Chapter 5. Main Study: Quantitative Data Analysis
5.9 Discussion
5.9.1 Performance due to Tasks and Boxes
The performance scores were dependent on the problem (Section 5.5.2, p.144).
Students performed well in Problem 2, which was an application problem. The scores
for the other application problem (Problem 1) and abstract problem (Problem 3) were
lower than that of Problem 2. Students may perform better in the application problems;
however, the experimental design did not allow an exhaustive look at the abstract and
application problems. Further, it was found that whilst mathematics confidence did not
influence how students performed in Problem 2, it did influence the performance in
Problems 1 and 3, where students with higher mathematics confidence were more likely
to do better. Therefore, students with higher mathematics confidence should mostly
perform better than lower mathematics confidence students as expected but there may
be problem types where mathematics confidence may not always influence
Performance was also found to be associated with tasks (Section 5.5.2, p.144).
The scores obtained for the interpretive and constructive tasks corroborated the
prediction and results of Galbraith and Haines (2000a) that students performed better in
interpretive tasks than constructive tasks. The study by Galbraith and Haines was based
on students solving these two task types with pen and paper. In the current study where
there is a technology-enabled environment, students performed better in the interpretive
tasks than in the constructive tasks. Although students had access to technology to solve
the procedural part of the constructive task, this did not provide any advantage to the
students, possibly as knowing what to do with the software box rather than being able to
use the software box played an important role in the performance of the constructive
task.
A comparison of scores obtained by Galbraith and Haines for their
undergraduate students solving tasks in polynomial algebra showed that their students
scored 37% less in the constructive task than in the interpretive task. In the present
study, the students performed even worse as they scored 60% less in the constructive
task than the interpretive task. This probably is because the two studies were in two
different mathematical domains or possibly because Galbraith and Haines used more
tasks (6 each).
The scores obtained for the interpretive and constructive tasks also varied with
Problem, where students performed better in the interpretive tasks for Problems 2 and 3.
The suggested reason for this was because the answers for these two interpretive tasks
had to be deduced from previously calculated values and these required interpreting or
reading off the variables carefully. For the constructive tasks, students all performed
significantly different on each problem, performing best on Problem 2’s constructive
problem type for the constructive task, with the students performing better on the
applications problems than the abstract problem.
Task types were also found to be affected by mathematics confidence with
students who had higher mathematics confidence performing better in the interpretive
tasks, whilst performing equally well in the constructive task regardless of mathematics
confidence. This probably meant that since interpretive tasks required conceptual
understanding that these students with higher mathematics confidence compared to
lower mathematics confidence students were more likely to make connections between
the answers ascertained from the mechanical task and then deduce what they meant,
such as in the cases of Problems 2 and 3.
It was interesting that the students regardless of mathematics confidence
performed the same in the constructive task, as this required finding relationships
between procedural and conceptual knowledge and then applying them together. It was
expected that the higher mathematics confidence students were probably more poised to
do these tasks since they had a higher likelihood of undertaking a deep processing level
and connecting their procedural and conceptual knowledge.
Now, looking at the first research question, that is:
Does the students' performance in solving the three task types depend upon the
software box they have access to?
What has been noted was that performance was dependent on task type but this
was already known from Galbraith and Haines. Now, was performance on the tasks
influenced by the software boxes? A cautionary yes is put forward. Performance on
either the interpretive or the constructive tasks was not significantly dependent on the
software boxes, although the difference in performance scores between the interpretive
dependent on the software boxes (Section 5.5.2, p.144). In particular, students using the
black-box software had the smallest disparity in their scores compared to the students
using the glass and open-box software. Further, marginal significance showed that
students using the glass-box software were doing better in the interpretive tasks than
those students in the black-box. Although there was no statistical significance, graphical
trends indicated that students using the black-box were performing best in the
constructive task. This suggested that for the constructive tasks, students using the
black-box software were more likely to grasp the conceptual-procedural knowledge
connection required for solving constructive tasks and perhaps the black-box software
influenced the ease in which the students were able to make this connection.