4.1 Pre- and Post-tests

The pre- and post-tests were identical in task and each comprised two parts.  First, participants were required to match pairs of objects rendered in three-dimensions according to their angular rotation.  They were given a minute each to complete five such matching exercises.
    
Table 8 presents the time taken to attempt to match the five pairs, both by the performance in the pre- and post-tests, as well as by gender.

It can be seen from the above table that there was a significant improvement in the durations needed to complete this part of the test.  However, there was no significant difference by gender, in their respective performances. 
   
Table 9 presents the number of correctly matched pairs, both by the performance in the pre- and post-tests, as well as by gender.

It can be seen from the above table that there was no significant improvement in the scores obtained, between the pre- and post-tests.  However, Tables 13 and 14 both show that male students scored consistently higher than female students (pre-test: F(1, 104) = 5.766, p = 0.018; post-test: F(1, 92) = 5.607, p = 0.020).  For all students, the pre-test showed a negative correlation between the number of accurately matched rendered objects, and the time taken to match them (males: correlation = -0.301, p = 0.027, N = 54; females: correlation = -0.398, p = 0.030, N = 52).
   
The second part of the pre- and post-tests required participants to orientate a QuickTime VR panoramic scene to match, within five minutes, a map of the same area.
   
Table 10 presents the time taken to match the orientation of the panorama to the map, both by the performance in the pre- and post-tests, as well as by gender.

It can be seen from the above table that there was no significant reduction in the time taken for this task.  However, among the female students, the pre-test showed a correlation between the time taken to match the rendered objects, and the time taken to orientate the panorama (correlation = 0.482, p = 0.000, N = 52).
   
Table 11 presents the angular deviation from the correct orientation of the panorama, both by the performance in the pre- and post-tests, as well as by gender.

It can be seen from the above table that there was no significant improvement in the angular accuracy in the orientation of the panorama.
   
Table 12 presents the distance deviation between the perceived axis of rotation of the panorama against the true axis of rotation, both by the performance in the pre- and post-tests, as well as by gender.

It can be seen from the above table that there was no significant improvement in the accuracy in the deduction of the true axis of rotation.  However, there was a correlation between the angular accuracy of the orientation, and the distance deviation from the axis of rotation.  This correlation held true regardless of gender, and for both the pre- and post-tests (pre-test males: correlation = 0.768, p = 0.000, N = 54; post-test males: correlation = 0.586, p = 0.000, N = 49; pre-test females: correlation = 0.473, p = 0.000, N = 52; post-test females: correlation = 0.378, p = 0.011, N = 45).

Similarly, for all students in the post-test, there was a correlation between the time taken in the panorama-to-map exercise, and the angular accuracy of the orientation (males: correlation = 0.293, p = 0.041, N = 49; females: correlation = 0.339, p = 0.023, N = 45).  For the post-test again, there was a correlation between the time taken by female students to orientate the panorama, and the distance deviation from the axis of rotation (correlation = 0.323, p = 0.030, N = 45).

Table 13 presents a one-way analysis of variance analysing the various aspects of pre-test performance against the gender of the participants.

Table 14 presents a one-way analysis of variance analysing the various aspects of post-test performance against the gender of the participants.

It can be seen from Tables 13 and 14 above that male participants scored consistently higher in the matching of rendered objects than their female counterparts. For example, for the pre-test, the mean number of correctly matched pairs of rendered objects was 3.8 for the male students, and 3.3 for the female students.  The respective standard deviations were 1.1 and 1.2 (Table 9).  The difference was significant at the p = 0.018 level, F(1, 104) = 5.766 (Table 13).  In the post-test, the mean scores were 4.0 and 3.6 respectively (standard deviation were 1.0 and 1.1 respectively).  The difference was significant at the p = 0.020 level, F(1, 92) = 5.607 (Table 14).  The better performance of males in this test is consistent with research by Grön et al (2000) which suggested that females are less able to mentally rotate spatial relations.

<- 4 Results and Discussion of Main Study          -> 4.2 Orienteering Task