Which course would you take to satisfy Upper-Division Writing Skills Requirement?

Dmitri Rogulkin
Senior Research Analyst
Office of Institutional Effectiveness
California State University, Fresno
The purpose of this visual display is to show variability in students’ outcomes across upper-division writing courses. Overall, course pass rates varied from 75% to 98%, while students enrolled in courses appeared to be equally prepared (based on their cumulative GPAs as of beginning of term). Thus, it looks easier for equally prepared students to pass some courses than other. In a few courses, female students were likely to do slightly better than male students (AFRS 104W, ENGL 160W, ANTH 116W). For some courses, comparison of pass rates becomes less meaningful because of small class sizes.
I’m using horizontal bars to display percentages of students passing courses. This makes it easier to read labels and compare data across courses.  The width of a bar is based on number of students enrolled in classes, which helps easily identify courses with just few students enrolled. I chose box-plots to show students GPA range and distribution, which allows for more information than averages or medians would give.
I decided to display only values for the pass rate variable to keep the image clean and make comparison between genders easier.  Sorting courses by overall pass rate makes it easier to see courses with the highest and lowest rates. Adding shades of blue color to the bars makes it easier to identify groups of courses with low/high pass rates, particularly in gender columns.



To add a comment, Sign In
Total Comments: 4
Gary posted on 6/6/2013 12:27 PM
I think this is a very successful graphic because it allows for quick and meaningful comparisons to be made, and intelligent next-step questions to be formulated. Color and bar thickness are both mapped to meaningful metrics, and the rug of distributed data points allows for finer scrutiny. Over-plotting in dense distributions could perhaps be improved by jittering the points or adding transparency, but it's a relatively minor issue. A pink(women)/blue(men) color scheme might create more visual interest, but that's a matter of taste. Good job and thanks for sharing!
Rebecca posted on 6/7/2013 5:53 PM
I like this display quite a bit. It takes more than a glance to understand all the pieces, but after figuring out the value, color, and size, it is extremely informative. The display provides a lot of data to answer the question posed. It both fits the intended message and uses an appropriate display for the type of data available. I like the monochromatic color scheme because currently all differences in color reflect meaningful differences in the data.

My only critique is that the GPA scale would be easier to use in the decision-making if it were wider. As narrow as it is, it's difficult to see differences.
Leslie posted on 6/10/2013 10:09 AM
Thank You for Sharing
Liz posted on 6/25/2013 12:13 AM
Clarity: I like this display – it has a lot going on but the information is all valuable. I was hung up on the title of the chart – this seems to suggest that the audience for the chart is students. If so, I’m assuming they can decipher the course and description well enough to know at first glance what course you are talking about. And frankly, since you have provided a footer, I'm assuming that if the user had more questions she or he could contact you.

Also, the bar and box headers are under the columns. I suspect this is a Tableau thing (may be some other software) but it feels counterintuitive to me to look below for the column headers.

Accuracy: The display allows me to be very accurate in assessing the percentage of students who pass each course. (Yes, you tell me the percent – it has to be pretty accurate!). The width of the lines give a sense for trend – it would be hard to be precise about the number of students. The trend data, however, is the point – not the look-up accuracy.

I think that the GPA display, while able to show a lot of information in a small space, may be too compact to be of much use. If I’m looking down the column, I can compare this measure of academic preparedness across courses with some precision. However, it is much harder to compare across columns, to compare men and women on a particular course, making this information broken out by gender only useful across a gender and not between. It’s a challenge –it may be that median GPA, while less nuanced than the quartiles, may be more informative in this space.

Fit with Intended Message/Appropriate Display for data and message: I think this is a great display – the use of color, width and line length provide interesting nuances to the data visually – there is a lot going on but it’s pretty easy to understand.

Aesthetic Appeal: I like the shades of blue, and the fact that you have tried to keep it simple and focused on the data – not a lot of colorful eye-candy. For me, frankly, the blue of the title and legend detracts a bit from the subtlety of the color palette (I really like the blue variations).

Thanks so much for sharing your chart – this complex chart feels remarkably clean and easy to understand.