Only 42% of Alpha School Students Hit Its 2x Learning Promise

A New York Times investigation found 42% of Alpha School students met its 2x learning promise, and critics dispute how the growth figure is calculated.

By USA 24 Editorial TeamOctober 25, 20264 min read

Empty classroom at dawn with rows of wooden desks, each holding a tablet, one screen glowing softly by tall windows. — Illustration: USA 24

ShareX Facebook LinkedIn Reddit Email

Key takeaways

The New York Times found that 42% of Alpha School students progressed twice as fast in both math and reading, the 2x standard featured in the chain’s marketing.
Only 37% of the roughly one-fifth of K-8 students who scored at or below the national 50th percentile in math last fall reached the 2x target by winter.
Critics say a student who gains 14 points when 7 are expected is labeled 2x, while the parent company says its students grow 2.6 times faster than peers on nationally normed MAP tests.
Tuition at most campuses is reported at about $40,000 per year, and districts in Texas and South Carolina have adopted the platform, reaching about 2,600 public school students.
A Pennsylvania Department of Education review of a related charter application called the model “untested” and said it lacked evidence of alignment with state standards.

What the Times found about the 2x claim

Alpha School 2x growth claims are now under national scrutiny. A New York Times investigation published October 23 found that only 42% of its students met that standard in both math and reading. The findings matter most to parents paying tuition, to public districts that have adopted the software, and to state regulators weighing similar AI tutors.

Dana Goldstein’s headline for the Times piece sums up the problem: “Students at an A.I. School Were Supposed to Learn Faster. Many Aren’t.” The figures in this story come from that reporting as summarized by other outlets, including an AI Weekly summary.

The weakest math students fared worst. About one-fifth of Alpha’s K-8 students scored at or below the national 50th percentile in math last fall. Only 37% of that group reached the 2x target by winter. A Substack essay by Benjamin Riley cites the same breakdown.

That pattern cuts against the central promise of a software-driven model, which is supposed to adapt most to students who are behind. The students who start furthest back are the ones least likely to hit the advertised target.

How the 2x figure gets calculated

The fight is less about the headline percentage than about the arithmetic behind the word “2x.” Critics point to a student who gains 14 points in math when 7 points were expected, and that student is labeled 2x. The same raw gain can read as impressive or ordinary depending on the baseline someone chooses.

Alpha’s parent company, 2 Hour Learning, says its students grow 2.6 times faster than peers on nationally normed MAP tests, according to the Evanston RoundTable. The school says those results rely on internal analyses of data that have not been independently reviewed, as Wikipedia’s entry on Alpha School describes its position.

One critic goes further. A piece in The Argument alleges that Alpha averages student results to produce its school-wide 2.6x figure. That is an allegation, not an established finding.

Kelsey Piper, an education writer, was blunter in her reporting for The Argument, saying the claims were built on numbers “that have been massaged to the point of being useless,” as the Evanston RoundTable quoted her. A ratio can look dramatic while the underlying change is small, so the headline number cannot carry a decision on its own.

A teacherless model built around a two-hour block

Alpha’s pitch is that students master core academics in about two hours a day. In its K-8 model, AI-based software covers the morning academic block, and afternoons go to workshops. The Times reporting describes the K-8 model as teacherless.

The “AI” label needs context. According to the founders, the technology resembles adaptive learning apps such as IXL or Khan Academy rather than large language models. That distinction matters to families, because adaptive practice software and a chatbot tutor make different claims about how students learn. Critics also note that few peer-reviewed studies validate the approach.

The network has grown quickly on paper. It reportedly operates about 50 campuses with roughly 1,500 students, and planned openings include Chicago, Atlanta, Charlotte, Raleigh, Santa Monica, Palo Alto and the East Bay. Tuition at most campuses is reported at about $40,000 per year.

Alpha is a for-profit company backed by software billionaire Joe Liemandt. According to a summary of the Times reporting by Suaragarut, it has also drawn attention from business leaders and the Trump administration.

Public districts in Texas and South Carolina are already buying in

The expansion is not limited to private tuition. A company document the Times reviewed shows that school districts in Texas and South Carolina adopted Alpha’s software platform, reaching about 2,600 public school students. Those districts now face the same question as parents: do the outcomes justify the marketing? Whether any district will change its use of the software is not yet known.

Regulators have raised questions before. A Pennsylvania Department of Education review of a related charter application called the model “untested” and said it lacked evidence of alignment with state standards. That review concerned a charter application rather than Alpha’s private campuses, but it shows the same questions are reaching public authorities.

Research credibility is also contested. MIT’s research on Alpha is funded by Alpha, and the Evanston RoundTable describes Harvard’s work as a collaboration with the school. The status of both projects is unresolved. The coverage used for this story does not include a response from Alpha to the findings.

How to check an AI school’s growth claims before you enroll

These steps are practical suggestions, not findings from the investigation.

Ask for independent outcome data, not only internal analyses.
Ask how the growth metric is calculated, including what baseline the expected gain is measured against.
Compare the school’s results with your state’s standardized test scores for the same grades.
Ask what share of students leave the program and the main reasons they give.
Ask whether the 2x figure covers all students or a selected group, and what happens to students who start behind.
If you work for a district, request the evaluation before renewing a contract, and confirm the software aligns with state standards.

A school that answers these questions with data is easier to trust than one that leads with a single ratio.

What to watch next

Several threads will determine how much this story matters. The first is whether Alpha releases the calculation behind its 2x figure in a form outside researchers can check. The second is whether the Texas and South Carolina districts review their contracts or publish their own results.

State officials are also testing AI claims in other settings. Texas has sued OpenAI and Sam Altman, joining eight states in a case over ChatGPT, which shows that state officials are willing to take AI companies to court. Expect similar scrutiny to reach AI tutoring models as more campuses open and more public districts sign contracts.

Frequently asked questions

Is Alpha School legit?

Alpha School is a for-profit network whose results are disputed rather than settled. The New York Times found that only 42% of students met the 2x standard, while the school says its students grow 2.6 times faster on MAP tests. Critics argue the 2x figure is calculated in a way that inflates results, so parents should ask for independent outcome data before enrolling.

How much does Alpha School cost?

Tuition at most campuses is reported at about $40,000 per year. The school pitches mastery of core academics in roughly two hours a day, with afternoons given to workshops. Families should compare that cost with state test results for the schools they would otherwise attend.

How is the Alpha School 2x figure calculated?

Critics say a student who gains 14 points in math when 7 points were expected is labeled 2x. The school’s parent company says its students grow 2.6 times faster than peers on nationally normed MAP tests. One critic alleges that Alpha averages results to produce its school-wide figure, which is an allegation rather than an established finding.

Are Texas and South Carolina public schools using Alpha’s software?

Districts in Texas and South Carolina adopted Alpha’s software platform, according to a company document the Times reviewed. The platform reaches about 2,600 public school students. Whether those districts will change how they use it is not yet known.

Sources

#AI in Education #Alpha School #EdTech #Charter Schools #Student Outcomes

About this article. Reported and produced by the USA 24 newsroom using the sources cited above. Read our editorial policy. Spotted an error? Tell us and we will correct it.