GiveBetter x New York City

Per Scholas

Tuition-free technology job training and placement for adults

Research time: 3 min on Claude Opus 5.5
  • Research — reviewed programs, finances and impact evidence.
  • Modeling — estimated costs, QALYs and uncertainty.
  • Source audit — checked claims, assumptions and calculations.

Updated: 2026-10-18

Donate

Summary

What do they do? Per Scholas, founded in the Bronx, runs free, full-time technology training courses for adults who are unemployed or in low-wage work. It now trains over 4,600 learners a year at more than 20 campuses, and helps graduates find jobs with employer partners. This in-depth review prices the organization’s whole cost per learner against earnings gains from the randomized WorkAdvance trial at its Bronx site.

Why we’re interested in this organization:

  • In a randomized trial of about 690 Bronx applicants, Per Scholas raised earnings in years 2–5 and 7–8; in year 6 participants earned $6,281 (20%) more than the control group.

  • Arnold Ventures rates the program Near Top Tier, and its effects were the strongest of the four WorkAdvance providers.

  • The trial cost only about $5,700 per participant to run, low for job training.

Our main reservations:

  • Our working estimate is about $90K per DALY; all of it is income, and by year 10 the earnings gain had disappeared.

  • Today’s whole-organization cost is about $13,497 per learner, more than double the trial’s inflation-adjusted cost, after rapid expansion to over 20 campuses.

  • The three other WorkAdvance providers showed no lasting effect, so the result may not transfer to every campus or to today’s curriculum.

What do you get for your dollar? $90K per DALY in New York City. DALY is our one-year comparison unit. QALY-based health gains and income-welfare equivalents retain their original assumptions; they are not measured clinical DALYs. Whole-organization cost per learner against randomized earnings gains at the Bronx site; income-only welfare comparison, not a verified marginal gift offer.

Annual earnings in years 1–10 from wage records; year-10 effect null.

1. What do they do?

Per Scholas (Per Scholas Inc, EIN 04-3252955) offers free, full-time technology courses, mainly in IT, to adults who are unemployed or in low-wage work. It works with employers on curricula and hiring. Founded in the Bronx, it now operates more than 20 metro campuses. It reported over 4,600 learners and more than 2,600 graduates in 2023. A donation supports the whole organization.

2. Monitoring and information sharing

Per Scholas was one of four providers in MDRC’s WorkAdvance randomized trial, which randomly assigned 2,564 applicants in 2011–13; about 690 were at the Per Scholas Bronx site. Earnings come from administrative wage records, now followed for ten years. Per Scholas publishes annual reports with learner counts and placement wages. We did not find New York-only learner counts or a current cost per learner; we have not contacted the organization.

3. Qualitative assessment

The Per Scholas site produced the strongest effects in WorkAdvance: earnings rose in years 2 to 5 and 7 to 8, with a year-6 gain of $6,281 (20%) over a $32,122 control mean. MDRC’s 10-year report (2025) found no year-10 effect at Per Scholas, while St. Nicks Alliance, another New York provider, showed later gains. The other two sites showed only short-lived effects. Arnold Ventures rates Per Scholas Near Top Tier. The trial version included post-placement advancement coaching; today’s larger, multi-campus program may differ.

4. What do you get for your dollar?

Our central estimate is $90K per DALY, using one modeled equivalent year. It is an income-only comparison: we credit no health gain.

Cost. CY2023 total expense was $62.1M for about 4,600 learners: $13,497 each, for the whole organization. In the trial, operating cost was about $5,700 per participant (net $3,500 after services controls received), or about $7,524 today.

Benefit. We credit an average gross gain of $5,000 a year in years 2–8, between the significant early gains and the year-6 peak of $6,281, and nothing after, because the year-10 effect was zero. 60% counts as net household welfare, on a $32,000 base. We keep 0.7 for transfer to today’s program, 0.85 for displacement and 0.9 for NYC residence.

Range. At the trial cost the estimate is $50K per DALY. Using the year-6 gain throughout gives $73K per DALY; a $3,000 average gives $148K per DALY; halving the transfer gives $181K per DALY. If the trial effect does not hold at current scale, there is no credited benefit.

Model, assumptions and sensitivity

CY2023 whole-organization expense divided by about 4,600 learners.

NYC-metro learners; 0.9 of earning years in the metro.

Price = cost per learner ÷ [0.5 × discounted years(2–8) × ln(1 + 0.6 × gain ÷ base) × transfer × non-displacement × NYC share]. Health = 0.

CY2023 total expense
62085666 USD (observed). Form 990. [propublica]
Learners 2023
4600 learners (observed). Over 4,600 trained in 2023, citing the annual report. [wiki] [ps-2024]
Trial operating cost
5700 USD per participant (observed). WorkAdvance cost analysis summary. [mdrc-costs]
Average gross gain, years 2–8
5000 USD per year (judgment). Between significant early gains and the year-6 $6,281; year 10 null. [av-workadvance] [mdrc-10yr]
Base earnings
32000 USD per year (observed). Year-6 control-group mean $32,122. [av-workadvance]
Net welfare share
0.6 share (judgment). After taxes and lost benefits.
Transfer to current program
0.7 share (judgment). Same organization and city, but a larger, changed program. [ps-2024]
Non-displacement
0.85 share (judgment). Some entry-level IT jobs are taken from other applicants.
NYC residence share
0.9 share (judgment). Adult learners mostly stay in the metro.

Whole-organization cost per learner, trial gains for years 2-8: Cost: $1.3M; New York City health QALYs: 0.00; all-population health QALYs: 0.00. Income-equivalent healthy years: 14.95. CY2023 expense over about 4,600 learners; $5,000 average gross gain in years 2-8 (60% net), none after; 0.7 transfer, 0.85 non-displacement, 0.9 NYC residence.

Program cost measured in WorkAdvance: Cost: $752K; New York City health QALYs: 0.00; all-population health QALYs: 0.00. Income-equivalent healthy years: 14.95. Operating cost of $5,700 per participant at Per Scholas in the trial, inflated to 2025.

Year-6 gain ($6,281) in every year 2-8: Cost: $1.3M; New York City health QALYs: 0.00; all-population health QALYs: 0.00. Income-equivalent healthy years: 18.57. Upper bound using the largest reported annual impact.

Average gain $3,000: Cost: $1.3M; New York City health QALYs: 0.00; all-population health QALYs: 0.00. Income-equivalent healthy years: 9.13. Lower average across years with and without significant impacts.

Gains last through year 10: Cost: $1.3M; New York City health QALYs: 0.00; all-population health QALYs: 0.00. Income-equivalent healthy years: 18.68. Ignores the null year-10 result.

Today’s program half as effective: Cost: $1.3M; New York City health QALYs: 0.00; all-population health QALYs: 0.00. Income-equivalent healthy years: 7.47. Scale-up to 24 campuses and changed curriculum dilute the Bronx trial effect.

More displacement of other IT job seekers: Cost: $1.3M; New York City health QALYs: 0.00; all-population health QALYs: 0.00. Income-equivalent healthy years: 10.55. Entry-level IT jobs partly taken from other applicants.

No effect at current scale: Cost: $1.3M; New York City health QALYs: 0.00; all-population health QALYs: 0.00. Income-equivalent healthy years: 0.00. The other three WorkAdvance sites found no lasting effect.

Counterfactual: Randomized control applicants with access to other training and jobs.

Attribution: A gift is credited with the transferred average effect of an added learner at whole-organization average cost; whether a gift adds a New York learner is unverified.

Dominated by transfer from the 2011–13 Bronx site to today’s scaled program and by current cost per learner.

Sensitivity

  • Program cost measured in WorkAdvance: $50K per DALY. Operating cost of $5,700 per participant at Per Scholas in the trial, inflated to 2025.
  • Year-6 gain ($6,281) in every year 2-8: $73K per DALY. Upper bound using the largest reported annual impact.
  • Average gain $3,000: $148K per DALY. Lower average across years with and without significant impacts.
  • Gains last through year 10: $72K per DALY. Ignores the null year-10 result.
  • Today’s program half as effective: $181K per DALY. Scale-up to 24 campuses and changed curriculum dilute the Bronx trial effect.
  • More displacement of other IT job seekers: $128K per DALY. Entry-level IT jobs partly taken from other applicants.
  • No effect at current scale: no finite positive estimate. The other three WorkAdvance sites found no lasting effect.

Unresolved inputs

  • New York learners and cost per added learner.
  • Evidence for today’s curriculum and campuses.
  • Whether added funding creates added New York cohorts.

Health and income breakdown

For the modeled $1.3M: 0.00 health QALYs + 14.95 income-equivalent healthy years in New York City: $90K per DALY. Income equivalents are welfare comparisons, not observed QALYs or DALYs.

Method and caveats.

5. Funding and previous grants

Per Scholas reported CY2023 revenue of $83.3M, including $63.5M of contributions and $17.8M of program service revenue, and expenses of $62.1M, up from $55.0M in 2022 and $36.1M in 2021. Training is free to learners; funding comes from philanthropy, employers and public workforce grants. We have not seen a costed plan for added New York learners.

Annual expenses

Organization-level spending, including programs, administration and fundraising. The research list averages three comparable, consecutive full fiscal years when available.

  • FY 2021: $36.1M; Per Scholas Inc, EIN 04-3252955, 12-month period, IRS Form 990 total functional expenses, calendar year. Source
  • FY 2022: $55.0M; Per Scholas Inc, EIN 04-3252955, 12-month period, IRS Form 990 total functional expenses, calendar year. Source
  • FY 2023: $62.1M; Per Scholas Inc, EIN 04-3252955, 12-month period, IRS Form 990 total functional expenses, calendar year. Source

6. Sources

  1. Effects of Sector-Focused Training After 10 Years: Findings from the WorkAdvance Evaluation. MDRC (Yusim, Schaberg, Tessler and Ubalijoro). Published: 2025-09-01; retrieved: 2026-10-18.
  2. Long-term follow-up of the WorkAdvance study. Arnold Ventures. Published: not stated; retrieved: 2026-10-18.
  3. An exemplary training programme: will it work here? (WorkAdvance costs summary). What Works Centre for Local Economic Growth. Published: not stated; retrieved: 2026-10-18.
  4. Can sector strategies promote longer-term effects? Three-year impacts from WorkAdvance. U.S. Department of Labor CLEAR. Published: not stated; retrieved: 2026-10-18.
  5. Per Scholas (2023 learners and graduates, citing the annual report). Wikipedia. Published: not stated; retrieved: 2026-10-18.
  6. Per Scholas 2024 Annual Report. Per Scholas. Published: not stated; retrieved: 2026-10-18.
  7. Per Scholas Inc, Form 990 filings (EIN 04-3252955). IRS, hosted by ProPublica. Published: not stated; retrieved: 2026-10-18.
  8. Cost-effectiveness: health and log-income comparison. Coefficient Giving. Published: not stated; retrieved: 2026-10-18.