are many factors that can influence individual learning process and outcomes. External factors include instructional design and learning environment whereas internal factors are personal characteristics such as motivation, prior knowledge, cognitive style, and apprehension. Prior knowledge and learning environment were two variables to be manipulated in this study. This study employs prior knowledge to group learners to homogeneous and heterogeneous groups in which learners of homogeneous group learners are having similar prior knowledge level and learners of heterogeneous group learning are having different prior knowledge level. 2. Research Method This study focuses on two different levels of prior knowledge (high and low) and 2 grouping types (homogeneous and heterogeneous) and investigates how they differently impact on learning performance. Independent variables include prior knowledge and grouping types. Dependent variable includes learners’ learning performance after a series experiment of collaborative learning. Learning performance will be measured by two domains: cognitive and affective. ( 1 ) Prior knowledge Dochy et al. (1996) described that before taking lessons, learners already possessed specific knowledge related to the learning subjects which is called prior knowledge. In this study, C programming language was chosen to be the prior knowledge to JAVA programming language. So, this study employed subjects’ previous C programming language as the indicator of their prior knowledge. Prior knowledge was divided into high prior knowledge and low prior knowledge based on the average score of C programming language class. The cutting score is 75. The scores over 75 were classified as high prior knowledge, and the rest others were classified as low prior knowledge. ( 2 ) Grouping types According to the level of prior knowledge (high prior knowledge and low prior knowledge), this study randomly grouped the subjects with similar prior knowledge level into homogeneous groups, and those with different prior knowledge into heterogeneous groups. The homogeneous prior knowledge group includes two types: high and low. (3) Learning performance Learning performance refers to knowledge, skills, abilities and attitudes that learners will attain after learning (Gordon, 2010). From literatures review, there are three educational objectives about learning performance. The first one is cognitive domain. It refers to cognition and understanding about learning objectives. The second is affective domain which is about learning attitude, liking, or impression. This study measured subject’s learning enjoyment after the experiment as the indicator of affective domain of learning performance. In order to eliminate the effects of pre-test scores from post-test scores, this study employs the pre-test scores as a covariate variable. This study proposes the following two hypotheses: Hypothesis 1. Learners with the same prior knowledge level but are in different grouping types will have significant different learning performance and learning enjoyment. Hypothesis 2. There are significant gender differences in learning performance and enjoyment. 3 The participants in this study were 134 undergraduates majoring in information management from a university in Taiwan, with 87 males and 47 females. All participants have already studied C programming language for one semester, and possessed basic logical concepts, related knowledge, and programming skills. The students were randomly grouped with 3 members in each group based on their prior knowledge level. A total of 45 groups was made and was employed in the field experiment. The learning material in this experiment was selected topics of JAVA programming language focusing on control structure, specifically, control statement and looping statement. The control statements included “if…else” and “switch case” two control statements, and “while”, “do…while”, and “for” three looping statements. This course consists of a three-hour lecture and a two-hour laboratory session in each week. These topics are very fundamental and important to programming language. The study followed a 2 × 2 quasi-experimental design with two independent variables “prior knowledge” and “grouping types”. Prior knowledge was separated into high prior knowledge and low prior knowledge and grouping types contained homogeneous group and heterogeneous group. A total of forty-five groups with 29 homogeneous groups (including 14 groups with three low prior knowledge members, and 15 groups with three high prior knowledge members) and 16 heterogeneous (mixed) groups (consisting of one/two high prior knowledge and two/one low prior knowledge members in each group) except the last group with two students only. The data of two-member group was not used for further analysis. The duration of the experiment was six weeks (including two weeks for pre-test and post-test), from March 2, 2011 to April 13, 2011. In the first week, the researcher introduced the purposes, procedure, and schedule to the experiment. A pre-test on JAVA knowledge was administered to all participants afterwards. The test scores would be contrasted to the post-test scores. The experiment lasted for four consecutive weeks and was held in the two-hour laboratory session with each included two major activities. The first activity was a group-based collaborative learning process to solve a learning task which was related to the last week’s instruction. This task could engage students in communication with team members and collaboration with each other to effectively solve the task. The solution to the programming task should be typed in computer and compiled after group collaborative discussion. After finishing the task, each subject filled out the questionnaire which included questionnaire measuring on learning enjoyment and personal profile. Weekly learning achievement test was administered on the first twenty 4 minutes of the lecture class in each of the four experimental weeks. And the achievement test items are related to the laboratory collaborative task and class instruction given in the previous last week. The post-test was conducted in the sixth week to examine the treatment effect. Learning performance was operationalized by the post-test scores with pre-test scores as covariate. 3. Results The descriptive statistics of four-week achievement test scores revealed that high level of prior knowledge learners could get higher score ( ranged from 83.50 to 96.89) than low level of prior knowledge learners ( ranged from 71.50 to 87.71) on average. The descriptive statistics of learning enjoyment for collaborative learning in laboratory class revealed that different types of group were all pretty high ( ranged from 3.94 to 4.36). But high prior knowledge learners in homogeneous groups changed dramatically and became much more joyful ( were from 3.94 to 4.27). Two-way ANCOVA with pre-test scores as the covariate and learning performance as the dependent variable was employed to test H1. Pre-test score had a significant effect on learning performance (p value of 0.012). Learners with different level of prior knowledge had significant different scores in learning performance (p value of 0.014). The average score for high prior knowledge learners and low prior knowledge learners were 87.87 and 77.59 respectively. Nevertheless, different grouping types did not have significant impact on subject’s learning performance but in learning enjoyment. Regarding the gender difference, the results revealed gender did not make significant differences in learning performance (male: = 83.45; female: did have significant differences in learning enjoyment (male: 4.38; p < 0.01). = 82.53; p > 0.1) but = 3.92; female: = 4. Conclusion and Implications This study applied two different levels of prior knowledge and three grouping types to conduct a six-week experiment and investigated how the above factors impact on learning performance. This study found that there were significant differences in learning performance between different prior knowledge levels. This finding is consistent with Dochy’s (1996), Shin, Schallert, and Savente’s (1994) and Alexander’s (1996) research that prior knowledge could affect learning outcome. Different grouping types did not have significant differences in learning 5 performance. But most learners in any grouping types could gradually perform better. This implies that teamwork is in fact beneficial to all participants no matter what groupings they are in. Grouping types did not have significant differences in learning enjoyment either. Mixed groups perceived higher learning enjoyment than either high prior knowledge or low prior knowledge groups do. In addition, there were no significant gender differences in learning performance. Nevertheless, this study found that female students obtained significant higher learning enjoyment than male students did. Implications There is no doubt that prior knowledge had an effect on learning performance. This study confirmed that learners with different prior knowledge need different types of grouping. In specific, arranging low level of prior knowledge learners in heterogeneous groups could help those learners attain helps from high level of prior knowledge learners. Although enhancing learners’ performance in cognitive and psychomotor domains is important, affective domain of learning performance cannot be ignored during the entire collaborative learning process. This study found that female participants had higher learning enjoyment during collaborative JAVA learning process. Therefore, mixing genders in groups may help learners attain better learning enjoyment. Lastly, this study found that low prior knowledge learners who were grouped together would spend more time in lab tasks. It is suggested that low prior knowledge learners be grouped with more capable learning partners so that they would have more knowledge source to learn from. Research Limitations The first limitation is that this study used convenient sample to conduct experiment. Because of the space constraints, only 44 groups were included in this study. It is suggested that future study replicates the experiment with a larger sample size. Secondly, this study employed a six-week experiment to collect data which may not be long enough to foster a sense of stable collaborative atmosphere. In addition, this study only analyzed quantitative data, some group dynamics phenomena might not be reflected on a quantitative-based way. These qualitative phenomena could only be observed during the experiment. It is suggested that in the future, researchers can take qualitative approach to complement quantitative results, such as record and analyze the 6 conversation contents, observe group dynamics, and conduct in-depth interviews with participants. References Alexander, P. A. (1996). The past, present, and future of knowledge research: A reexamination of the role of knowledge in learning and instruction. Educational Psychologist, 31, 89-92. Choi, K.S., Deek, F.P., & Im, Il. (2009). Pair dynamics in tem collaboration, Computers in Human Behavior, 25, 844-852. Dochy, F. J.R.C. (1996). Prior knowledge and learning. In T. Husen & T. N. Postlethwaite (Eds.). International encyclopedia of education (2nd ed., pp. 4698-4702). Oxford/New York: Pergamon. Gagné, R. M. (1992). Tryout of an organizing strategy for lesson design: Maintenance procedure with checklist. Brooks AFB, TX: Armstrong Laboratory, Human Resources Directorate. Gordon, A. (2010). Writing instruction objectives and tests. 10/10/2011 Retrieved from World Wide Web: http://instructionaldesign.gordoncomputer.com/Objectives.html. Laughlin, P. R., Bonner, B. L., & Miner, A. G. (2002). Groups perform better than the best individuals on letters-to-numbers problems. Organizational Behavior and Human Decision Processes, 88, 605–602. doi:10.1016/S0749-5978(02)00003-1. Shin, E., Schallert, D., & Savenye, W. (1994). Effects of learner control, advisement, and prior knowledge on young students' learning in a hypertext environment. Educational Technology Research and Development, 42(1), 33-46. doi: 10.1007/bf02298169 Van Boxtel, C. A. M., Van der Linden, J., & Kanselaar, G. (2000). Collaborative learning tasks and the elaboration of conceptual knowledge. Learning and Instruction, 10, 311–330. doi: 10.1016/S0959-4752 (00)00002-5. 7