 |
Asia-Pacific Forum
on Science Learning and Teaching, Volume 9, Issue 2, Article 3
(Dec., 2008) |
Based upon the three main research questions investigated in this study, all the learning outcomes revealed that CA + modeling ability instruction could improve students’ modeling ability, in terms of general perspective or context-based. Besides, the concept test also showed the results of highly increased enhancement after CA + modeling ability instruction. However, one remarkable result disclosed that students’ performance regarding general modeling ability, concept learning and context-based modeling ability were all improved significantly in the group receiving modeling ability only instruction. This discovery brings us hope of cultivating students modeling ability by a simple and efficient modeling ability instruction. The detailed outcomes are presented as follows.
The performance of general modeling ability between different instructional designs
There are three dimensions embedded in general modeling ability questionnaires, which include ontology, epistemology and methodology. Regarding the results from the three different instructional designs, it was discovered that all three groups of students had no difference in terms of these three dimensions of general modeling ability before instruction, but performances were all enhanced significantly after instruction. From the ANCOVA analysis, it was revealed a significant effect for different instructions with regard to ontology (F [2, 145] = 7.533, p < .05), epistemology (F [2, 145] = 16.147, p < .05) and methodology (F [2, 145] = 9.199, p < .05). The detailed result is presented in Table 3.
Moreover, a Scheffe post-hoc test indicated that students in the both groups of CA + modeling ability and modeling ability only instructions did significantly better than the control group (p < .05) in all three dimensions of general modeling ability. However, there were no significant differences between the CA + modeling ability and the modeling ability only instructions in all dimensions of ontology (p = .975), epistemology (p = .784) and methodology (p = .357) (Table 4).
Table 3. The results of pre and post-tests of the general modeling ability test from different instructions.
Three dimensions of general modeling ability Three different instructions Mean df 1 df 2 F value Sig. Pre-test Post-test Ontology
CA + modeling ability
3.01 (S.D.=0.30)
3.31 (S.D.= 0.36)
2
145
7.533
0.001*
Modeling ability
3.06 (S.D.=0.63)
3.30 (S.D.= 0.34)
Control
2.96 (S.D.=0.31)
3.02 (S.D.= 0.38)
Epistemology
CA + modeling ability
3.02 (S.D.=0.27)
3.41 (S.D.= 0.40)
2
145
16.147
0.000*
Modeling ability
2.95 (S.D.=0.33)
3.35 (S.D.= 0.42)
Control
2.96 (S.D.=0.38)
2.29 (S.D.= 0.68)
Methodology
CA + modeling ability
3.22 (S.D.=0.29)
3.48 (S.D.= 0.38)
2
145
9.199
0.000*
Modeling ability
3.11 (S.D.=0.33)
3.37 (S.D.= 0.43)
Control
3.14 (S.D.=0.35)
3.13 (S.D.= 0.42)
* The mean difference is significant at the 0.05 level.
Table 4. The post-hoc analysis of students’ performance from the post-test of general modeling ability among three different instructional groups
Three different instructions Three dimensions of general modeling ability Ontology Epistemology Methodology CA + modeling ability
Modeling ability
0.975
0.784
0.357
Control
0.001*
0.000*
0.001*
Modeling ability
CA + modeling ability
0.975
0.784
0.357
Control
0.003*
0.001*
0.028*
Control
CA + modeling ability
0.001*
0.000*
0.001*
Modeling ability
0.003*
0.000*
0.028*
* The mean difference is significant at the 0.05 level. Concept test performance between different instructional designs
In terms of the concept test, it was disclosed that all three groups of students had no difference before instruction, but the performances were all highly improved after instruction with the significant differences between the groups. From the ANCOVA analysis, it showed a significant effect for different instructions for the concept test (F [2, 145] = 21.593, p < .05) (Table 5.) Besides, a Scheffe post-hoc test indicated students in the both groups of CA + modeling ability and modeling ability only instructions did significantly better than the control group (p < .05). Still, there was no significant difference between CA + modeling ability and modeling ability only instructions (p = .314) (Table 6).
Table 5. Pre- and post-test results of the concept test from three groups of students
Three different instructions Mean df 1 df 2 F value Sig. Pre-test Post-test CA + modeling ability
1.97 (S.D.= 1.36)
7.16 (S.D.= 1.89)
2
145
21.593
0.000*
Modeling ability
2.02 (S.D.= 0.88)
7.58 (S.D.= 1.12)
Control
1.87 (S.D.= 0.57)
5.40 (S.D.= 1.10)
* The mean difference is significant at the 0.05 level.
Table 6. The post-hoc analysis of students’ performance on concept tests among three different instructional groups
Three different instructions Sig. CA + modeling ability
Modeling ability
0.314
Control
0.000*
Modeling ability
CA + modeling ability
0.314
Control
0.000*
Control
CA + modeling ability
0.000*
Modeling ability
0.000*
* The mean difference is significant at the 0.05 level. The performance of the context-based modeling ability between different instructional designs
Regarding the performance of the context-based modeling ability from the results, it was indicated that all three groups of students had no difference before instruction, but the performance was all significantly better after instruction. The ANCOVA analysis pointed out a significant effect among different instructions from the context-based modeling test (F [2, 145] = 5.383, p < .05) (Table 7.) Besides, a Scheffe post-hoc test displayed that those students in the both groups of CA + modeling ability and modeling ability only instructions performed significantly better than the control group (p < .05). Yet, no significant difference between CA + modeling ability and modeling ability only instructions (p = .695) (Table 8). Besides, most students still lack the context-based modeling ability after modeling ability instruction.
Table 7. Pre- and post-test results of context-based modeling ability among three groups of students
Three different instructions Mean df 1 df 2 F value Sig. Pre-test Post-test CA + modeling ability
0.84 (S.D.= 0.79)
1.47 (S.D.= 1.04)
2
145
5.383
0.006*
Modeling ability
0.67 (S.D.= 0.81)
1.32 (S.D.= 1.07)
Control
0.77 (S.D.= 0.63)
0.77 (S.D.= 0.50)
* The mean difference is significant at the 0.05 level.
Table 8. The post-hoc analysis of students’ performance on context-based modeling test among three different instructional groups.
Three different instructions Sig. CA + modeling ability
Modeling ability
0.695
Control
0.006*
Modeling ability
CA + modeling ability
0.695
Control
0.046*
Control
CA + modeling ability
0.006*
Modeling ability
0.046*
* The mean difference is significant at the 0.05 level.
