Consulting for Primary Analyses: What’s in a p-Value?

Once the survey was administered, we had to make certain the data were accurately coded so results would stay true to the survey data. For instance, we had to create a summation of Likert scale responses that was accurate in relation to the amount of agreement with a certain feeling to make our Likert indexes. We also had to omit responses from those who may have not been truthful on the survey, such as one respondent who reported “mango” for their gender.

After the coding was complete, the data was ready to be analyzed and we were called upon for a consultation. It might be fair to say that, at this point, many students believed their primary goal of data analysis was to obtain a p-value of < 0.05 and our influence might make this happen. Most students also groaned a bit when we mentioned the need to first look at the data through graphical representations, such as boxplots and scatterplots, and to check for assumptions (not everyone’s favorite part of data analysis). Fortunately, the sociology students confessed how “fun” it was to see the final results once we got into our various parametric and nonparametric tests.

We could tell how disappointed some groups were to find null results, so we made sure to stress the importance of null findings. For example, it was interesting that there was no difference between males and females in the number of hook-ups they have engaged in at St. Olaf. Another surprising null finding was that the perceived level of intimacy in couples was not associated with the total number of romantic relationships a person had been in, unless we controlled for the number of romantic relationships between St. Olaf students.

This portion of the consultation was probably one of the most rewarding for us. Many students, undergraduates in particular, are initially intimidated by anything related to math or statistics. However, the sociology students came to enjoy the data analysis because they finally got to see the results they had been waiting for.

The Next Stages: Deeper Analyses and Formal Presentations

The next unit of the sociology research methods course was to create a research poster and the first manuscript of a research paper. We explained to our peers the goal of a research poster in statistics: to create an eye-catching representation of the knowledge gained from the study. In a final meeting, we proofed their research manuscripts, happy to find that everyone’s hard work had paid off. By the end of the semester, many of the groups had wrapped up work on their projects and were looking forward to presenting at a regional conference in the spring.

Because CIR projects run the entire school year, we continued with the projects by conducting analyses of our own after the sociology research methods course was complete. This gave us a chance to investigate further relationships and more sophisticated modeling techniques. Each of us contributed an appendix to the sociology students’ scientific papers with our own results and discussion sections.

We concluded our work by attending the Midwest Sociological Society Conference and presenting with the sociology students to an audience of students and faculty. Initially, we had concerns about stepping on toes, as the projects were not ours to begin with. However, the responses we received from the groups we worked with were overwhelmingly appreciative. The groups even expressed relief in having us at the conference as “stat experts.” The positive reactions were not limited to our groups, either. Many conference attendees expressed interest in having undergraduate statistical consultants for their schools or businesses in the future.

Final Thoughts on a Significant Statistics Undergraduate Experience

In addition to helping our fellow students through the process of conducting a research project, we were awarded valuable experience as statistical consultants. We made a few mistakes along the way, but our contributions provided many learning opportunities for each of us. Working with peers in a different discipline forced us to familiarize ourselves with the basics in their field, as well as new techniques in ours. Through this experience, we found it helpful to think about how to explain basic statistical terms and techniques to an audience without a strong background in statistics, strengthening our own understanding of statistics through the process. Most importantly, undergraduate statistical consulting gave us the opportunity to apply the tools we have learned in the classroom to real-world research that we were able to follow from start to finish.