• August 27, 2015

At Elite Colleges, Legacy Status May Count More Than Was Previously Thought

Family connections help you get into college. And a new paper suggests that at highly selective colleges, they may count even more than was previously thought.

A researcher at Harvard University recently examined the impact of legacy status at 30 highly selective colleges and concluded that, all other things being equal, legacy applicants got a 23.3-percentage-point increase in their probability of admission. If the applicants' connection was a parent who attended the college as an undergraduate, a "primary legacy," the increase was 45.1-percentage points.

In other words, if a nonlegacy applicant faced a 15-percent chance of admission, an identical applicant who was a primary legacy would have a 60-percent chance of getting in.

The new study is sure to add fuel to the debate over the role of legacy admissions, particularly in determining who gets into the country's most-sought-after colleges. And it sheds light on advantages that colleges themselves may not have even been fully aware of. The author, Michael Hurwitz, controlled for a broader range of variables, such as student character and high-school activities, than had traditional analyses. In doing so, he found that the other, more-common method underestimates the advantage for legacies.

"Some colleges may think this admissions advantage is justifiable or they may use the findings to reshape their policies," says Mr. Hurwitz, a doctoral candidate in quantitative policy analysis at the Harvard Graduate School of Education.

He also looked at the difference between legacies with a primary connection and those with looser connections—a parent who attended graduate school, or a sibling, grandparent, aunt, or uncle who attended as a graduate or undergraduate. He found that the tighter connection, while less common, provides a much larger benefit.

"The takeaway to me is that here's a study that seeks to control for a number of factors and finds that legacy status is even more important than previously thought," says Richard D. Kahlenberg, a senior fellow at the Century Foundation and the editor of Affirmative Action for the Rich: Legacy Preferences in College Admissions. "It's more evidence that this is not a feather on the scale."

For an individual applicant, legacy or nonlegacy status may indeed matter a lot. But Mr. Hurwitz cautions that because of the size of the applicant pools at the sample colleges, legacy admits don't greatly decrease other students' already-long odds of acceptance. Of the 290,000-plus applications he studied, only about 6 percent had legacy status.

An article on his study, "The Impact of Legacy Status on Undergraduate Admissions at Elite Colleges and Universities," was published last month in the journal Economics of Education Review. The data come from 133,236 unique applicants for freshman admission in the fall of 2007 at 30 highly selective private colleges and universities.

Mr. Hurwitz's research found that legacy students, on average, had slightly higher SAT scores than nonlegacies. But he was able to control for that factor, as well as athlete status, gender, race, and many less-quantifiable characteristics. He also controlled for differences in the selectivity of the colleges.

He was able to do so by focusing on the large number of high-school students (47 percent) who submitted applications to more than one of the colleges in the sample. A given applicant's characteristics, like the wealth of their family or strength of their high school, wouldn't vary from college to college. But their legacy status would, and so too might their admissions outcomes. (Mr. Hurwitz also ran an analysis that showed that students who applied to multiple colleges were representative of the overall pool.)

He found that traditional analyses, which control for some of the major quantifiable measures, like SAT scores, but for fewer variables over all, underestimated the legacy advantage. What that means, he says, is that some unquantifiable aspects of legacies' applications—such as life experiences, type of high school, or extracurricular activities—must otherwise work against their chances of admission. But he says, "the data aren't rich enough to tell what that is."

Thomas J. Espenshade, a professor of sociology at Princeton University who has done key research on legacies, says that the new estimates are not widely different from those in previous studies, but that, nonetheless, the larger advantage is notable. Previous studies also had not looked at differences between various familial connections, he says.

A Handed-Down Benefit

Across the board, primary legacies got a greater advantage than secondary legacies. The difference seemed to matter the most at the most-selective colleges in the sample, those with an average base acceptance rate of just under 10 percent. Secondary legacy status at those top-tier colleges conferred an estimated advantage of 8.7 percentage points, while primary legacy gave a 51.6-point advantage.

Legacy status of any kind mattered more at the most-selective and least-selective colleges than it did at those in the middle tiers. The data didn't reveal why, but Mr. Hurwitz thinks that, because such a small proportion of qualified applicants are admitted at the most-selective colleges, any edge over another applicant is magnified—while the less-selective colleges may be most eager to cultivate alumni loyalty and giving.

The data set did not contain information on giving, so Mr. Hurwitz could not look at how much of the legacy advantage comes simply from having relatives who attended a college versus from having relatives who not only attended a college but also donated to it.

Mr. Hurwitz also looked at how students within certain SAT ranges fared against one another. There wasn't a clear-cut pattern, but generally the higher the SAT score, the more legacy status mattered. That finding, Mr. Hurwitz says, seems in line with colleges' argument that legacy status matters the most in deciding between two highly-qualified candidates. "It's easier to justify nudging the student if they're really strong academically," he says.

Legacy and the SAT

The legacy advantage varies depending on what SAT range an applicant falls in. Below are the percentage-point increases in admissions probability for three different types of legacies.

SAT Score

Any Legacy

Primary Legacy

Secondary Legacy

Source: Michael Hurwitz, Economics of Education Review


































Richard H. Shaw, dean of undergraduate admission and financial aid at Stanford University, says his office considers students to be legacies only if one of their parents earned a degree from the university, with an emphasis on undergraduate degrees. Such status is taken into account among many factors, he said, but it certainly does not trump competitive expectations. Mr. Shaw said the university's legacy admits are generally stronger than the median of the admitted class, based on quantitative measures, like test scores, rigor, and grades.

"Stanford also has a high percentage of admitted and enrolling first-generation students each year whose parents did not graduate from a four-year college or university," Mr. Shaw wrote in an e-mail. "We consider access and opportunity a very important principal. We also value intergenerational connections to the Stanford experience."

Several other highly selective colleges declined to comment or did not respond to requests for comment. Mr. Hurwitz will not name the colleges in his sample; he signed an agreement pledging not to do so in exchange for what would otherwise be private data.

Very few colleges, however, have admissions rates approaching anything as low as 10 percent. The study also references other research that has relied on similar data from a group called the Consortium on Financing Higher Education, which comprises all the Ivy League institutions and two dozen other highly competitive private universities and liberal-arts colleges.

Whatever their identity, the colleges in the study are very selective. And Mr. Hurwitz says the findings are most likely to be of relevance to officials and would-be students at similarly-competitive colleges.


1. 22228715 - January 06, 2011 at 07:04 am

Although fairly balanced, there is a slight tone in the article (and perhaps from the researcher) that legacy advantage is a bad thing, a pernicious force that we don't fully realize but should be shocked by. Why? If intergenerational participation is a stated value of the private institution... if a certain segment of higher ed values student participation in the culture... if legacies would be already partially socialized, and that's desired... If most businesses give advantages to repeat customers, even family members of customers... why is it wrong for a handful of selective private institutions to do this for a tiny percentage of their admits?

2. dogood1776 - January 06, 2011 at 08:23 am

Maybe it's why they are elite universities...

3. 11147066 - January 06, 2011 at 08:31 am

This report is incomplete and puzzling. It ignores the fact that the majority of legacy applicants to Ivy League schools, between two thirds and three quarters, are not admitted. What distinguishes between those legacies granted this advantage and those that are rejected? Is it grades, SAT scores, or is it related to the type of legacy connection? The study demonstrates that undergraduate legacies carry more weighit. That has always been a given. What effect does income, level of donations of the parents, and financial need have on whether legacies are given the "boost" this study describes?

4. dank48 - January 06, 2011 at 08:34 am

"Than Was Previously Thought" by whom?

What's next? A study to find out whether on average a Mercedes costs more than a Chevy? This definitely belongs in the NFS category.

5. 11121641 - January 06, 2011 at 10:08 am

I amthe first person in my family to go to college. I amslo the only one EVER to have earned a PhD. And I did it allthrogh public universities. The Ivies rejected me en masse, at both undergraduate and graduate level. The plutocracvy takes care of its own. Nothing new here.

6. research989 - January 06, 2011 at 11:05 am

Congratulations and bravo to Michael Hurwitz on this impressive contribution! This study is methodologically sound and greatly adds to our understanding of the effect of legacy in the admissions process. It is also extremely important, given that elite colleges are major gatekeepers of power in our society and given the rancorous debate around affirmative action.

7. fofo14 - January 06, 2011 at 11:06 am

I used to work as an admissions officer in an elite university. And, what is reported doesn't match up with what I remember. Legacy status did not trump the need for the student to be qualified. There were a handful of legacy and development interest students who were admitted for those reasons, but generally, they met the basic requirements.

I find this work intriguing because it is quantifying the data, but I also think there are social effects that come from having high achieving parents that trickle down to the children, whereby the value to attend a selective university shows up in the kids. So they are prepared and ready when the admissions time comes.

8. ppatrikis - January 06, 2011 at 11:11 am

Are we assume that so-called legacy applicants are automatically unworthy, as is suggested by the study? Might is not be the case that the children of highly educated parents do well?
My double primary legacy son got into the Ivy League institution of his choice because of his brains and his recommendations and not because of the piddling $200/year that his parents donate.

9. mchag12 - January 06, 2011 at 12:13 pm

Some of these comments suggest that there are those that simply skim the article then rush to make comments. Read the article. please. I hope the past admissions officer was not evaluating my kid's application. As for ppatrikis-- yes, the article deals with that too. But it doesn't answer how many students were in the same position without legacy status and did not get in.

10. 12357 - January 06, 2011 at 12:30 pm

Agreed, mchag12.

I don't think the article suggests that legacies are unworthy at all. Besides the fact that there are several references to the interesting patterns between legacy status and other acceptance-relevant characteristics, the methodological point here is, I think, that these applicants' admissions outcomes are being compared to their own admissions outcomes at other schools they applied to. Or at least that is my reading of this paragraph:

"He was able to do so by focusing on the large number of high-school students (47 percent) who submitted applications to more than one of the colleges in the sample. A given applicant's characteristics, like the wealth of their family or strength of their high school, wouldn't vary from college to college. But their legacy status would, and so too might their admissions outcomes. (Mr. Hurwitz also ran an analysis that showed that students who applied to multiple colleges were representative of the overall pool.)"

I also imagine some of the questions posed here in the comments section are addressed more fully in the Economics of Education Review article.

11. bolmanl - January 06, 2011 at 01:07 pm

Is this sentence accurate? "In other words, if a nonlegacy applicant faced a 15-percent chance of admission, an identical applicant who was a primary legacy would have a 60-percent chance of getting in." That would imply that the majority of legacy applicants at elite schools would be admitted, which is not true from reports that I've seen. Another reading is that the legacy applicant's chances increase from 15% to about 22% -- .15 + (.15 * .45).

That seems more likely to me, particularly after last year when my son was rejected at both of the two schools where he was a primary legacy -- Princeton and Yale -- but accepted at Harvard, where he was not.

12. 12104700 - January 06, 2011 at 01:37 pm

bolman1, yes, that sentence is correct (this is the reporter). The researcher looked at percentage-point increases in the probability of admission, not percent increases. The 45-percentage-point advantage for primary legacies does not, however, mean that a majority of them get in. In the researcher's sample of regular and early action applicants, 40.6 percent of primary legacies were admitted. My understanding is that the difference comes from looking at an individual applicant's relative advantage, versus looking at admissions outcomes for an entire pool of students with varying SAT scores, GPAs, essays, life experiences, etc.

13. princeton67 - January 06, 2011 at 01:51 pm

I hope Mr. Hurwitz turns his statistically sophisticated attention to the roles of race or ethnicity, extra-curricular, income, parental education, etc. For example, other variables controlled or equal, is the all-state wrestler in, but the all-state violinist out? the student from the $50,000 family in, but his twin from the $150,000 family out?

14. 88854333 - January 06, 2011 at 02:42 pm

Two comments:

Any university has the right to cultivate its alumni by accepting its graduates' family members. It's a solid means to build endowment support and ongoing loyalty, which strengthens the university.

Affirmative Action is intended to ensure that non-legacies get a fair chance to a quality education, too. Why is AA necessary? Legacies.

15. gavinmoodie - January 06, 2011 at 02:59 pm

Why are so many posters so defensive about legacy admissions. There are 2 problems with the practice, in my view. First, it is a particularly stark example of inter generational transmission of advantage; and second: nepotism.

16. czander - January 06, 2011 at 03:32 pm

Well this clears things up. I always wondered how someone who could not pronounce nuclear was admitted to Yale and Harvard.

17. skocpol - January 06, 2011 at 04:12 pm

There might be nonlinear effects here. A is a primary legacy at X, making her more likely to get into X. College Y decides that this also makes it less likely that she would accept an offer from Y if she gets into X. That would hurt the "yield" of Y, a prestige factor, so they are less likely to admit her, thereby increasing the relative importance of legacy as measured in this study.

We went to Michigan State. Our son applied to University of Michigan. U of M turned him dowm. That suggests that institutional rivalries might also lead to a differential effect.

Of course one of us teaches at an Ivy, and the other at large urban private where he could go for free. He turned down acceptance by our Ivy, and applying for a solid free education, then went to a different Ivy that he liked better -- maybe because it was just the right distance away from home.

A complete understanding of the factors that determine application, admission, and acceptance choices would indeed need to take into account many factors beyond just legacy... Of course I am sure that the author and commenters are aware of this.

18. 11266895 - January 06, 2011 at 05:00 pm

I suspect you are misinterpreting the results of this study a bit. From the abstract it states that: " I estimate that the odds of admission are multiplied by a factor due to legacy status." That's typically how a logit-type analysis is reported and what that means is there is no single increase in probability due to the measured effect. Rather, it is a conditional increase the actual amount of which depends on the measured values of other variables in the model. In other words, a student with virtually no chance of admission to one of the studied institutions who happens to be a primary legacy does NOT receieve a 45 percent bump. Rather they receive a 45 percent increase of their existing probability. To simplify grossly, a kid with a 1 percent chance of admission without being a legacy ends up with a 1.45 percent chance of admission while a kid with a 10 percent chance of admission without being a legacy ends up with a 14.5 percent because they are a legacy. Same percentage increase, but one leads to an overall .45 percent increase while the other leads to a 4.5 percent increase.

19. shallot - January 07, 2011 at 02:12 pm

I also wonder about the yield issue, which I believe is something World News counts in its rankings. I'd think a legacy would be more likely to accept than a non-legacy.

20. pschmidt - January 07, 2011 at 08:18 pm

Re: comment 18, by "11266895." Please see comment 12 and/or the full paper. The story states things correctly. We are talking about a legacy advantage translating into a 45.1 percentage-point increase in odds of admission. A student with a 1 percent chance of being admitted without such a legacy hook would have a 46.1 percent chance of admission with that legacy hook, and thus would have a decent chance of edging out applicants who are a lot brighter and more motivated and have much more to contribute to the college and to society.

21. jherrera - January 07, 2011 at 08:29 pm

The reason why this research is so interesting, is that often, those folks who are admitted based on this legacy are the first ones to scoff at people who are admitted due to affirmative action. Some are ready to eliminate affirmative action while benefitting from this type of advantage themselves. We need to remember how this advantage, often unearned, disadvantages students of color, who don't get in because these folks are.

22. tomian - January 09, 2011 at 01:45 am

Legacy admission is unfair. Not that I wouldn't use it if I had it. But it's always nice when a privileged person can smile and say "I was lucky", rather than casting about for justifications.

If a legacy kid gets into a highly selective college based on their own hard work and talent, he or she will know it. If that same kid gets in with lesser grades and scores, he or she knows that too.

But don't greedy about it and expect those whom you were boosted over to think it's fair. That's unseemly.

23. 11266895 - January 10, 2011 at 11:49 am

Pschmidt, there is simply no way the statistical tool used to model admission, logistic regression, can lead to the statement that ALL legacy students receive a 45.1 percent increase in their overall probability of admission. This article is a mess and fundamentally misrepresents the study.

Add Your Comment

Commenting is closed.

  • 1255 Twenty-Third St., N.W.
  • Washington, D.C. 20037
subscribe today

Get the insight you need for success in academe.