This is a 2006 study by Clifford Adelman where he updates his model from "Answers in the Tool Box. Academic Intensity, Attendance Patterns, and Bachelor's Degree Attainment" using the data from the NELS:88/2000 study, which again follows nationwide students, this time from 8th grade in 1988 through to 2000. While this gives slightly less time to graduate from college than the first study, it is still a useful comparison. Adelman is attempting to test whether his results hold up, how graduation rates have changed, and to address any criticisms of his first piece. (If you have not read the previous post on this paper, you should do so now.)
Academic resources is still a predictor of success, but less so. High school rank/gpa became more important than test scores, and completing Algebra 2 no longer doubled the chance of degree obtainment, now it is required to go beyond Algebra 2. The reason this factor is weaker is because the model takes more into account of college academics. The NELS:88/2000 included more information on courses taken, which led to some interesting results, such as the fact that 5 or more credits earned over summer drastically improves graduation rates, especially for African-Americans.
Once again, transferring is a mixed bag. For those transferring from a 2 or 4-year institution to a 4-year institution and earn at least 10 credits from each, transferring increases graduation rate. But multiple schools lowers graduation rates. Adelman describes the later case as brought down by students "swirling" back and forth between 2 and 4-year systems, wandering aimlessly. These are the transfers that the system desperately needs to avoid.
Compared to the original model, this one added multiple schools, whether a student was ever part time, and percent of classes dropped after deadline as negatives and summer term credits and cumulative college math credits as positives. Parenthood was no longer a factor, nor was earning less than 20 credits in the first full year. Overall, the model was about as predictive as before, this time continuous enrollment, first year grades, and percent of classes dropped after the deadline were the biggest predictors of graduation. The model still finds race and gender insignificant.
Over these two papers, Adelman's main points are that college success comes from a strong high school foundation and continuous enrollment in school. If we want to study graduation rates, we need to examine students, not institutions, and we need to use hard evidence, not survey data. The system needs to focus on students, setting high expectations, encouraging summer credits, making sure they have a plan and the resources needed to carry it out.
We should discuss causality here as you make sense of this sort of research. One issue is whether measured variables are causal or if, instead, they are only proxies to other variables which really determine what is going on. If the latter, then the question are what are the underlying variables to consider and what is the correlation between the underlying variables and the measured ones.
ReplyDeleteI will give a simple example based on my own experience. I transferred from MIT to Cornell in the middle of my sophomore year. Before I transferred I was going downhill mentally and emotionally. Had I stayed at MIT I likely would have dropped out of school.
So consider a simple two-state situation. The first state has the student happy and the student finds school nurturing. The second state has the student unhappy and school appears a contributing factor to that. Those two states cover all the possibilities for enrolled students.
Now you can envision about how those states evolve over time. The simplest possible model is Markovian. There is some probability that the student remains in the state they are in. There is a complementary probability that they transition to the other state. This assumption is not made for realism. It is made to keep things simple. You might also assume that entering freshmen are all in the first state.
Then ask yourself what happens over time and what choices do students make. Again, simplicity dictates how to model this. Students who like school stay enrolled and proceed to complete their degrees. Students who are unhappy about school have a tripartite decision to make: (a) stick it out and hope things get better, (b) transfer and hope things are better at the next place, or (c) drop out.
In this simple model then, students who transfer are unhappy, but they are hopeful that the new place will be better for them than the old place. That was my situation exactly.
In reality there may be other reasons for transferring. If you transfer from a private to an in-state public, tuition is lower. You might transfer to be closer to home. You might transfer because of an academic interest where the new school is better in that area than the old school.
I don't know if you can do this sort of thing for the other variables in Adelman's study or not. But, for example, you write that high school ranking becomes more important than test scores in predicting success. One might ask, cynically, so what? Or one might ask, more earnestly, what does that signify?
Until you do this sort of exercise, I'm not sure what you are learning from reading this stuff.
You might frame the issues this way. Suppose you are talking to a parent of a student in high school or you are talking to a high school guidance counselor. In each case assume they haven't read the research. So you are going to advise them based on your reading of it. What would you tell them?