| Summary: | 碩士 === 國立臺灣師範大學 === 資訊教育學系 === 94 === In this research we investigated if the use of textual or iconic programming languages would result in different learning effects. Fifty-two six graders from a local elementary school participated in this research. A class of 26 students learned to program in MSWLogo and the other class of 26 students learned Drape. Both MSWLogo and Drape are Logo-style languages; however, programmers use textual commands to program in MSWLogo while iconic commands are used in Drape.
The quantitative data collected in this research included grades of five paper-based tests and a hands-on programming test. Each paper-based test included questions assessing such cognitive abilities as knowledge, comprehension, application and analysis. The programming test consisted of 24 problems: 16 of them required students to write complete programs while the remaining eight problems were for students to modify existing programs. Analyses of test grades revealed that students of the MSWLogo group performed significantly better than those of the Drape group in the first three paper-based tests. However, there was no significant difference between the two groups in the last two tests. It seems that there is a steeper learning curve for students to program in Drape (iconic language) than in MSWLogo (textual language). Nevertheless, when students become more and more familiar with language syntax, the difference diminishes. The hands-on programming test did not show a significant difference between performances of the two groups, whether with write-complete-program problems or with modify-program problems.
The qualitative data collected in this research included automatically captured screen logs of students’ programming processes while they were working on the programming test, results of a questionnaire survey conducted at the end of the experiment, and journals kept by students and the instructor. A qualitative analysis of these data indicated that students of the MSWLogo group made syntactic errors frequently while students of the Drape group were often confused by icons that were not intuitively clear or those icons that looked similar to each other. Apparently, each language has its own disadvantages in usability. With regard to students’ attitudes toward programming, there were no significant differences between the two groups in how much students liked programming or if they would like to learn to program again in the future.
To sum up: This research concludes that there are no differences between the two groups of six graders who learned MSWLogo (textual language) and Drape (iconic language) respectively, whether in students’ performance or their attitudes toward programming. Many people may tend to think that iconic languages are easier to learn and would be preferred by elementary school students. Our findings may help to shed such a delusion and provide some useful hints to primary school teachers when it comes to select a suitable programming language for their students to learn.
|
|---|
