Oldal 40 [40]
OCR
3.4 THE EXPERIMENT 39 — We muddled the lines of a given problem’s algorithm, and the students had to establish the correct line order (3 points). — Students had to determine the loop skeleton of a difficult problem’s algorithm (without explicitly writing the algorithm) (2 points). — Students had to check the loop skeleton of a fairly difficult problem’s algorithm and then write explicitly the C/C++ code that resolves the problem (2 points). — The 10* point was received for filling in the test. 3.4.1 Results and discussion The diagrams below show the results of the experiment (see figures 3.6-7). Students (identified by their position in the sorted list of the group they belong to) are represented on the horizontal axis, and their results on the vertical axis. The white and grid columns represent the first semester average (End_of_semester_mark) of the experimental and the control group, respectively, and the grey and black columns represent their test results after the experiment (Number_of_test_points). The means of the points the students belonging to the experimental (12 + 12 members) and control (12 + 12 members) groups received for the test are 6.27 and 5.02 respectively. Comparing these values with the independent samples’ t-test, we found that the difference between them is significant (p = 0.038 < 0.05) (favouring the experimental group). As we can discern from the figures, the points the students received for the test are consistently below the marks they had on the basis of their whole semester performance. This was a direct consequence of the nature of the test. Although the lower results as a phenomenon can be observed in the case of all students, the size of these differences (End_of_semester_mark —Number_of_test_ points) varies from student to student. Since we expected that the new method was going to enhance the students’ abilities to apply their knowledge even in “unfamiliar circumstances”, we also decided to analyse these differences. The means of the differences for each group in part are: 2.56 (control group in class H), 1.53 (experimental group in class H), 3.21 (control group in class G), 1.75 (experimental group in class G). Comparing these values with the independent sample t-test (in the case of each experimental-control group pair), we found that the differences are significant (favouring the experimental groups): for class H, p = 0.016; for class G, p = 0.002. Then again, when comparing the performances of the two control groups or the two experimental groups, we did not receive significant differences. (The better test results in the case of the experimental groups cannot be explained by the contingent differences between the teachers.) In conclusion, it can be stated that the results of this didactical experiment support our expectation that the multi-sensory method presented above improves students’ skill to analyse and design algorithms.
