I did a teaching methods study last year in two of my classes. I developed a survey with Likert*-response format questions** testing some self-reported values, which I then correlated to course performance. Since my survey did not create a true Likert scale (e.g. a survey with multiple questions testing the same variable to create a scale), it complicates the data handling. Before the study I had scribbled many notes and ideas in my lab notebook, so now I am in the process of carrying out my intended analyses.
One analysis that I am currently working on when I can find time involves rank testing. I have been struggling with how to present the data in a fashion that accurately tells an honest story. Tonight I made a prototype chart that helped me take a step forward. It’s still very rough and not completely developed, but the core idea has been planted and I am excited to create the polished version of the chart soon.I decided to present counts of the responses from students in four categories. My previous plan had been to present percentages on the Y-axis, but I felt that it was deceptive. Using counts allows the reader to remain aware of the size of the samples in each category.
The final chart will look like the prototype above, yet will have four different variables (i.e. be four times the size of the chart above). The labels and titles will be clearer as well (this is currently unpublished data, so I want to be vague with it right now).
My goal is to submit this manuscript by the end of the winter break. I don’t know if it is a reasonable goal yet or not–but it is a goal nonetheless. . .
* This study taught me I did not know how to pronounce Likert.
** This study taught me that I had a misconception of Likert-scales.