Scamming Their Way into College — But Why?

The college cheating scandal in the news recently has highlighted the competition for slots at “elite” universities.

Thirty-three wealthy parents were charged with paying a college preparatory company to doctor their children’s scores on standardized tests, pay bribes to college coaches and university administrators, or falsify the students’ ethnicities, disability statuses, or athletic profiles — all with the goal of getting their children coveted spots in the freshman class of Georgetown, Yale, or one of six other universities.

Of course, these are allegations and the parents are presumed innocent until proven guilty, but the incident raises the question: How much college-admissions fraud is there?

No one knows. The only time one learns about such fraud is when a law enforcement agency or journalists expose it. This particular investigation involved more than 200 FBI agents nationwide; had its budget been cut, it is likely no one would have ever been charged. According to CNN, the FBI discovered this scheme while investigating a different case about securities fraud: a financial executive told FBI investigators about a coach taking bribes in return for admitting students.

The irony in this story is that there are many, many colleges where students can get an excellent education without cheating on their admissions applications. Frank Bruni argues in Where You Go Is Not Who You'll Be: An Antidote to the College Admissions Mania that numerous college paths end in career success — and the paths don’t all go through Georgetown or Yale.

Bruni posits that public perception of college status is influenced by articles that mention the undergraduate degree of a successful person only when it is from an “elite” institution. Peter Osterlund looked up which college had awarded the degree of each person on Forbes magazine’s “30 under 30” lists and wrote: “We found that Forbes tells of one ‘30 under 30’ honoree's experience as an undergraduate at Duke, but doesn't mention Arizona State University undergraduate degrees carried by three of its young stars … We found that most Forbes ‘30 under 30’ honorees attended, well, ordinary colleges—in some cases, obscure places, in other cases, state schools like the University of Where-They-Just-So-Happened-To-Live-At-The-Time.”

This started me wondering about my own field of statistics. One of the highest honors in statistics in the United States is the designation as Fellow of the American Statistical Association (ASA), awarded to persons who “have made outstanding contributions to statistical science.” Where did ASA Fellows obtain their undergraduate degrees?

I selected a stratified random sample of about 20% of the persons awarded ASA Fellow between 2010 and 2018.* I was able to find the undergraduate degree-granting institution from web-posted CVs or other publicly available information for 92 of the 106 Fellows in the sample (see spreadsheet).** The spreadsheet also lists the institutions that awarded the PhD degree (when applicable) for all 106 Fellows in the sample. The analysis below concerns undergraduate degrees only; the graduate-degree-granting institution is often an important factor in hiring decisions, and more than a quarter of the Fellows in the sample received their PhD from the University of Michigan, University of California-Berkeley, Harvard, Iowa State, or Carnegie Mellon.

About half of the 47 Fellows who received an undergraduate degree in the US graduated from public universities. Harvard, Columbia, UCLA, the University of Florida, and the University of Wisconsin-Madison each had 2 undergraduate alumni in the sample of Fellows — and so did Tabor College, a small liberal arts college in Hillsboro, Kansas.

Some college commentators use a college’s acceptance or admissions rate as a measure of its selectivity and until 2019 acceptance rates were a component in the U.S. News & World Report college rankings. The acceptance rate is, of course, a flawed measure of quality — and the more it is used for ranking, the more flawed it can become. There is intense competition among colleges to achieve better rankings, and a college can manipulate its acceptance rate by making the admissions process easier or cheaper, or encouraging applications from students who are unlikely to be accepted. Such practices increase the number of applications, thereby lowering the acceptance rate and making a college appear more selective.

Since the measures are so widely used, however, let’s look at admission rates for the schools attended by the ASA Fellows in the sample. The US Department of Education’s College Scorecard Data contains admissions rates (defined as the number of admitted undergraduates divided by the number who applied for Fall semester), SAT/ACT scores, graduation rates, student debt, post-college employment and salaries, and a wealth of other information for US colleges and universities. The histogram displays the admissions rates from the 2010 Scorecard data for the institutions attended by the 47 Fellows who studied in the US. The histogram shows about the same number of Fellows attending schools with admissions rates below 30% as schools with admissions rates above 70%.

How do these compare with the 2010 admissions rates for colleges attended by persons who are not Fellows? I do not have college information on other ASA members. But the 2010 College Scorecard gives information on all colleges, and I used that to create a rough comparison set. Since most statisticians receive their undergraduate degree in the mathematical sciences, I estimated the number of undergraduate students receiving degrees in math or statistics from each institution and assigned the college’s admission rate to each of those students.***

My estimated distribution of admissions rates for the colleges of persons with “math undergraduate degrees” in 2010 shows relatively few students in colleges with low admissions rates, with most attending colleges with admissions rates between 30% and 80%; the average admissions rate is 58%. The average admissions rate in 2010 for colleges attended by the ASA Fellows is 50% (with 95% confidence interval from 42% to 58%).

This sample of Fellows is small, and one must be careful not to overgeneralize. And of course we cannot conclude that attending a specific college “caused” a person to become a statistician or ASA Fellow — we don’t know what would have happened had he or she attended a different institution.****

But it appears that, while ASA Fellows may have been modestly more likely to have received their undergraduate degrees from a “selective” university, they hail from a wide range of undergraduate institutions. A recent article reviewing research on the relationship between college attended and later life outcomes concluded that “college selectivity is not a reliable predictor of student learning, job satisfaction, or well-being” and “engagement in college is more important than where you attend.” In other words, what you get out of college depends on what you put into it — good news for all the non-cheaters.

Copyright (c) 2019 Sharon L. Lohr

References and Footnotes:

Bruni, Frank (2015). Where You Go Is Not Who You'll Be: An Antidote to the College Admissions Mania. New York: Grand Central Publishing.

*I stratified the sample by year and gender. This means that I took a separate random sample of about 20% of the men and 20% of the women awarded Fellow for each year between 2010 and 2018. Taking a separate sample for each year and gender ensured that the full sample would have the same percentage of persons from each year, and the same percentage of men and women, as the complete list of Fellows.

**It is possible that the Fellows who do not list undergraduate institutions on their CVs attended different types of institutions (perhaps with higher admissions rates) than the Fellows who posted that information.

***To be precise, I restricted the data to the 1,174 institutions that awarded undergraduate degrees in math or statistics. To get the data on a per-undergraduate-math-major basis (to correspond to the histogram from the Fellows) I calculated the product z = (number of undergraduates) x (percent of degrees in math and statistics) x (6-year college graduation rate) for each institution. I then looked at statistics for admissions rates assuming the number of math undergraduate degrees from each institution is proportional to z (in statistics terminology, I weighted the data by z). Of course this is just a rough estimate, but is probably a better approximation than would be obtained by considering all types of undergraduate majors. I also estimated the distribution using weights (number of undergraduates) x (percent of degrees in math and statistics) and the shape was similar but with a slightly higher average admission rate of 62%.

****There has been a lot of research on estimating the monetary return from attending a highly selective college. See, for example, the paper by economists Stacy Dale and Alan Krueger, who looked at earnings, 14 years later, of persons who had entered college in 1989. Arguing that students with greater earnings potential are likely to apply to more selective colleges, they compared the earnings of students who had attended a selective college with those of students who had applied to selective colleges but had attended college somewhere else. They found that the earnings premium for most students was associated with having applied to selective colleges rather than having actually attended them.