Student Retention at Community Colleges

An Exercise in Data Visualization

America's community colleges play an essential role in preparing people for the jobs of the twenty-first century. Students, young and old, come to community colleges for courses in a remarkable array of subjects. They can gain basic skills in math and reading, take college level courses in calculus and chemistry, learn job-related requirements in auto repair, welding, nursing, and many other subjects. Faculty and adminitrators are continually working to improve the prospects for retention and progression of the many students who are lacking in basic skills, pressed by the responsibilities of work and supporting families, and otherwise at jeopardy for dropping out.

The Community College Research Center (CCRC) at Columbia University Teachers College has been one of the centers for quantitative research in community college retention studies. They have issued a series of CCRC Briefs, one of which is the subject for this page's exercise in data visualization. Brief 45, Student Progression Through Developmental Sequences in Community Colleges, dated September 2010 and referred to herein as CCRC 45, is a particularly useful meta-study. CCRC 45 presents several tables, well laid out but challenging to understand. I saw this as an opportunity to use on of my favorite programming languages, Python, to quickly develop helpful interpretive visualization.

Here is the data from Table 1 of CCRC 45:

Table 1: Enrollment in and Completion of Developmental Sequences

Developmental Courses Never Enrolled Enrolled - Lost Enrolled - Failed or Withdrew Completed Sequence Total (N) Total (%)
1 level Below 37% 2% 17% 45% 56,551 42%
2 levels below 24% 13% 32% 32% 38,153 27%
3+ levels below 17% 23% 44% 17% 43,886 31%
Total 27% 11% 29% 33% 141,590 100%

Developmental Courses Never Enrolled Enrolled - Lost Enrolled - Failed or Withdrew Completed Sequence Total (N) Total (%)
1 level Below 33% 5% 12% 50% 5,341 69%
2 levels below 21% 13% 24% 42% 16,983 32%
3+ levels below 27% 19% 25% 29% 8,825 9%
Total 30% 8% 16% 46% 78,149 100%

Personally, I find it difficult to compare and contrast using these two tables. I hope you will find the two graphics below to be of assistance in understanding Table 1. They are built for information, not for beauty.

Math:  Math visualization  Reading: 

The Python program for generating these two graphics can be found here. It is written in the spirit of Python, that it is able to do simply whatever it can do.

You can reach me at my hotmail dot com account: cassiodorus

last update 24Mar2012