Data Results and Analysis

Methods

In order to assess the effects of our teaching, we administered a survey to students designed to measure their attitudes and beliefs about science. In addition to surveying the students exposed to PBL (experimental group), we also administered the survey to students in classrooms that were not exposed to PBL (comparison group).

Instrument

The Student Attitude Survey was obtained from Dr. David E. Kanter, an expert in designing and testing PBL interventions in science. The instrument measures five aspects of students' beliefs and attitudes about science. Specifically:

• Science self-concept (Self-concept)
  
• Sense of efficacy doing general science tasks (Efficacy 1)
  
• Sense of efficacy doing science tasks related to the curriculum (Efficacy 2)
  
• Interest in science (Interest)
  
• Perception of the value and relevance of science (relevance)
  

Each factor includes several statements with which students were asked to indicate their level of agreement. (For example, the statement, 'I enjoyed what we did in science class this year' contributes to the factor of 'Interest in science.')

Data Collection

The survey was administered to the 7th grade experimental group three times (January of 2013, March of 2013, and  May of 2013). The survey was administered to the 8th grade experimental groups twice (February of 2013 and May of 2013). The comparison groups for both grades took the survey in the spring of 2013. 

Results

In Grade 7:

• Within the experimental group, females score significantly lower on indicators of attitudes toward science (on a scale of 1-5) than their male classmates (p < 0.01).
  

• Comparison group students had a significantly higher attitude towards science than students in the experimental group. This difference is significant at the 0.10 level (p = 0.067); however, its practical significance is questionable since the difference in average attitudes is 0.18 on a scale of 1-5.

• When comparing responses on the factors (i.e., efficacy, interest, relevance, self-concept), there is a significant difference by sex (p < 0.01) and by factor (p < 0.01). This is most evident in the combined sex by factor result (f = 2.47, df = 4, p <.05), as seen in the figure below. Female students consistently score lower on all the factors, but especially lower on the efficacy and self-concept factors.

                              

In Grade 8:

• When considering individual factor scores, sex, and time, females score significantly lower than males (p < 0.01). There are also significant differences on how the experimental group responds to the factors (p < 0.001). By looking at the average scores for each factor, it seems that the group scores lower on Efficacy 2 (Sense of efficacy doing science tasks related to the curriculum) than all other factors. In other words, there is a significant difference between how the experimental group responds to Efficacy 2 than to all other factors.
  

• At time 3, the experimental group scored significantly higher than the comparison group on the combined factors (p < 0.01). This finding lends support to the hypothesis that students who experience PBL score higher on their attitudes towards science than students who did not experience PBL. As is the case with seventh graders, females report significantly lower attitudes towards science than males (p < 0.05). 

Analysis

Our data collection and design was in its preliminary phase, with some flaws within the experimental design. Therefore, the results from our action-research do not provide sufficient evidence to "prove" that project-based learning has a direct impact on students' attitudes toward science. However, the reliability scores (α 0.93-0.94) for the survey instrument were very strong, indicating that the survey can be reused with future incoming students, given that our sample population represents the overall student population in our school communities.

Implications for Further Research

During the next school year, our team plans on conducting the same action-research with an improved experimental design in hopes of collecting more reliable data that will allow us to answer our research question. The discrepancy between male and female attitudes towards science, with females scoring lower, is a topic worth researching further as well.

--> Return to Portfolio Home