Evaluation - Advances in Personalized Web-Based Education - page 112

Table 4.16 Multiple comparisons about reading times for group B
(I) Background

(J) Background Mean difference

(I-J)

Std.
error

Sig.

0.4925 * Computer science related 0.07250 0.000

Arts

0.2985 * Science (other than

computers)

0.07755 0.002

− 0.4925 *

Arts 0.07250 0.000

Computer science related

− 0.1940 *

Science (other than
computers)

0.07618 0.041

− 0.2985 * Arts 0.07755 0.002

Science (other than
computers)

0.1940 * Computer science related 0.07618 0.041

* The mean difference is significant at the .05 level

background on computer-related sciences (Sig. < 0.0005). On the other hand, the
results of the Table 4.16 show that for group B the mean value of reading times
is significantly different only between the learners with background on arts and
learners with background on computer-related sciences (Sig. < 0.0005). For group
B, the mean value of reading times does not differ significantly between neither
the learner with background on arts and learners with background on sciences
other than computers (Sig. = 0.002 > 0.0005) nor the learners with background on
sciences other than computers and learners with background on computer-related
sciences (Sig. = 0.041 > 0.0005). Therefore, the student model of the presented
programming tutoring system makes valid conclusions concerning the aspects of
the students' characteristics.

4.3.8 The Validity of the Adaptation Decision Making
of the Student Model

The results of the validity of the system's decision making were very encouraging.
It has been answered that the percentage of time that the learners spent to revise
a previous domain concept is at all waste of time. The results of the learners'
answers to the questionnaire F (Appendix B) are depicted in Fig. 4.14 .
In addition, learners' performance in a final test was measured and compared
for two groups (group A: the group of students which used the presented
programming tutoring system, and group B: the group of students which used a
similar programming tutoring system with the same organized content, in which
the learner chooses if s/he will return to revise the concept that the system
indicates her/him and/or if s/he will read or not the suggested concept each time).
The two-independent sample t-test was used to analyze the experiment's data. The
results of the experiment are depicted in Table 4.17 . The variability in two groups
is about the same (value “Sig.” = 0.106 > 0.05 Table 4.18 ). Therefore, the different
average scores did not occur by chance or due to differences on the education,
knowledge level and abilities of the learners of the two groups. Furthermore, the
differences between the mean values are statistically significant; since the value of