Read here
This 1999 study is one of the most cited studies on the subject of college graduation, and for good reason. Clifford Adelman uses data from High School & Beyond/Sophomore cohort files, a study that followed a national selection of students from 10th grade in 1980 through to 1993. Researchers tested these students in 10th and 12th grade, asked them about family background and future ambitions in both these grades, asked about degree obtainment and college experiences throughout the period, and, most importantly, tracked the high school and college transcripts of these participants over the length of the study. Adelman is interested in examining what are the main factors in college graduation, especially for bachelor's degrees, and what society can do to improve these rates.
This study is important for a number of reasons. First, it reconciles what participants report against what the records say. Only asking students leads to faulty results, either because students do not know answers, as is often the case with parental education or salary, because they misunderstand the question, such as those who study abroad for a semester claiming they've only ever attended one school, or they lie by over-reporting the good, such as grade point average (gpa) or under-reporting the bad, such as remedial classes taken. Second, it track students as they move throughout the higher education system, not just how they fair at any particular institution, and it tracks them for an extended period of time. Many surveys, especially before this one, focused only on first or second year performance at the same institution where the participant started, but the goal is not persistence, but graduation. Universities may want to focus on retention rates, but for students, what matters is whether the receive their degree or not.
The first factor Adelman considers is academic resources, a term he borrows from previous research, which attempts to measure academic performance in high school and how it affects college graduation rates. This factor, which is an index obtained by combining intensity of high school classes, test scores, and class rank, becomes one of the major determinants of graduation. Students in the lowest two socio-economic status (SES) quintiles but highest academic resources quintile averaged higher graduation rates than those in the highest SES quintile overall. A great high school education is a strong predictor of college success. Especially important is math; completing Algebra 2 in high school more than doubles the odds of receiving a bachelors degree. SES still has some significance in the model, but academic resources is a much more powerful indicator.
At the time of the study, the phenomena of attending multiple institutions was growing, as it does to this day, and researchers were still behind on how they measured its effects. Transfers can have very different effects on graduation rates depending on the schools being transferred from and to, how many credits earned at each, and whether the original school is transferred back to. For example, students who attend a community college where they earn 10 or more credits, then transfer to a 4-year institution where they earn more than 10 credits have higher graduation rates than those who start of attending a 4-year institution, even though those who attend a 4-year institution first have higher graduation rates overall. Transferring to a school in a different state also increases graduation rates over transfers within a state. But no matter where the student is, the biggest factor is that they are continuously in school, never missing more than 2 semesters or 3 quarters, and this is the second big factor in graduation rates.
Other factors that they study found were significant were whether they had children before 1986 (age 22 to 23), completed less than 20 credits in first complete year, or transferred and did not return to the same institution all of which had negative effects, while improved gpa from freshman year to senior year and freshman gpa, as well as previously mentioned continuous enrollment, academic resources, SES quintile, and transferred from a 2-year college to 4-year college and earned 10 credits at each all had positive effects. Overall, the model had about a 43% explanatory power, mainly coming from continuous enrollment and academic resources. What is interesting is that race and gender have no significant affect on the model, despite numerous attempts to include it, and while SES quintile matters, it is much weaker than expected.
My suggestion is that while you read these background pieces to make heads or tails of what the literature says you simultaneously try to work through a simple choice-theoretic model on whether to continue with schooling or drop out. This might help you make heads or tails of what your reading. Here are a few preliminary thoughts on such a model.
ReplyDeleteFirst, there are some very famous dropouts - Bill Gates, Steve Jobs, and Mark Zuckerberg in the computer area - I'm sure you can come up with names of athletes who turned pro and left college early. So one reason to drop out is that the value of the outside option exceeds the value of staying in college. At a societal level, this reason is likely of little importance, but it is helpful to get started with the modeling.
Next, Higher Ed is segmented. You can segment along many different dimensions. One of those is age of student as compared to age at the time of high school graduation. Traditional students go to college right out of high school (or if they take a year off they are still close enough in age to other traditional students). Non-traditional students are older, may have a family, and may be working a full time job.
The composition of the student population has changed over time. It used to be majority traditional students. Now it is majority non-traditional students.
There are many other dimensions of students. Parental income is one. First language non English is another. Standardized test scores is one more.
There there are different types of college. Residential - full time. Non-residential full time. Non-residential part time. And more recently, there are for profits versus not-for profits.
College attendance is known to increase during periods of high unemployment, but is that rational? There are now for profit providers. Some of these are not above board. They have no intention of graduating students. They make money because the students get loans for tuition, then drop out. They are referred to as diploma mills, though they typically don't give a degree.
If a student is at a diploma mill, the reason for dropout is likely provider moral hazard. If a student is at a legitimate institution, and if the degree would seem to confer value in the labor market, then it is something else that encourages dropout. Getting a list of those explanations at a micro level, would then help in reading papers like this one.
It is okay for you to refine this model over time as you come to understand things better. But I'd get into doing it early on.