The statistical software made available to thesis students in the Reed Psychology Department generally includes a user-friendly program that is not capable of certain types of analyses (e.g., logistic or hierarchical regression), and a less-friendly advanced program that can do these analyses, but requires knowledge of a programming language. These software options have worked well for many thesis projects, but not for research in educational and developmental psychology. Oftentimes in these areas, students will examine a combination of continuous and categorical variables, requiring the use of regression analyses. For example, do students cheat or not, as a function of individual variables (e.g., self-esteem) and contextual factors (e.g., performance- versus mastery-oriented classrooms)? Can we predict the likelihood of women becoming math, humanities, or social science majors as a function of perceived stereotypes in the educational environment? Not having accessible and appropriate software to conduct these analyses limits the sorts of questions that my students and I can pursue.
Therefore, I used funds from the Murdock Grant to purchase an advanced regression module for the SPSS program in my lab in May 2002. This software has been useful in a number of ways. For my own research, I was able to analyze data examining the role of student-teacher relationships as a moderator of the effects of praise on children’s intrinsic motivation. The analyses revealed that, in the context of a supportive and involved relationship, praise enhances children’s intrinsic motivation; in the context of a less-supportive relationship, praise tends to have little to no beneficial effects. Without the ability to conduct hierarchical regression analyses, I would have been forced to dichotomize the continuous variable of relationship quality.
The software is also beginning to be extremely useful for thesis research in my lab. One thesis student is currently examining the relative importance of parent, peer, and teacher attachment relationships as predictors of several academic and mental health outcomes in an early-adult population. User-friendly statistical software that allows for regression analyses will be essential for the data analysis phase of this project, which will begin in the early spring of 2003. I anticipate many more thesis projects over the coming years that will demand the use of this software. Given the limited time-frame in which thesis students must conduct their work, the powerful but user-friendly nature of this software is especially critical.
Finally, during the summer of 2002, the teaching assistant for our research design and data analysis (RDDA) course also became familiar with our newly-purchased SPSS program. Our department has now decided to replace our existing user-friendly program with SPSS as the main data analysis program taught in the RDDA course, beginning in 2004. I expect that this will greatly assist our students in the data analysis process for their thesis and other research projects. Thus, the support of the Murdock grant has been helpful for student and faculty research, and for curriculum development in our RDDA course.