The results of the study will be presented in five sections. The first section is about overall performance of the sample in the diagnostic test. The next section deals with the students’ confidence in the electromagnetism questions. The third section considers the motivational orientations of the PBL students. The fourth section is about the learning strategies used by the PBL students. The final section is concerned with the relationships between the PBL students’ performance in electromagnetism and their motivational orientations and learning strategies.

Overall performance in the diagnostic test of electromagnetism

The scores for each electromagnetism item were correlated with the relevant total scores to identify the items that were providing unexpected contribution to the total scores. All items except one correlated positively with the total scores at least at the 5% level. So, in general, the items were expectedly contributing to the achievement scores. However, question 1, which was about the force on a stationary charge in a uniform magnetic field, did not correlate statistically significantly with the total scores. On this question, many students confused electric effects on a stationary charge with magnetic effects, and they were confident of their incorrect answers (see Table I). The mean achievement score was 5.5 out of 16 with a standard deviation of 3.0, suggesting that the students found the test quite difficult. The internal reliability of the test was investigated by calculating the KR-20 reliability coefficient. In the calculation, a student who gave at least 50 points to the correct answer got the question right. The test had a KR-20 value of 0.67, suggesting that the test items were reliable measures of the understanding of electromagnetism.

PBL students’ confidence

Table I presents students’ confidence in each electromagnetism question. In the table, a non-confident response occurred when students were not confident in their answer, and so divided 100 points between at least two options. The ‘non-confident right’ category is for non-confident responses which include the correct answer to the questions. Likewise, the ‘non-confident wrong’ category is for non-confident responses, which do not include the correct answer to the questions. A student was categorized as a confident right when s/he gave 100 points to a correct answer, and categorized as a confident wrong when s/he gave 100 points to an incorrect answer.

Table I. Students’ confidence in the basic ideas of electromagnetism.

	Confidence
	Confident right		Non confident right		I do not know		Non confident wrong		Confident wrong
Question	f	%	f	%	f	%	f	%	f	%
1	7	9.0	12	15.4	7	9.0	26	33.3	26	33.3
2	33	42.3	6	7.7	2	2.6	10	12.8	27	34.6
3	9	11.5	6	7.7	8	10.3	13	16.7	42	53.8
4	11	14.1	11	14.1	3	3.8	17	21.8	36	46.2
5	29	37.2	8	10.3	7	9.0	13	16.7	21	26.9
6	23	29.5	10	12.8	6	7.7	20	25.6	19	24.4
7	33	42.3	12	15.4	9	11.5	2	2.6	22	28.2
8	10	12.8	18	23.1	10	12.8	12	15.4	28	35.9
9	12	15.4	25	32.1	5	6.4	10	12.8	26	33.3
10	33	42.3	12	15.4	5	6.4	3	3.8	25	32.1
11	14	17.9	17	21.8	8	10.3	6	7.7	33	42.3
12	17	21.8	20	25.6	8	10.3	5	6.4	28	35.9
13	3	3.8	14	17.9	9	11.5	15	19.2	37	47.4
14	40	51.3	10	12.8	10	12.8	5	6.4	13	16.7
15	28	35.9	15	19.2	7	9.0	2	2.6	26	33.3
16	9	11.5	18	23.1	13	16.7	9	11.5	29	37.2
Total	311		214		117		168		438

Note: f is for frequency, and %, for percentage.

A noticeable finding from Table I is that about half of the non-confident responses (17.1% of all responses) were non-confident right, indicating that many students were in a transition state, and the correct and incorrect electromagnetism ideas existed in their minds together. The correct answers were given at least 50 points in about half of the non-confident right responses. Therefore, the students who provided a non-confident right response may benefit more from a conceptual change intervention than the students who did not use the correct idea when reasoning since they are more likely to question their knowledge when they are confronted with conflicting evidence. The fact that 60.0% of the responses were confident in their answer suggests that the students in the sample were inclined to be confident in their answers to the electromagnetism questions. However, 35.1% of the responses were confident wrong (i.e., overconfidence), suggesting that many students did not know that the answers they provided were incorrect. We can expect these students to hold onto their incorrect ideas even if they are faced with conflicting evidence. These students also are not likely to question their existing knowledge/beliefs about the basic ideas of electromagnetism.

The motivational orientations of the PBL students

Descriptive statistics of the motivation scales of the MSLQ is presented in Table II. The Cronbach’s alpha internal reliability coefficients of the scales are similar to the ones obtained by Buyukozturk, et al., (2004). The values of the coefficients suggest that the scales are reliable measures of the PBL students’ motivational orientations for physics. The first three scales in Table II are the value components of the motivation scales. The means of these scales are about one point above the mid-point of four, suggesting that the students had high interest in physics after being taught with the PBL approach for a year. However, we would expect the geology and geophysics students to be more interested in physics since they will need considerable physics knowledge to better understand the topics in geology and geophysics.

Table II. Descriptive statistics of the motivation scales of the MSLQ (n= 78).

Sub-scale	Number of items	Mean	Standard deviation	Cronbach’s Alpha
Intrinsic goal orientation	4	5.15	1.04	0.66
Extrinsic goal orientation	4	5.11	1.31	0.67
Task value	6	4.99	1.04	0.77
Control of learning beliefs	4	5.15	1.03	0.53
Self-efficacy for learning and performance	8	4.93	1.15	0.87
Test anxiety	5	3.85	1.37	0.74

The second component of the motivation scales, the expectancy component, consists of the sub-scales of “control of learning beliefs” and “self-efficacy for learning and performance.” The mean values of these scales indicate that the students believed that their efforts to learn physics by the PBL approach could produce positive outcomes, that they would perform well in physics, and that they were able to accomplish a physics task. On the other hand, in the affective component, many students felt anxious when taking physics tests. This may be related to the assessment and evaluation methods adopted by the Faculty of Engineering. The students were assessed through multiple-choice and open-ended questions and they had to score at least 60 out of 100 to provide a good contribution to their final grades. Giving more credit to the assessment and evaluation of the PBL processes rather than students’ physics knowledge might take some of the pressure from the students, and they would feel less anxious about their performance in physics.

The learning strategies of the PBL students

Descriptive statistics of the learning strategies scales of the MSLQ is presented in Table III. The Cronbach’s alpha internal reliability coefficients of the scales are similar to the ones obtained by Buyukozturk, et al., (2004). It seems that the scales are reliable measures of the PBL students’ use of different learning strategies for physics.

Table III. Descriptive statistics of the learning strategies scales of the MSLQ (n= 78).

Sub-scale	Number of items	Mean	Standard deviation	Cronbach’s Alpha
Rehearsal	4	4.78	1.19	0.63
Elaboration	6	5.03	0.99	0.78
Organisation	4	5.06	1.13	0.69
Critical thinking	5	4.69	1.21	0.83
Metacognitive self-regulation	12	4.74	0.89	0.78
Time and study environment	8	4.50	1.05	0.76
Effort regulation	4	4.18	1.24	0.62
Peer learning	3	3.82	1.33	0.65
Help seeking	4	4.46	1.25	0.51

The cognitive and metacognitive strategies component of the learning strategies includes the sub-scales of rehearsal, elaboration, organisation, critical thinking, and metacognitive self-regulation. The mean values of these scales suggest that many students will experience difficulties in constructing internal connection between physics ideas to be learned, applying their prior knowledge to new situations in order to solve problems, reach decisions, and having control of their cognition. One would expect the PBL students to use cognitive and metacognitive strategies more frequently since the PBL approach requires students to read many physics materials to be successful in the course. It is quite possible that the students’ inadequate use of cognitive and metacognitive strategies contributed to their poor understanding of the basic concepts of electromagnetism as indicated by the mean electromagnetism score of 5.5 (out of 16) with a standard deviation of 3.0. As a student-centred approach, the PBL approach requests students to use their resources, such as students in their PBL group, effectively. However, as suggested by the sub-scales of time and study environment, effort regulation, peer learning, and help seeking, the PBL students in the sample were not very good at using the resources available to them. This also may be a factor contributing to poor conceptual understanding of the PBL students in electromagnetism.

The relationships between the PBL students’ performance in electromagnetism and their motivational orientations and learning strategies

The PBL students’ performance in electromagnetism was correlated with their scores in motivational orientations and learning strategies scales to investigate if there is a relationship between these variables. There was a weak association (0.252) between the students’ performance in electromagnetism and the scores of the self-efficacy for learning and performance scale at the 5% level. Surprisingly, the other variables such as intrinsic goal orientation and critical thinking did not have a significant statistical correlate with the students’ achievement scores. The fact that the students found the diagnostic test difficult may be one of the reasons why many motivation and learning strategies scales did not correlate with the students’ achievement scores.