Waitlist Mechanics: Calibration Data for College Admissions Simulation

waitlist_mechanics.md

Waitlist Mechanics: Calibration Data for College Admissions Simulation

Overview

This document compiles waitlist statistics for the 55 colleges in the simulation, drawing primarily from Class of 2029 (2024-2025 CDS) and Class of 2028 (2023-2024 CDS) data. Waitlist behavior varies enormously by school and year, making it one of the hardest parameters to calibrate.

Key aggregate finding: According to NACAC, colleges on average admit 20% of students off the waitlist. At the most selective institutions (under 25% overall admit rate), that figure drops to roughly 7-8%. For the Class of 2028, 26% of students accepting a waitlist spot were admitted nationally, up from 23% for the Class of 2027.

Data Sources

• IvyWise waitlist admission rates (Class of 2029 CDS data)
• College Kickstart Class of 2028/2029 waitlist notification dates and stats
• Individual university CDS reports (2023-2024 and 2024-2025)
• College Transitions Dataverse waitlist statistics
• Ivy Coach historical waitlist data
• Harvard Crimson, Brown Daily Herald, Yale Daily News reporting

Tier 1: HYPSM

Notes: - Harvard does not report waitlist numbers in CDS. The 75 admitted for C/O 2029 was unusually high (vs 41 for C/O 2028, 27 for C/O 2027), partly due to international visa uncertainty causing extra admits. - Yale admitted 0 from waitlist for C/O 2028, then 23 for C/O 2029 -- highly volatile. - Stanford's C/O 2028 was an outlier year (15% pull rate); C/O 2029 returned to ~6%. - MIT's waitlist is small (~500-620 offered) but pull rate is very low (1.8% for C/O 2029). - Princeton admitted 40 for C/O 2029, consistent with historical range of 0-164.

Tier 2: Ivy+

Notes: - Brown's C/O 2029 waitlist activity was unusually high; Herald poll data suggests ~230 admitted, far above the historical range of 15-120. This may reflect post-SFFA enrollment shifts. - Columbia does not include waitlist data in its CDS. Estimated ~3-4% pull rate based on peer comparison. - UChicago does not report waitlist data in CDS. Estimated based on peer institutions. - Caltech's 24% pull rate is an outlier -- small class (235) means even small yield misses require significant WL pulls. - Cornell has the largest Ivy waitlist by volume (~6,000 acceptors, ~350-390 admitted). - Dartmouth swings from 0 (C/O 2028) to 29 (C/O 2029) -- extremely volatile.

Tier 3: Near-Ivy

Notes: - Carnegie Mellon stands out with an enormous waitlist (~10,000 acceptors) but admits only ~32, giving a 0.3% pull rate. This is the lowest among all 55 schools. - Vanderbilt states ~10% of its enrolling class (1,600-1,700) comes from the waitlist, implying ~160-170 admitted annually. - Georgetown's 8.1% pull rate is unusually high for its selectivity tier, likely reflecting its non-binding EA round creating yield uncertainty. - WashU's pull rate (7.6%) is consistent with its strong ED reliance + yield protection needs.

Tier 4: Selective Privates & Top LACs

Notes: - Tufts' 35.7% pull rate for C/O 2029 is remarkably high and may reflect yield protection dynamics -- Tufts admits conservatively in RD (expecting melt), then pulls heavily from WL. - Tulane swings from 2.0% (C/O 2028) to 18.9% (C/O 2029) -- extreme volatility likely driven by its heavy EA/ED strategy. - USC is unique among selective schools in not maintaining a waitlist at all. The simulation currently has waitlistPullRate:0.05 for USC which should be set to 0. - Swarthmore consistently admits 0 from waitlist. - Williams swings from 3 (C/O 2028) to 25 (C/O 2029).

Tier 5: Top Public / LAC

Notes: - UC Berkeley swings wildly: 24.7% pull rate (C/O 2028) to 0.3% (C/O 2029). The UC system's waitlist behavior is heavily influenced by system-wide yield dynamics. - UCLA consistently pulls large numbers (1,211 for C/O 2029) due to its enormous class size and relatively low yield (18%). - Wesleyan shows extreme volatility: 201 admitted (C/O 2028) to just 5 (C/O 2029). - Michigan's massive waitlist (~18,000 acceptors) yields ~950-970 admits consistently (~5.2%).

Tier 6: Selective Public

Notes: - UW-Madison shows extreme swings: 61.4% (C/O 2028) to 6.4% (C/O 2029). The 2028 figure was likely driven by an enrollment shortfall. - UW Seattle similarly volatile: 72.4% to 20.0%. - Virginia Tech: 0 admits (C/O 2028) to 1,524 (C/O 2029) -- complete reversal. - Purdue: 466 admits (C/O 2028) to just 9 (C/O 2029). - Public universities show the most year-to-year volatility in waitlist behavior due to large class sizes and unpredictable yield from in-state vs out-of-state students.

Waitlist Yield Rates

Waitlist yield (the fraction of WL admits who actually enroll) is generally lower than Regular Decision yield for most schools. The gap varies by tier:

Why WL Yield Is Lower Than RD Yield

1. Double deposits: WL admits have already committed elsewhere and paid a deposit (~$500). Switching requires forfeiting that deposit and disrupting housing/roommate plans.
2. Timing: Late offers (May-July) mean students have emotionally committed to their May 1 choice.
3. Alternative options: WL admits have strong profiles and are enrolled at competitive alternatives.
4. Financial aid uncertainty: WL admits may receive less favorable aid packages since institutional aid budgets are largely committed by waitlist season.
5. Logistical friction: Orientation scheduling, housing assignments, and course registration may already be underway at the committed school.

Estimated Waitlist Yield by Tier

Key exception: At HYPSM, WL yield approaches RD yield because few students would turn down Harvard or Stanford for wherever they initially committed. The yield gap widens as prestige decreases.

Implications for Simulation

The current resolveWaitlist() function does not model WL yield separately -- it assumes students always accept a WL offer from a higher-tier school (line 2740: if (wlTier < currentTier || !s.committed_to)). This is approximately correct for tier upgrades but ignores same-tier lateral moves and financial considerations. A more realistic model would apply a WL yield rate that is ~75-95% of RD yield depending on tier.

Waitlist Timeline

When Waitlist Offers Are Made

Waitlist decisions go out simultaneously with RD decisions, typically in late March. Students are asked to opt in (accept the WL spot) within 1-2 weeks.

When Waitlist Admits Are Notified

Most waitlist admits at elite colleges hear back within the first two weeks of May, immediately after May 1 when schools can assess their yield shortfall. At public universities with larger classes, waitlist activity can extend into July.

May 1 Dynamics

The May 1 deadline creates a cascade: 1. Students deposit at their preferred school 2. Schools tally deposits vs targets 3. Under-enrolled schools activate waitlists 4. WL admits who accept release their seat at their May 1 school 5. That school may then activate its own waitlist (cascade effect)

This cascade is modeled in the simulation's resolveWaitlist() function through its 5-iteration loop (line 2696).

Demographic Patterns on Waitlists

ALDC Students Are Rarely Waitlisted

Research from the SFFA v. Harvard trial (Arcidiacono & Kinsler, Journal of Labor Economics, 2022) revealed that hooked applicants receive substantial admissions advantages in the initial round, making waitlisting rare for these groups:

Implication: The waitlist pool is disproportionately composed of non-hooked, academically strong applicants who narrowly missed the admit threshold. ALDC students are almost always admitted outright or rejected -- they are rarely placed in the ambiguous middle ground of the waitlist.

Who Gets Pulled from the Waitlist?

While colleges state they do not rank waitlist applicants, they use WL pulls strategically to address class-shaping needs:

1. Institutional needs: Undersubscribed majors (e.g., classics, physics) may receive priority
2. Gender balance: Schools pull to balance gender ratios
3. Geographic diversity: Under-represented regions get priority
4. Financial considerations: Full-pay students may receive quiet preference, as aid budgets are largely committed
5. Demonstrated interest: Students who submit LOCIs signal enrollment intent, reducing yield risk
6. Specific talents: Musicians, athletes in non-recruited sports, specific extracurricular needs

Post-SFFA Considerations

After the 2023 Supreme Court ruling ending race-conscious admissions, some schools have used waitlists as a tool for maintaining class diversity through race-neutral means (geography, socioeconomic status, first-generation status). Brown's unusually high WL activity for C/O 2029 may partly reflect this dynamic.

Simulation Implications

The simulation should model waitlist composition as: - Lower rate of hooked students on WL (~5-10% vs ~25-40% in admitted class) - Higher academic credentials on average (near-miss admits) - WL pulls slightly favor full-pay students and those with demonstrated interest

LOCI (Letter of Continued Interest) Effectiveness

What Is a LOCI?

A formal letter sent to an admissions office after being waitlisted, expressing continued interest and providing updates on achievements since the original application.

Effectiveness Data

• LOCIs are fairly rare -- most waitlisted students treat a waitlist as a rejection and move on
• Because they are rare, they are actually read and considered by admissions officers
• Schools that track demonstrated interest (e.g., Tufts, WashU, Boston College) give more weight to LOCIs
• Some schools explicitly invite LOCIs: Georgetown, Vanderbilt, Emory, Notre Dame
• Others discourage additional materials: MIT, some Ivies

Timing Best Practices

• Send initial LOCI within a few days of waitlist notification
• Brief update in late April with new achievements
• Do not send more than 2-3 communications total -- excessive contact is counterproductive

Schools That Value Demonstrated Interest via LOCI

Simulation Implications

LOCI effects are too granular to model directly. The concept maps to demonstrated interest, which the simulation captures through ED/EA multipliers and the utility-function fit score. Students who apply early already demonstrate interest; waitlist-stage interest is implicitly captured by the yield model.

Historical Trends: Waitlists Are Growing

Application Volume Driving Waitlist Growth

The explosion in college applications has made yield prediction harder, increasing reliance on waitlists:

• Average applications per student: 6.8 (up from ~3 in the early 2000s)
• Total applications to selective schools have roughly doubled in 10-15 years
• Common App, Coalition App, test-optional policies, and direct-admit programs all contribute
• More applications = more cross-admits = more yield uncertainty = larger waitlists

NACAC Data on Waitlist Usage

• 43% of all colleges used a waitlist (2018-19 NACAC survey)
• 82% of schools with <50% acceptance rate maintained a waitlist
• 48% of private vs 34% of public institutions use waitlists
• National avg WL admission rate: ~20% (2018-2020)
• Most selective institutions: ~7% average

Year-Over-Year Trends (Class of 2027 vs 2028)

COVID Effect (C/O 2024, enrolling fall 2020)

The pandemic created unprecedented yield uncertainty, leading to massive waitlist usage:

Post-COVID, waitlist activity has not fully returned to pre-pandemic levels -- schools learned that larger waitlists provide better insurance against yield uncertainty.

Growing Waitlist Offer Sizes

Schools are offering WL spots to more applicants over time:

This expansion is partly strategic (more class-shaping options) and partly a consequence of larger applicant pools.

Summary of Waitlist Dynamics

Key Patterns

1. Extreme year-to-year volatility: Most schools show 3-10x variation between years. UC Berkeley went from 1,191 admits to 26. Virginia Tech from 0 to 1,524. This makes any single year's data unreliable for calibration.

2. Inverse relationship with yield: Schools with lower yield rates (WashU 35%, Rice 43%, Colgate 28%) tend to use waitlists more actively. High-yield schools (Harvard 84%, MIT 87%, UChicago 86%) rarely need to pull.

3. Small classes amplify volatility: LACs with 400-600 seat classes (Williams, Pomona, Swarthmore) can swing from 0 to 60 admits based on just a few yield percentage points.

4. Public universities have the largest waitlists: Michigan (~18,000 acceptors), UW-Madison (~7,000), Virginia Tech (~11,000). But pull rates are unpredictable.

5. Some schools never use their waitlist: Swarthmore consistently admits 0. UIUC admitted just 1 for C/O 2029.

6. USC has no waitlist: Unique among the 55 schools -- students are admitted or denied with no waitlist option.

Schools That Don't Report Waitlist Data in CDS

• Harvard (does not report any WL stats)
• Columbia (omits WL section from CDS)
• UChicago (omits WL section from CDS)
• NYU (hasn't published WL stats since C/O 2017)
• Duke (sporadic reporting)
• Wake Forest (does not report)
• Northeastern (does not publish)
• Vanderbilt (delays reporting)

Recommendations for Simulation Calibration

Recommended waitlistPullRate and wlTypicalPulled Values

Based on multi-year averages (prioritizing C/O 2029 data where available, cross-referenced with C/O 2028 and historical patterns), here are recommended values. The waitlistPullRate is the fraction of WL acceptors admitted; wlTypicalPulled is the absolute number admitted in a typical year.

Tier 1: HYPSM

Tier 2: Ivy+

Tier 3: Near-Ivy

Tier 4: Selective Privates & Top LACs

Tier 5: Public Elite / LAC

Tier 6: Selective Public

Critical Corrections Needed in sim.js

1. USC must have waitlist disabled (waitlistPullRate: 0.0, wlTypicalPulled: 0) -- USC does not maintain a waitlist.

2. MIT is over-calibrated: Current pullRate: 0.08, wlTypicalPulled: 40 vs actual 1.8% and ~10 pulled. Reduce significantly.

3. Michigan is under-calibrated: Current wlTypicalPulled: 60 vs actual ~960. The simulation's small model scale means this needs proportional adjustment, but the direction is clear.

4. UC Berkeley is under-calibrated: Current wlTypicalPulled: 26 reflects only C/O 2029 (an outlier low year). Multi-year average is ~600.

5. UW-Madison and UW Seattle: Current wlTypicalPulled values are far below actual averages. Public universities use waitlists as major enrollment management tools.

6. Tufts needs major upward revision: Its yield protection strategy leads to aggressive waitlist pulls (35.7% pull rate for C/O 2029), far above the current 5%.

7. Brown's pull rate and typical pulled are both underestimated based on C/O 2029 data.

Simulation-Specific Considerations

The wlTypicalPulled value in the simulation is scaled by SIM.config.studentsPerSchool / 20 (see resolveWaitlist() at line 2676). This means the simulation already adjusts for its smaller scale. The recommended wlTypicalPulled values above are real-world absolute numbers that the simulation's scaling factor will handle.

The 5% deficit threshold (deficitPct <= 0.05) in resolveWaitlist() means schools must be under-enrolled by at least 5% before activating their waitlist. This is reasonable -- in reality, schools typically don't go to the waitlist unless they're meaningfully under-enrolled.

Volatility Factor

Given the extreme year-to-year volatility (2-10x swings for many schools), the simulation could benefit from adding randomness to waitlist behavior. Consider:

// Add noise to wlTypicalPulled to model year-to-year volatility
const volatilityFactor = 0.5 + rand() * 1.0; // Range: 0.5x to 1.5x
const adjustedPulled = Math.round(wlData.typicalPulled * volatilityFactor * scaleFactor);

This would produce more realistic variation across Monte Carlo runs.

References

Primary Data Sources

• IvyWise. "Waitlist Admission Rates." https://www.ivywise.com/blog/waitlist-admission-rates/
• College Kickstart. "Class of 2029 Waitlist Notification Dates and Stats." https://www.collegekickstart.com/blog/item/class-of-2029-waitlist-notification-dates-and-stats
• College Kickstart. "Class of 2028 Waitlist Notification Dates and Stats." https://www.collegekickstart.com/blog/item/class-of-2028-waitlist-notification-dates-and-stats
• College Transitions. "Waitlist Statistics." https://www.collegetransitions.com/dataverse/waitlist-statistics
• SupertutorTV. "Chances of Getting Off the Waitlist in 2024." https://supertutortv.com/videos/chances-of-getting-off-the-waitlist-in-2024-at-top-universities/ (multi-year CDS data for 50+ schools)
• Collegiate Gateway. "College Waitlists: What Are Your Chances?" https://collegiategateway.com/college-waitlists-what-are-your-chances-of-being-accepted/ (5-year tables per school)

Historical and Ivy-Specific

• Ivy Coach. "Ivy League Previous Years Waitlist Acceptance Rates & Statistics." https://www.ivycoach.com/the-ivy-coach-blog/ivy-league/ivy-league-waitlist-acceptance-history/
• Ivy Coach. "USC Waitlist Acceptance Rate." https://www.ivycoach.com/the-ivy-coach-blog/college-admissions/usc-waitlist/ (USC does not maintain a waitlist)
• Harvard Crimson. "Class of 2029 yield tops 83%." October 2025.
• Brown Daily Herald. "Brown admits more waitlisted students for fall 2025." September 2025.
• Yale Daily News. "Yale admits 4.59 percent of applicants." March 2025.
• Vanderbilt Hustler. "Record-low 4.7% admitted to Class of 2029." April 2025.

NACAC and National Data

• NACAC. "State of College Admission Report." https://www.nacacnet.org/state-of-college-admission-report/
• NACAC Chapter 2 (2019): 43% of colleges used waitlists; 82% of most selective schools. https://nacacnet.org/wp-content/uploads/2022/10/soca2019_ch2.pdf

Demographic and ALDC Research

• Arcidiacono, P. & Kinsler, J. "Legacy and Athlete Preferences at Harvard." Journal of Labor Economics, Vol 40, No 1 (2022). https://public.econ.duke.edu/~psarcidi/legacyathlete.pdf
• XFactor Admissions. "Who Actually Clears a Top College's Waitlist?" https://xfactoradmissions.com/basic-guide-to-college-admissions/who-actually-clears-a-top-colleges-waitlist

LOCI and Demonstrated Interest

• College Essay Guy. "How to Write a Great Letter of Continued Interest." https://www.collegeessayguy.com/blog/letter-of-continued-interest
• IvyWise. "When Should You Write a Letter of Continued Interest?" https://www.ivywise.com/blog/when-should-you-write-a-letter-of-continued-interest/

CDS Reports

• Individual university Common Data Sets (2023-2024, 2024-2025).
• Cornell CDS: https://irp.cornell.edu/common-data-set
• Bowdoin CDS: https://www.bowdoin.edu/ir/common-data/index.html
• NYU CDS: https://www.nyu.edu/content/dam/nyu/institutionalResearch/documents/cds-on-website/
• Vanderbilt Admissions Blog: https://admissions.vanderbilt.edu/apply/waitlist/