Source: Malcolm Gladwell, "Getting In," The New Yorker / Krueger & Dale (2002) / Fred Glimp, Harvard
Tags: treatment effect, selection effect, Krueger-Dale study, income premium, selective colleges, Penn vs Penn State, happy-bottom-quarter, Fred Glimp, college prestige, does college matter, elite education value
The popular belief is that attending an Ivy League school transforms students into higher achievers (a treatment effect). Research by economists Krueger and Dale suggests the opposite: the Ivies function more like a modelling agency than like Marine Corps basic training. When you compare students who got into top schools but chose less selective ones, the income premium from the elite school disappears. The student, not the school, is the active ingredient.
Treatment effect
The change in outcome caused by an experience or intervention itself. The Marine Corps is a treatment-effect institution: it takes raw recruits and, through basic training, transforms them into soldiers. The experience does the work.
Selection effect
The outcome explained by who was chosen in the first place, rather than by what happened to them afterwards. A modelling agency is a selection-effect institution: you do not become beautiful by signing with the agency; the agency signs you because you are already beautiful.
Krueger-Dale study (2002)
Research by economists Alan Krueger and Stacy Dale that compared students who were admitted to elite schools but chose to attend less selective ones against students who enrolled at the elite school. Finding: the income difference between attending a selective and a less selective school disappears when you control for where the student was admitted, rather than where they enrolled.
Happy-bottom-quarter policy
Admissions strategy attributed to Fred Glimp, Harvard's dean of admissions after Bender. Glimp recognised that every class will have a bottom quarter and tried to fill those spots with students psychologically equipped to thrive there, rather than driven types who would be miserable at non-top rank. In a candid moment, he wondered whether the answer was "Harvard sons" (legacies).
The popular American belief treats elite universities as treatment-effect institutions: brilliant professors, motivated peers, and a prestigious degree combine to confer career advantages a state university cannot match.
Naive studies seem to support this. Comparing two students with the same SAT scores and grades, the one who attends Harvard earns significantly more ten or twenty years later.
Gladwell's point: those studies are flawed, because matching on SAT scores and grades does not make two students equivalent. The one Harvard admitted may also be more ambitious, energetic, and personable, and those intangible qualities, not the school, may explain the income gap.
To isolate the school's effect, Krueger and Dale compared students who were accepted at a top school with students who were accepted at the same top school but chose to attend a less selective one.
Within that matched group, attending the more selective school made no measurable difference to later income.
Krueger's example: students choosing between the University of Pennsylvania (Ivy) and Penn State (state school). Penn graduates earn more on average, but among students admitted to both, those who chose Penn State did just as well.
One might expect the Penn-choosers to be the more ambitious group, which should bias results in Penn's favour. Even so, no gap appeared.
Krueger's interpretation: families see that the average top-school graduate earns around $120,000 versus $90,000 from a moderately selective school, and they attribute the difference to the school. In reality, they are "assigning to the school a lot of what the student is bringing with him to the school."
Krueger noted one important exception: students from the very lowest economic backgrounds do appear to benefit from attending an Ivy League school.
For everyone else, the general rule holds: a hardworking, intelligent person tends to do well regardless of which school they attended.
A strong student can make valuable contacts at Penn State just as at Penn; and if you were good enough to get into Penn, doors tend to open for you regardless of where you enrolled.
Glimp, who succeeded Bender as Harvard's dean of admissions, focused on a practical problem: what happens to the students who end up in the bottom quarter of every class?
His concern: if every admitted student is intensely driven and status-conscious, the inevitable bottom quarter will be miserable. Better to admit some students with "the psychological or what-not tolerance" to be content while not ranking at the top.
This became the "happy-bottom-quarter" policy, sometimes described as notorious.
Glimp's thinking anticipated the Krueger-Dale findings. He believed implicitly that a class's outcomes are largely determined at the point of admission: "if you want to graduate winners you have to admit winners." The school is a selection-effect institution, and its job is to select well.
⚠️ The treatment-effect vs selection-effect distinction is the conceptual spine of Gladwell's argument. Expect exam questions that ask you to apply these terms to the Ivy League context and explain why the distinction matters for evaluating elite education.
⚠️ The flaw in the naive comparison (matching on SAT and GPA alone) is a methodological point worth knowing precisely. Krueger and Dale's fix, comparing students admitted to the same school who chose differently, is the stronger design.
⚠️ The low-income exception is easy to overlook but important: it means the "school doesn't matter" conclusion is not universal. For disadvantaged students, the Ivy League may function as a genuine treatment-effect institution.
⚠️ Glimp's happy-bottom-quarter policy illustrates how seriously Harvard's own administrators took the selection-effect model. They were not trying to transform students; they were trying to select the right ones for each slot.
Q: Explain the difference between a treatment effect and a selection effect, using Gladwell's examples.
A: A treatment effect is the change caused by the experience itself. The Marine Corps is a treatment-effect institution: basic training transforms recruits into soldiers. A selection effect is the outcome explained by who was chosen. A modelling agency is a selection-effect institution: it selects people who are already beautiful. Gladwell argues the Ivy League operates more like the modelling agency.
Q: What was the key methodological improvement in the Krueger-Dale study compared to earlier research on the income premium from elite schools?
A: Earlier studies compared students with the same SAT scores and grades who attended different-tier schools. Krueger and Dale instead compared students who were admitted to the same selective school but chose different institutions, isolating the school's effect from the student's underlying qualities. The income premium disappeared.
Q: For which group of students does attending an Ivy League school appear to confer a genuine advantage, according to Krueger?
A: Students from the very lowest economic backgrounds. For most other students, the general finding is that personal qualities rather than institutional prestige drive long-term outcomes.
Q: What was Fred Glimp's "happy-bottom-quarter" policy, and what does it reveal about Harvard's self-understanding?
A: Glimp tried to ensure the bottom quarter of each Harvard class was filled with students psychologically comfortable not being at the top, rather than intensely driven types who would be distressed by their rank. It reveals that Harvard's own admissions leadership operated on the assumption that outcomes were largely determined by who was admitted, not by what the school did to them, a selection-effect model.
treatment effect vs selection effect, Krueger and Dale 2002, income premium elite colleges, does college prestige matter, Penn vs Penn State study, happy-bottom-quarter policy, Fred Glimp Harvard, selection bias college outcomes, Ivy League value, college earnings research, modelling agency vs Marine Corps analogy, elite education ROI