Coined in 1973 by design theorists Horst Rittel and Melvin Webber, a Wicked Problem is a complex, ambiguous social or cultural issue that is almost impossible to solve because of nebulous, shifting, and increasingly murky success criteria. I previously wrote about using some of Scott Carlson’s and Ned Laff’s ideas from Hacking College to create Wicked Problem Studios that could be taught in high school. That article specifically singled out independent schools as a forum for many of the concepts, primarily because of the higher percentage of students in grades 9-12 who attend college.

I had a conversation with Ned Laff, co-author of Hacking College, who liked my idea of building a Wicked Problem Studio for independent school (and public school) students in high school. He suggested that I consider finding ways to build a localized focus on each Wicked Problem to make it more interesting to students at that age level who may not have developed a regional, national, or global perspective.

I used ChatGPT Pro to develop the graphic design for the Wicked Problem Studio overview portrayed in that article. The driving question was: “How do we prepare people to thrive in the age of artificial intelligence while protecting human creativity, judgement, and opportunity.”   I directed ChatGPT Pro to organize the topics into six units with a timeline of six weeks per unit. This aligned with a 36-week academic year that most schools organize under and could be split into trimesters (two six-week units each) or two semesters (three six-week units each).

The only part of the Wicked Problem Studio included in that article was the graphic design outlining the organization into six units. After receiving feedback from several colleagues and friends including Ned, I opted to see if I could build out the rest of the studio with the assistance of ChatGPT Pro.

I used the six unit descriptions generated in the design and prompted ChatGPT Pro to develop specific assignments for each week of the six-week units. The units and their week-by-week assignments are listed below in sequence. I reviewed the output and chose not to edit any of it. I thought it would be helpful for  readers to view the actual output from my ChatGPT Pro prompts and evaluate it before any edits were made.

For anyone interested in the concept of creating and teaching Wicked Problem Studios, I propose creating an overview similar to the two-page illustration below and sending it to a panel of interdisciplinary teachers for review to ensure it covers all the major components of the problem-solving framework. After such a review, an AI tool like ChatGPT Pro could be used to create assignments for each unit that teachers could review and edit, or teachers could create the assignments the old-fashioned way.

Since I believe that people using AI tools are collectively smarter than the AI tools themselves, I think exercises like this illustrate how teachers can leverage these tools to develop a curriculum in hours rather than weeks.

Assignment Framework Developed

Below is a specific assignment framework for the Artificial Intelligence, Work, and Human Learning Wicked Problem Studio, organized as six six-week units for grades 10–12.

Each unit assumes one major assignment per week, building toward a unit deliverable and then a final capstone. Students should maintain an AI Studio Portfolio throughout the course.

There are three standing expectations for every assignment:

1. Students must keep an AI-use log documenting tools used, prompts entered, outputs received, revisions made, and human decisions added.
2. Students must not enter private personal information, school records, or confidential data into AI tools.
3. Students must distinguish clearly between AI-generated material, human revision, verified evidence, and original judgment.

Wicked Problem Studio Assignments

Artificial Intelligence, Work, and Human Learning

Driving Question: How do we prepare people to thrive in the age of artificial intelligence while protecting human creativity, judgment, and opportunity?

Unit 1: How AI Works

Unit focus: Students build foundational understanding of AI systems, data, prompts, outputs, strengths, and limitations.

Major unit deliverable: AI explainer guide or mini lesson.

Week 1 Assignment: AI Vocabulary Glossary

Students create an illustrated glossary of 15–20 core AI terms. Suggested terms include artificial intelligence, algorithm, machine learning, training data, model, generative AI, large language model, prompt, output, hallucination, bias, automation, chatbot, neural network, and human-in-the-loop.

Student product: A digital or print glossary with student-friendly definitions, one example per term, and an icon or visual metaphor.

Assessment focus: Accuracy, clarity, use of examples, and ability to explain technical ideas in accessible language.

Week 2 Assignment: Data Pattern Memo

Students examine a simple dataset, such as movie preferences, school lunch choices, weather records, or mock hiring data. They identify patterns in the data and explain what a machine might “learn” from those patterns.

Student product: A one-page memo titled “What This Dataset Teaches.”

Assessment focus: Pattern recognition, understanding of training data, recognition of limitations, and explanation of possible bias.

Week 3 Assignment: Prompting Lab

Students ask an AI tool the same question using three different prompts: a vague prompt, a specific prompt, and a role- or audience-based prompt. They compare the outputs.

Student product: A prompt comparison chart with the original prompts, summaries of outputs, and a reflection on what changed.

Assessment focus: Prompt design, comparison, clarity of analysis, and understanding that AI output depends heavily on user instructions.

Week 4 Assignment: Hallucination Hunt

Students generate or examine AI responses on a teacher-approved topic, then fact-check the responses using reliable sources. They annotate errors, unsupported claims, omissions, or misleading language.

Student product: An annotated AI response with at least five comments and a short reliability rating.

Assessment focus: Factchecking, source evaluation, skepticism, and evidence-based critique.

Week 5 Assignment: Tool Comparison Lab

Students compare two AI tools or two AI-assisted workflows on the same task, such as summarizing an article, creating a study guide, generating research questions, or organizing notes.

Student product: A comparison matrix rating accuracy, usefulness, bias risk, transparency, ease of use, and need for human revision.

Assessment focus: Comparative evaluation, responsible tool use, and judgment about appropriate use cases.

Week 6 Assignment: Peer Teach-Back

Students design and deliver a five-minute mini-lesson explaining one AI concept to classmates.

Student product: A short slide deck, handout, visual explainer, or live demonstration.

Assessment focus: Conceptual understanding, communication, accuracy, and ability to teach peers.

Unit 1 culminating assessment:
Students compile their glossary, prompt lab, hallucination hunt, and tool comparison into an AI Explainer Guide for students or families.

Unit 2: AI in School and Work

Unit focus: Students investigate how AI is changing learning, writing, research, productivity, and workplace tasks.

Major unit deliverable: Case-study presentation on AI in one field.

Week 1 Assignment: Writing Comparison

Students complete a short writing task twice: once without AI assistance and once with AI assistance. The topic should be low-stakes and teacher-approved. They compare the two processes and outcomes.

Student product: A reflection titled “What AI Changed and What It Did Not Change.”

Assessment focus: Metacognition, writing process, ethical disclosure, and quality of reflection.

Week 2 Assignment: Workflow Demonstration

Students choose one workflow where AI might assist: brainstorming, coding, editing, data analysis, design, scheduling, or note-taking. They document the workflow step by step.

Student product: A workflow diagram showing where AI helps, where humans remain necessary, and where risks appear.

Assessment focus: Process mapping, practical understanding, and recognition of human oversight.

Week 3 Assignment: Industry Case Study

Students select one field—healthcare, law, education, journalism, finance, engineering, marketing, agriculture, entertainment, government, or nonprofit work—and research how AI is affecting it.

Student product: A short case-study slide deck.

Required slides should include:
field overview; current AI uses; benefits; risks; jobs or tasks affected; skills workers will need; one unresolved question.

Assessment focus: Research quality, field-specific understanding, and balanced analysis.

Week 4 Assignment: Time and Quality Study

Students compare a human-only workflow with an AI-assisted workflow. They track time, quality, ease, frustration, and revision needs.

Student product: A brief report with a chart or table comparing the two approaches.

Assessment focus: Evidence-based comparison, honest reflection, and ability to avoid simplistic “AI is better” or “AI is worse” conclusions.

Week 5 Assignment: Expert Interview

Students interview a teacher, employer, professional, college faculty member, or technology leader about AI use in school or work.

Student product: Interview notes plus a one-page synthesis.

Suggested interview questions:
How is AI changing your work?
What tasks does it help with?
What risks concern you?
What skills will young people need?
What should schools teach about AI?

Assessment focus: Quality of questions, professionalism, synthesis, and connection to course themes.

Week 6 Assignment: Case-Study Presentation

Students present their industry case study to classmates.

Student product: A 5–7-minute presentation with visuals and a short audience Q&A.

Assessment focus: Evidence, organization, clarity, field-specific insight, and balanced discussion of benefits and risks.

Unit 2 culminating assessment:
Students submit a revised AI in One Field Case Study, including research, interview evidence, workflow analysis, and recommendations for students interested in that field.

Unit 3: Skills That Endure

Unit focus: Students identify, practice, and document the human capabilities that remain essential in an AI-enabled economy.

Major unit deliverable: Personal future-skills portfolio and growth plan.

Week 1 Assignment: Judgment Memo

Students analyze a scenario in which AI gives a recommendation, but the human decision is not obvious. Examples might involve hiring, medical triage, school discipline, college advising, financial aid, or news moderation.

Student product: A one-page decision memo answering: What does the AI recommend? What information is missing? What human values matter? What decision would you make and why?

Assessment focus: Ethical reasoning, judgment, clarity, and attention to context.

Week 2 Assignment: Persuasion Task

Students create a short speech, editorial, video script, or public message persuading a specific audience about an AI-related issue.

Possible topics:
Students should disclose AI use.
Schools should teach prompt literacy.
AI should not replace human feedback.
Workers need lifelong learning opportunities.
AI tools should be audited for bias.

Student product: A persuasive communication piece.

Assessment focus: Audience awareness, argument, evidence, tone, and communication skill.

Week 3 Assignment: Creativity Challenge

Students redesign a product, service, classroom routine, school process, or community experience using design thinking. AI may be used for brainstorming, but students must make the final design decisions.

Student product: A prototype sketch, storyboard, concept map, or design brief.

Assessment focus: Creativity, problem definition, iteration, and explanation of human choices.

Week 4 Assignment: Teamwork Simulation

Students complete a collaborative challenge, such as designing an AI-use policy for a classroom, planning a community workshop, or evaluating a fictional AI product.

Student product: Team output plus individual collaboration retrospective.

Reflection questions:
What role did I play?
Where did our team communicate well?
Where did we struggle?
How did we resolve disagreement?
What would I do differently next time?

Assessment focus: Collaboration, leadership, accountability, and reflective honesty.

Week 5 Assignment: Adaptability Plan

Students research how one career or field is changing because of AI and identify what a person in that field would need to keep learning.

Student product: A personal learning plan with three skills to build, three resources to use, and three habits to develop.

Assessment focus: Career awareness, adaptability, realistic planning, and self-direction.

Week 6 Assignment: Future Skills Profile

Students create a personal profile of their current strengths and growth areas in relation to the AI era.

Suggested categories:
critical thinking; communication; creativity; collaboration; ethical judgment; technical fluency; adaptability; initiative.

Student product: A future-skills portfolio and growth plan.

Assessment focus: Self-knowledge, specificity, evidence from prior assignments, and realistic next steps.

Unit 3 culminating assessment:
Students submit a Personal Future-Skills Portfolio connecting their work from the unit to college, career, and civic readiness.

Unit 4: Ethics, Bias, and Trust

Unit focus: Students examine fairness, privacy, misinformation, intellectual honesty, and accountability in AI systems.

Major unit deliverable: School AI ethics guide or policy brief.

Week 1 Assignment: Bias Audit

Students examine a dataset, AI-generated output, recommendation system, or fictional algorithmic decision for possible bias.

Student product: A bias audit worksheet identifying who benefits, who may be harmed, what data may be missing, and what safeguards are needed.

Assessment focus: Fairness, evidence, perspective-taking, and ability to detect hidden assumptions.

Week 2 Assignment: Privacy Case Brief

Students analyze a case involving student data, facial recognition, online tracking, surveillance, health data, or workplace monitoring.

Student product: A two-page case brief.

Required sections:
case summary; stakeholders; privacy risks; benefits claimed; ethical concerns; recommended guardrails.

Assessment focus: Ethical analysis, privacy reasoning, and policy awareness.

Week 3 Assignment: Deepfake and Misinformation Analysis

Students examine examples of synthetic media, manipulated images, fake audio, AI-generated text, or misinformation campaigns.

Student product: An annotated example plus a short guide titled “How to Evaluate Whether This Is Trustworthy.”

Assessment focus: Media literacy, verification strategies, and understanding of trust breakdown.

Week 4 Assignment: Academic Integrity Policy Draft

Students draft guidelines for responsible AI use in schoolwork.

The draft should address:
when AI use is allowed; when it is not allowed; how students should disclose AI use; what counts as meaningful human contribution; how teachers should evaluate work.

Student product: A one-page student-facing AI-use policy.

Assessment focus: Practicality, clarity, fairness, and alignment with learning goals.

Week 5 Assignment: Structured Debate

Students participate in a formal debate on an AI ethics or policy question.

Possible motions:
Schools should allow AI for brainstorming but not drafting.
Employers should disclose when AI screens job applicants.
AI-generated political content should be labeled.
Students should be graded partly on how they use AI responsibly.

Student product: Debate preparation notes and post-debate reflection.

Assessment focus: Evidence, reasoning, listening, rebuttal, and respectful disagreement.

Week 6 Assignment: Responsible-Use Framework

Students synthesize the unit into a responsible-use framework for a school, club, classroom, or community organization.

Student product: A policy brief or ethics guide.

Required sections:
purpose; acceptable uses; unacceptable uses; disclosure expectations; privacy protections; bias safeguards; human oversight; review process.

Assessment focus: Ethical reasoning, policy design, clarity, and feasibility.

Unit 4 culminating assessment:
Students submit a School AI Ethics Guide or Policy Brief that could be reviewed by school leaders.

Unit 5: Solutions in Action

Unit focus: Students design practical, human-centered ways schools or communities can use AI well.

Major unit deliverable: Prototype, workflow design, or solution brief.

Week 1 Assignment: Needs Assessment

Students identify a real school or community need where AI might help, but only if used carefully.

Possible needs:
student study support; college counseling preparation; language learning; tutoring; accessibility; teacher workload; community information; career exploration; library research support.

Student product: A needs-assessment memo.

Required sections:
problem; affected users; current pain points; why AI might help; why AI might not help; risks to avoid.

Assessment focus: Problem definition, user awareness, and realistic thinking.

Week 2 Assignment: Solution Scan

Students research three existing AI-supported tools, workflows, policies, or programs related to their chosen need.

Student product: A solution scan chart.

Required columns:
solution; target users; benefits; limitations; equity concerns; privacy concerns; evidence of usefulness; adaptation ideas.

Assessment focus: Research, comparison, and ability to learn from existing models.

Week 3 Assignment: Prototype Sprint

Students brainstorm and prototype a tool, workflow, policy, training module, or guide.

Examples:
student AI-use checklist; teacher feedback workflow; career exploration chatbot protocol; responsible prompting guide; AI-supported study routine; parent information guide; school club AI policy.

Student product: Low-fidelity prototype, workflow map, storyboard, or draft guide.

Assessment focus: Design thinking, creativity, connection to user needs, and feasibility.

Week 4 Assignment: User Feedback Test

Students test their idea with classmates, teachers, staff, parents, or community members.

Student product: Feedback summary with at least five user comments or observations.

Required reflection:
What did users understand?
What confused them?
What did they value?
What risks did they notice?
What should change?

Assessment focus: Feedback collection, humility, listening, and revision planning.

Week 5 Assignment: Revision Memo

Students revise their solution based on feedback and evaluate it through ethical, practical, and equity lenses.

Student product: Revision memo.

Required sections:
what changed; why it changed; privacy safeguards; bias safeguards; accessibility considerations; implementation challenges.

Assessment focus: Iteration, ethical design, feasibility, and inclusion.

Week 6 Assignment: Solution Brief

Students produce a final solution brief explaining their proposed AI-supported intervention.

Student product: Three- to five-page solution brief or equivalent presentation document.

Required sections:
problem; users; proposed solution; implementation plan; risks; safeguards; evidence; next steps.

Assessment focus: Coherence, practicality, ethical safeguards, and communication.

Unit 5 culminating assessment:
Students submit a Prototype and Solution Brief suitable for presentation to a school or community audience.

Unit 6: Capstone Design Challenge

Unit focus: Students synthesize the studio by proposing an AI-readiness plan, tool, or policy for a real audience.

Major unit deliverable: AI-readiness action plan and final presentation.

Week 1 Assignment: Challenge Proposal

Student teams select a final challenge connected to AI, work, learning, ethics, or community readiness.

Possible capstone topics:
school AI-use policy; AI and college readiness; AI and academic integrity; AI and career exploration; AI and student mental workload; AI and media literacy; AI and future skills; AI and teacher support; AI and equitable access.

Student product: One-page challenge proposal.

Required sections:
driving question; target audience; problem statement; why it matters; proposed product; early research questions.

Assessment focus: Focus, relevance, audience, and feasibility.

Week 2 Assignment: Research Matrix

Teams research models, examples, policies, expert views, and constraints related to their challenge.

Student product: Research matrix with at least eight sources or examples.

Suggested columns:
source or model; key idea; evidence; relevance; limitation; how it informs our project.

Assessment focus: Research depth, source quality, synthesis, and connection to project design.

Week 3 Assignment: Draft Action Plan

Teams create a first full draft of their capstone product.

Possible products:
policy proposal; training module; student toolkit; parent guide; teacher workflow; career-readiness resource; community workshop plan; human-AI collaboration guide.

Student product: Draft action plan or prototype.

Required sections:
audience; goals; proposed actions; timeline; needed resources; risks; safeguards; success measures.

Assessment focus: Structure, feasibility, audience alignment, and evidence base.

Week 4 Assignment: Feedback Round

Teams present their draft to peers and at least one outside reviewer, such as a teacher, administrator, counselor, librarian, employer, alumnus, or technology professional.

Student product: Feedback synthesis.

Required sections:
who reviewed it; what feedback they gave; what patterns emerged; what the team will revise; what the team will not revise and why.

Assessment focus: Responsiveness, professionalism, stakeholder engagement, and revision judgment.

Week 5 Assignment: Final Revision and Presentation Design

Teams revise their written proposal and prepare a final presentation.

Student product: Final written proposal plus presentation deck, poster, demonstration, workshop plan, or policy document.

Required elements:
clear problem statement; evidence; proposed solution; implementation steps; ethical safeguards; cost or resource considerations; success indicators.

Assessment focus: Quality of revision, clarity, persuasiveness, design, and readiness for public presentation.

Week 6 Assignment: Public Presentation and Reflection

Teams deliver their capstone to an authentic audience.

Possible audiences:
school leadership team; faculty committee; technology committee; parent association; student government; local employers; alumni panel; board committee; community organization.

Student product: Public presentation plus individual reflection.

Individual reflection questions:
What did I contribute?
What did I learn about AI?
What did I learn about human judgment?
What would I revise if this project continued?
How has my thinking about college, work, and learning changed?

Assessment focus: Presentation quality, evidence, feasibility, ethical reasoning, teamwork, and personal reflection.

Unit 6 culminating assessment:
Students submit and present an AI-Readiness Action Plan for a real school, career, or community context.

Suggested Portfolio Structure

By the end of the studio, each student’s AI Studio Portfolio could include:

1. AI vocabulary glossary
2. Prompting lab and hallucination hunt
3. Tool comparison matrix
4. AI in one field case study
5. Expert interview synthesis
6. Future-skills profile
7. Bias audit
8. Privacy or misinformation case brief
9. Responsible-use framework
10. Solution brief or prototype
11. Capstone proposal and presentation
12. Final personal reflection

Simple Assessment Categories Across the Studio

A common rubric could evaluate students on:

Conceptual understanding: Does the student understand how AI works and where its limits are?

Evidence and verification: Does the student check claims, use reliable sources, and avoid unsupported conclusions?

Ethical reasoning: Does the student consider fairness, privacy, bias, transparency, and human responsibility?

Human skill development: Does the student demonstrate judgment, communication, creativity, collaboration, and adaptability?

Practical design: Are the proposed solutions realistic, useful, and appropriate for the intended audience?

Reflection: Can the student explain how their thinking changed and what they still need to learn?

The overall purpose is to help students leave the studio not merely knowing about AI, but knowing how to work with it, question it, govern it, and preserve the human capabilities that matter most.

Final Thoughts

Since ChatGPT was first made available to the masses in November 2022, I have utilized output from ChatGPT and other AI tools in numerous articles in my blog. I have provided transparency in the form of disclosure regarding the AI tool and the extent to which I used it for my published content.

It was easy to identify AI-generated content in the earlier versions of LLMs. It’s getting harder as the LLMs gain sophistication and minimize their hallucinations. At some point, many professionals will be expected to use these tools as part of their routine work.

It took me less than five minutes to generate the graphics for this course and approximately the same time for the list of assignments in each of the six six-week units. Are the assignments perfect? No. What would I do to make them perfect? I’d sit down with a group of colleagues (assuming I’m a teacher at a high school committed to this Wicked Problem Studio idea) and discuss each of the units and the specific assignments, tweaking them along the way.

ChatGPT and other AI tools are designed for rapid iteration. They don’t get tired or frustrated (yet). I am sure that people are using these tools to design courses, both in person and online. Some may choose to disclose their use of the AI tools, and some may not.

This is another example of a situation where knowledge is transitioning from a scarce resource to a commodity. The only difference may be in the effectiveness of the teacher/instructor who teaches it. There are many downstream repercussions for courses that can be constructed, designed, and implemented by people who may not have the prerequisite knowledge of the subject matter experts who teach those courses now. One of the biggest downstream impacts may be that educational institutions will only be able to charge tuition based on the value of their community and not the content of their courses. And that idea could be the basis for a future article.

The post The Wicked Problem Studio: Preparing for a Future of Learning and Working Influenced by Artificial Intelligence appeared first on Wally Boston.

The findings summarized by Guidewell Education point to a clear shift among highly selective institutions. The article highlights three lessons:

1. Test scores have remained powerful predictors of college performance.
2. Many students misunderstand what “optional” really means.
3. Some colleges gain institutional advantages from keeping tests optional, including larger applicant pools and higher reported score ranges (since students with lower scores will opt not to submit them).

Yale’s internal review, as reported by Guidewell, found that students who did not submit scores had a much lower admit rate than students who did, raising a transparency concern. “Optional” may have sounded neutral, but in practice, the admissions office saw scores as meaningful evidence.

That is why the next phase of test-optional admissions will not be a broad, uniform return to the pre-pandemic norm. Instead, it will be a split market. More highly selective colleges, STEM-heavy institutions, and competitive public flagships are likely to return to required or “test-flexible” policies. However, most U.S. colleges will likely remain test-optional or test-free because their enrollment incentives, applicant pools, and access goals point in the opposite direction.

The Data Decides

The strongest argument for reinstating the SAT or ACT is not an insistence from seasoned faculty members. It is data. Dartmouth, one of the most visible schools to reverse course, said its research showed that high school grades combined with standardized testing (SAT by itself explains 22% of the variance) were the most reliable indicators of success in Dartmouth’s curriculum (combined variables explain 25% of the variance), and that scores could help identify high-achieving students from low- and middle-income backgrounds, first-generation students, and students from less familiar schools. Dartmouth then reactivated its testing requirement beginning with applicants to the Class of 2029.

UT Austin reached a similar conclusion from a public-university perspective. After four years of test-optional admissions, the university announced that it would again require scores beginning with Fall 2025 applications. Its internal analysis found that students who opted to have scores considered had a median SAT of 1420, compared with 1160 among those who did not, and that students who opted in had an estimated first-semester GPA 0.86 points higher, controlling for factors such as high school GPA and class rank. UT also found that those students were 55% less likely to have a first-semester GPA below 2.0.

Those numbers matter because selective admissions offices are not only choosing students; they are trying to predict who will thrive in demanding programs. At institutions where many applicants have near-perfect GPAs, rigorous course loads, polished essays, and strong recommendations, standardized tests provide one more comparative data point. That is especially useful where grade inflation has weakened the ability of GPA alone to distinguish academic readiness. UT’s explanation was blunt: With so many high school GPAs clustered around 4.0, the SAT or ACT can serve as a differentiator for admission, major placement, and early academic support.

Transparency is Key

The second argument for requiring scores is transparency. MIT made this point early, reinstating its SAT/ACT requirement in 2022. MIT argued that tests helped assess academic preparedness and identify socioeconomically disadvantaged students who lacked access to advanced coursework or other enrichment opportunities. MIT also stated a requirement was “more equitable and transparent” than a test-optional policy because it reduced uncertainty about whether students should submit scores.

This transparency issue is central. In a test-optional system, students with strong counseling often know when to submit and when to withhold. Students without that advice may under-submit, mistakenly assuming that “optional” means “irrelevant.” Harvard made a similar point when it reinstated required testing for the Class of 2029, saying that students who withhold scores may be leaving out information that could help their applications when interpreted in context. Harvard also cited research suggesting that other application components, such as essays, extracurriculars, and recommendations, can be at least as shaped by family resources as test scores.

That does not mean test scores are perfect. They are not. Access to test prep, repeat testing, transportation, fee waivers, and informed advising still varies widely. But the emerging view among many selective colleges is that scores are flawed but useful, especially when interpreted in context. Brown’s review reached that conclusion. The university reinstated testing for the Class of 2029 and said academic outcomes were strongly correlated with test scores across subgroups, including students from less-advantaged high schools and historically underrepresented groups.

What Increasing Policy Reversals Mean

The policy reversals have continued beyond the first wave. Caltech reinstated its standardized-test requirement for applicants beginning in fall 2024, saying standardized tests provide admissions officers and faculty with useful information about academic preparedness. It also noted that more than 95% of its most recently enrolled class had taken standardized exams, even while scores were not visible to admissions. Penn reinstated testing for students applying for fall 2026 admission, saying the shift would bring “clarity and transparency” to the process. Cornell will require SAT or ACT scores for first-year applicants beginning with Fall 2026, after a multiyear study found that scores, reviewed in context, helped complete the picture of an applicant.

Princeton’s decision is especially telling because it gives the market a longer runway. Princeton remains test-optional for applicants seeking to enroll in fall 2026 or fall 2027, but it will return to required SAT or ACT scores for the 2027-28 admission cycle, after a review of five years of test-optional data found stronger academic performance among students who submitted scores. Stanford and Johns Hopkins have also moved back toward required testing for fall 2026 admissions, adding to the sense that the most selective tier is normalizing a return to scores.

Does that mean test-optional admissions is collapsing? No. Nationally, the opposite is still true. FairTest reports more than 2,085 accredited bachelor’s degree-granting institutions with ACT/SAT-optional or test-free policies for students seeking to enroll in Fall 2026 or beyond, including more than 2,010 test-optional institutions and more than 75 test-free institutions. FairTest also says more than 90% of colleges have retained or extended test-optional or test-free policies.

The key distinction is that the most visible colleges do not represent the entire market. Highly rejective universities can require scores and still fill their classes many times over. Less selective private colleges, regional publics, and institutions facing demographic pressure may have a stronger incentive to keep barriers low. Test-optional policies can increase applications, reassure anxious students, and help colleges compete for applicants in a shrinking or uneven high school population.

Guidewell’s article makes this point well. Test-optional admissions can deliver institutional benefits such as higher reported average scores because lower scores are withheld, larger applicant pools, lower admit rates, and a competitive advantage over colleges that require testing. Those incentives are difficult for many colleges to give up, even if internal data shows that scores have predictive value.

That creates the most likely forecast that more reversals are coming, but they will be concentrated. Expect additional highly selective private universities, competitive engineering schools, and flagship publics with capacity constraints to adopt required, test-flexible, or test-preferred policies. These colleges have the most to gain from additional academic information and the least to lose from smaller applicant pools. They can also frame the change as an equity move, arguing that required scores prevent under-submission by talented low-income students.

At the same time, most colleges will not abandon test-optional admissions. For many institutions, the SAT and ACT are not central to enrollment management. Their challenge is not choosing among tens of thousands of academically indistinguishable applicants. Their challenge is attracting, admitting, and enrolling enough qualified students. For them, requiring tests may create friction without enough predictive payoff to justify the loss of applicants.

The End of Test-Optional?

The better phrase for the future is not “the end of test-optional.” It is “the end of one-size-fits-all testing policy.” Students applying to the most selective colleges should increasingly assume that testing matters, even where scores remain technically optional. Students applying to a broader range of colleges will still find many places where scores are optional, irrelevant, or not considered at all.

In short, more colleges will revert to test requirements, but not most colleges. The return of testing will be real, influential, and heavily concentrated at the top of the selectivity ladder. Test-optional admissions will remain the national majority, but at the most competitive institutions, the word “optional” is losing its power.

The post Test-Optional Colleges Are Not Going Away—But the Admissions Market Is Splitting appeared first on Wally Boston.

But the world today’s high school graduates enter has changed. College itself is changing. Some students may arrive on campus with AP credits, dual enrollment credits, online learning experiences, internships, research projects, professional and industry certifications, entrepreneurial ventures, and clearer expectations about the value of higher education. At the same time, other students still struggle to understand how college works once they get there. They may know how to get admitted, but not how to use the institution well.

This is a discrepancy that needs attention.

A student can enter college with 18 credits (AP and dual enrollment) and still not know how those credits apply. A student can have a declared major and still not understand how to build relationships with faculty, pursue internships, explore careers, or connect academic choices to long-term goals. A student can attend a prestigious college and still move through the curriculum passively, choosing courses because they fit a schedule rather than contributing to a larger intellectual and professional direction.

This is where Scott Carlson and Ned Laff’s Hacking College: Why the Major Doesn’t Matter offers a useful framework for independent schools. Although the book is written for college students, many of its most important ideas belong in grades 9–12. If independent schools want to prepare students not just for admission to college, but for genuine college and career success, they should begin teaching students how to design their education before they leave high school.

We don’t need to turn high school into college, but we can help students arrive at college as active, informed, curious designers of their own learning.

Beyond “What Do You Want to Major In?”

High school students are often asked premature questions: What do you want to major in? What career do you want? What college do you want to attend?

These questions are not useless, but they can narrow student thinking too early. Many students do not yet know enough about themselves, the world of work, or the structure of higher education to answer them well. Worse, they may come to believe that choosing the “right” major is the key to future success.

One of the most useful ideas in Hacking College is that students should focus less on the major itself and more on building a broader “Field of Study.” A Field of Study is not simply a department label. It is a student’s evolving map of interests, questions, skills, experiences, relationships, and opportunities.

For a high school student, this shift is powerful. Instead of asking, “Do I want to major in biology, business, engineering, or political science?” the student might ask, “What kinds of problems do I want to understand?” or “What subjects do I return to even when no one assigns them?” or “What kinds of people are doing work that fascinates me?”

A student interested in climate change, for example, may discover connections among environmental science, economics, public policy, engineering, architecture, data science, ethics, and community organizing. A student interested in health may discover connections among biology, psychology, public health, technology, inequality, communication, and design. A student interested in storytelling may discover pathways through literature, film, journalism, marketing, history, digital media, and entrepreneurship.

It’s time for majors to evolve. Instead of putting hard lines around learning, students should be encouraged to view majors as one part of a larger educational design.

Teaching Students to Use the “Blank Spaces”

Another valuable concept from Hacking College is the idea of “blank spaces.” In college, these are the electives, general education courses, minors, internships, independent studies, and co-curricular opportunities that students often fill casually. They choose a course because it is convenient, because a friend recommended it, because it seems easy, or because it checks a requirement.

High school has blank spaces, too.

High school blank spaces include electives, summer programs, service projects, internships, clubs, independent studies, capstones, arts commitments, athletic leadership, online courses, dual enrollment options, and research opportunities. In many schools, students choose these experiences without much intentionality. They may accumulate impressive activities without developing a coherent sense of what they are learning or why it matters.

Independent schools can change that.

Beginning in grade 9, students could maintain a simple “Learning Map” that tracks courses, major projects, extracurricular commitments, summer experiences, service, interviews, and emerging interests. The purpose would be to help students notice patterns.

• Where have I done my best thinking?
• Which assignments stayed with me?
• What kinds of problems do I like solving?
• What skills am I building?
• What do I want to test next?

By junior year, students should be able to explain not only what they have done, but what they are learning about themselves through what they have done.

I created an example of a Learning Map using ChatGPT Pro and appended it below. I can envision how completing this with actual courses and projects would comprise a portfolio of learning, academically and experientially.

The Research Investigative Inquiry: A High School Essential

Perhaps the most practical idea independent schools can borrow from Hacking College is the Research Investigative Inquiry. In the college context, students are encouraged to research fields of interest, identify people doing meaningful work, conduct conversations, and use those insights to shape their educational choices.

This belongs in high school.

Imagine if every student completed a Research Investigative Inquiry by the end of junior year. Each student would identify a broad question, field, profession, or social problem of interest. The student would then research people and organizations connected to that area, conduct several informational interviews, synthesize what they learned, and identify the academic disciplines and experiences that could help them explore further.

This would teach several lessons at once.

Students would learn how to ask better questions. They would learn how adults’ careers develop, which is often in nonlinear and unexpected ways. They would learn that professional relationships are built through curiosity and generosity, not just networking for personal advantage. They would learn that academic subjects connect to real people doing real work in the world.

Just as important, students would begin college with more confidence. A student who has already interviewed a civil engineer, a city planner, and a climate policy analyst will ask better questions when choosing courses. A student who has spoken with a physician assistant, a hospital administrator, and a public health researcher will understand that “health careers” include many paths. A student who has interviewed artists, designers, writers, and entrepreneurs will see creative work as both craft and profession.

This kind of inquiry helps students develop social capital before they need it urgently.

Wicked Problems and the Case for Interdisciplinary Learning

Independent schools are well-positioned to help students understand that the most important problems do not fit neatly into academic departments. Climate change, artificial intelligence, migration, public health, inequality, food systems, mental health, misinformation, and democratic participation are all “wicked problems.” They are complex, messy, interdisciplinary, and resistant to simple solutions.

A high school curriculum that prepares students for college and careers should give them practice working across disciplines. This does not mean abandoning traditional subjects. Students still need strong foundations in mathematics, science, history, literature, language, and the arts. But they also need opportunities to bring those disciplines together.

A school could create “Wicked Problem Studios” in grades 10–12. These might be semester-long or year-long interdisciplinary experiences focused on topics such as water, housing, AI and ethics, public health, climate resilience, or media and democracy. Students would read, research, analyze data, write, debate, design, interview, and present. They would learn that disciplines are not isolated boxes. They are tools for understanding complicated realities.

This approach also strengthens the liberal arts. It shows students that communication, critical thinking, ethical reasoning, quantitative analysis, creativity, and collaboration are not abstract virtues. They are survival skills for complex problems.

Using ChatGPT Pro, I created several graphic overviews of a Wicked Problem Studio. For use in this article, I opted to use the question of how we can prepare people to thrive in the era of artificial intelligence. The overview is illustrated below.

I thought the six-unit Studio Arc created by ChatGPT Pro was reasonable. I believe it’s reasonable to assume each unit lasts six weeks. If I were developing the studio, I could use this outline to create the units, then develop specific assignments for each week in each unit. I like the idea of the Capstone Design Challenge for the sixth and final unit. I may consider developing the six-unit, 36-week Studio for a future post.

Dual Enrollment With Purpose

As more students earn college credit during high school, independent schools need to be thoughtful about how those credits fit into the larger student experience. Dual enrollment can be valuable, but only when it is connected to a purpose.

The question should not be, “How many college credits can a student earn before graduation?” The better question is, “How does this course help the student build readiness, explore a field, develop a skill, or create future optionality?”

Some students may use early credits to accelerate their studies. Others may use them to make room in college for study abroad, research, internships, a double major, or a more demanding academic sequence. Still others may use a college-level course to test whether a field truly interests them.

Schools should also teach credit literacy. Students and families need to understand the difference between earning credit, receiving placement, and having a course apply to a major or graduation requirement. A credit that appears impressive on a transcript may not help much if it does not transfer or apply in a meaningful way.

Before enrolling in a dual enrollment course that is not part of the basic general education pathway, students should be able to explain how the course fits their emerging Field of Study, academic readiness, and future plans.

The non-profit Education Design Lab (EDL) (full disclosure: I am a Board member) initiated the Career Ready by Design project to strengthen dual enrollment pathways in the state of Georgia. Their design question for that project is: “How might we design clear, transparent dual-enrollment and learn-work pathways that make expectations explicit, align coursework with meaningful credentials, and eliminate the hidden complexities that leave learners feeling lost?”

The EDL project is currently working with five community colleges and 15 public high schools. The project’s goal of having high school students earn before graduation “at least 12 transferable credits with explicit connections to credentials, degrees, and career pathways aligned with regional workforce needs” is very much aligned with the purpose stated above. Paying attention to the outcomes from this work could be insightful for private and public schools alike.

College Navigation Is a Skill

One of the most overlooked forms of college readiness is knowing how college works. Many students arrive on campus without understanding degree maps, prerequisites, office hours, advising, add/drop deadlines, transfer credit, general education requirements, or the difference between a major, minor, certificate, and concentration.

Independent schools can teach this directly.

By senior year, students should know how to read a college catalog, email a professor, prepare for an advising appointment, ask whether a course fulfills a requirement, and think strategically about their first year. They should understand that college is not only a set of courses. It is an ecosystem of faculty, peers, advisers, offices, research opportunities, internships, alums, and support systems.

Students who know how to navigate that ecosystem will be better positioned to thrive.

A New Standard for College Preparation

The independent school promise should extend beyond college admission. It should include preparation for what happens after the acceptance letter.

A strong grades 9–12 program could include a Learning Map, Field of Study advising, Research Investigative Inquiry, Wicked Problem Studios, credit-literacy education, experiential learning, and a senior presentation in which students explain what they have learned about themselves and how they intend to use college well.

The final question for graduates should not only be, “Where are you going to college?” It should also be, “How are you prepared to make the most of it?”

That is the deeper opportunity. Independent schools can help students graduate not as passive consumers of a curriculum, but as thoughtful orienteers; specifically, young people who know how to explore, connect, question, adapt, and build purposeful lives.

The post Helping High School Students “Hack” College Before They Get There appeared first on Wally Boston.

Based on 9,516 responses from students who were prospects for Entering Class 2025 (students who entered college in the Fall of 2025), the survey shows that college choice is not a simple ranking exercise. It reveals a longer, more fluid process in which students discover institutions through new tools, compare those schools across practical measures of value, and keep their options open later into senior year.

Why Do Students Choose a College?

For colleges, the most important lesson is that “why students choose” has become less about a single campus feature and more about confidence. Students want confidence that a degree will lead somewhere, that the cost will be manageable, that the campus will feel right, and that the institution can answer highly personal questions quickly. The winning college is increasingly the one that reduces uncertainty better than its competitors.

The clearest finding is that career value is central. When students were asked what characteristics best represent the value of higher education, successful job placement after graduation ranked first at 44%. Internships, co-ops, and other active-learning experiences followed at 35%, tied with availability of scholarships. Generous financial aid awards came next at 30%, moderate tuition at 29%, strong career services at 24%, and a comprehensive set of academic programs at 23%. By comparison, “well-known school” came in at 19%. This matters because it suggests that institutional prestige still has value, but it is not the dominant proof point. Students are asking: Will this college help me get a job, gain experience, and justify the investment?

That emphasis is especially important because the report frames college choice against growing uncertainty about artificial intelligence. Students are not simply wondering whether college is “worth it” in the abstract. They are wondering whether the careers they are preparing for will exist in a recognizable form. In the survey, 42% of students said AI would influence the job or career they pursue, while another 23% were unsure. Students were also divided over whether AI would change the number of jobs requiring a college degree: 32% expected fewer jobs to require one, 25% expected more, and 29% were unsure. In other words, uncertainty itself has become part of the decision-making environment.

That uncertainty is already reshaping academic choices. About 10% of students reported changing their intended major due to AI. Among those reconsidering majors, concerns about job security were a major driver, especially in fields such as technology (39%), creative work (23%), business (14%), and engineering-related (9%) roles.

For enrollment leaders, the takeaway is not that students are abandoning career-oriented majors. It is that they need clearer evidence that a college is preparing them for the labor market they expect to enter. Messages about “career readiness” must become more concrete: examples of internships, employer partnerships, AI-integrated coursework, alumni pathways, entry-level job outcomes, and transferable skills.

Affordability is the second major reason students choose one institution over another. The report shows that financial concerns are not limited to tuition. Among students who opted out of college, “I couldn’t afford college” was selected by 25%, while 12% said they needed to work and college was not an option.

EAB also notes that practical financial concerns have grown more prominent, including a rise in students citing cost of living as a primary financial concern from 51% in 2025 to 67% in 2026. For enrolled students, cost still appears throughout the value equation as scholarships, financial aid, moderate tuition, total budget, and expected debt all factor into family decision-making.

Parents and guardians remain deeply involved in this part of the choice process. The report finds that parent influence is strongest in areas of total college budget and anticipated debt, more than in the number of applications submitted or even the major a student studies. This reinforces the fact that colleges are not only recruiting students. They are recruiting family decision systems. A student may love the campus, but a parent may be focused on debt, net price, and whether the degree will lead to a stable job. Colleges that communicate affordability clearly, early, and repeatedly will be better positioned than those that make families decode price late in the process.

The third reason students choose specific colleges is fit, but fit now has a more digital and comparative character. Students still appreciate communications from the college that assist their search process. Asked how they would like the school to share information at the start of their search, 68% preferred emails, 61% preferred college websites, 52% chose mail/personal letters, 49% chose college visits, and 29% chose text messages.

The report highlights virtual tours as an important tool for evaluating campus experience. Forty percent of students took a virtual tour during their college search, and among those who did, 40% viewed the same school’s tour more than once. Virtual tours served multiple purposes: 18% used them to replace an in-person visit, 30% used them to compare schools, 39% used them to prepare for an in-person visit, and 24% used them to show a school to family members. EAB connects this activity to key enrollment-decision factors such as campus environment and location.

This is an important nuance. Students may describe their final choice emotionally: “I could see myself there.” But the path to that feeling is increasingly mediated by digital tools. A virtual tour, a campus video, a residence hall walkthrough, a student Q&A, or an interactive map may help students test whether a place feels real to them. The effect is especially important for students who cannot easily travel. First-generation students were more likely than their peers to take virtual tours, and use was higher among Black, Asian, and Hispanic students than among White students. Students from households earning less than $60,000 were also among the most likely to use these tools.

AI’s Role in the College Choice Process

AI is becoming another tool for discovery and comparison. The report finds that 67% of students use AI tools at least a few times per week. Within college search specifically, the share using AI tools rose from 14% to 26% between the 2025 and 2026 surveys. The share naming AI as a top one or two helpful resource also increased, from 3% to 5%. Those numbers may still look modest, but the pace of change is the signal. Students are learning to ask tools for customized comparisons: Which schools are good for my major? Which colleges offer strong internships? Which campus is affordable for someone like me? Which school is close to home but still has a strong program?

The equity implications are notable. EAB reports that students of color are more likely than White students to use AI in college search, find it helpful, and discover a previously unknown school through AI. That means AI may expand the set of colleges some students consider, especially when they lack access to traditional counseling or travel resources. But it also creates a new visibility challenge for institutions: a college that is not accurately represented in AI-generated answers may lose a student before direct communication ever begins.

Finally, students are choosing later and comparing more. The report shows that students now submit an average of 7.1 applications, up from 6.1 five years earlier. They receive an average of 5.2 acceptances, while deposits have held at 1.4 for four consecutive years (students with family income greater than $120,000 have a deposit ratio of 1.5). This means applications and admits are weaker signals of intent than they used to be. More students are building portfolios of options, comparing offers, and delaying commitment while they wait for clarity on price, program fit, and personal confidence.

Demographic patterns add another layer. Students of color are applying to more schools than White students, with Black and Asian students submitting the highest average number of applications. Higher-income students are more likely to submit multiple deposits, suggesting that financial flexibility can extend the decision window. For colleges, the implication is straightforward: the work of persuasion does not end at admission. In many cases, it begins again there.

How Should Institutions Respond?

The report ultimately points to a more demanding enrollment environment. Students choose colleges that make value visible, cost understandable, fit tangible, and answers immediate. They are less willing to accept broad promises and more likely to compare institutions across evidence that feels personal to them. A beautiful campus, a strong brand, or a generous scholarship may still matter, but each is part of a larger question: Can I see myself succeeding here, academically, socially, and financially, in a world that feels uncertain?

For enrollment teams, the response should be equally clear. Lead with proof of career preparation (this may be easier said than done for campuses that have not tracked the career outcomes of their alums). Explain cost in family-friendly terms. Invest in digital experiences that help students imagine campus life. Make institutional information easy for both students and AI tools to find and interpret. Stay engaged after admission, because students are still comparing. The colleges that win will not be the ones that merely say they are a good fit. They will be the ones that help students and families believe it.

The post The College Choice Process Today: Value, Fit, and Confidence in an Uncertain Market appeared first on Wally Boston.

NBOA’s Financial State of the Industry: BIIS 5-Year Trend Report 2021–2025 (note: the report is free to NBOA members and $125 for non-members) tracks 18 financial and operating indicators for 279 day schools and boarding schools that submitted financial operations data every year from 2020-21 through 2024-25.

By following the same schools over time, the report provides a useful longitudinal view of what changed during (2020-2021) and after the pandemic, rather than a one-year snapshot of different schools in each period. The pool of schools was not statistically selected, but NBOA notes that the schools included are distributed geographically across the United States.

The headline is not a simple story of recovery or distress. It is both. Enrollment and demand signals are generally positive. Balance sheets have improved. Net tuition revenue has grown. Capital investment is beginning to recover. However, expenses are rising faster than tuition revenue, the per-student “gap” that must be covered by other revenue sources is widening, philanthropy is becoming more uneven, and deferred maintenance has increased. Independent schools appear to have gained financial runway, but they are also facing a sharper test of strategy amid demographic and economic shifts that are exerting pressure on their financial models.

Enrollment Patterns

Enrollment tells the first part of the story. Average enrollment among the 279 schools increased from 571 students in 2020-21 to 604 in 2024-25, a 5.7% gain. Median enrollment rose from 475 to 495 students, a 4.2% increase. Most schools grew: 76% reported higher enrollment over the four-year span, while 23% declined.

However, the pace of enrollment growth slowed materially, from a 2.8% increase early in the period to just 0.7% from 2023-24 to 2024-25. Larger schools fared better: those at the 90th percentile increased enrollment 8.1%, compared with just 1.6% growth at the 10th percentile. Regionally, the report notes that 44% of schools in New Jersey and New York and 35% in the Midwest experienced enrollment declines.

The broader demographic environment is not uniformly favorable. NCES data show private K-12 enrollment was 4.7 million in fall 2021, unchanged from fall 2019, with private schools representing 9% of combined public and private enrollment in both years. However, NCES’s broader projections for elementary and secondary enrollment expect both public and private enrollments to be lower in 2030 than in 2019, with private elementary and secondary enrollment projected to decline 12% between 2019 and 2030. The takeaway for independent schools is that demand exists, but it is likely to be more competitive, more regional, and more sensitive to value.

The second major story is balance sheet strength. Average total net assets per student increased from $86,979 to $104,630, a 20.3% gain. Median total net assets per student rose even faster, from $66,400 to $84,410, a 27.1% increase. Fully 80% of schools increased total net assets per student over the period. Expendable financial resources per student also improved, rising from $33,741 to $40,105 on average, an 18.9% increase; the median rose 26.4% from $25,302 to $31,988. However, that strength was not universal, as 37% of schools saw per-student expendable financial resources decline.

This distinction between average total net assets per student and expendable financial resources per student is important. Total net assets are a sign of long-term financial capacity, but they may be tied up in buildings, endowment restrictions, or other illiquid resources. Expendable financial resources are closer to the funds a school can use to handle shocks, seed initiatives, or manage downturns. The report suggests that many schools are better capitalized than they were in 2020-21, but not all are more flexible.

On pricing, the data are striking. Average gross tuition and fees per student increased from $31,563 to $37,517, an 18.9% gain. Median gross tuition and fees rose 18.7%. Net tuition and fees per student, after financial aid and tuition remission, rose even faster: average net tuition increased from $25,636 to $30,817, a 20.2% gain, while the median rose 20.9%. Over four years, net tuition and fees per student increased at 97% of schools.

That is a positive sign for the tuition-dependent independent school model. Schools were not simply raising sticker prices while discounting away the gains. Financial aid per student increased 14.9%, and 71% of schools increased financial aid per student, but financial aid as a percentage of gross tuition and fees declined slightly from 14.2% to 13.3%. The total discount rate, including tuition remission, also declined from 16.5% to 15.7%.

The risk is affordability. NAIS’s 2024 research on how parents pay school costs found that among parents who applied for financial aid, only 10% said they would still be very or extremely likely to send their children to private school without that support. NAIS also reported that 55% of parents felt stressed about paying for private school, up from 47% in 2018. NBOA’s findings show pricing power, but schools should not confuse pricing power with unlimited household capacity.

Expense Growth

The most important warning sign in the report is the rate of expense growth. Average total operating expenses per student increased from $30,486 to $38,798, a 27.3% increase, with an average annual increase of 6.8%. The median rose 28.6%. Nearly every school in the study (95%) saw operating expenses per student increase.

Because net tuition per student rose 20.2% while expenses rose 27.3%, the gap between what schools charge in net tuition and what they spend per student widened significantly. The average gap per student increased from $4,833 to $7,469, up $2,636, and 80% of schools saw the gap increase.

This is the core structural challenge. Tuition increases have been strong, but expense increases have outpaced those gains. One analysis that might be useful is whether or not the operating expense increases were due to overall inflation or increases in faculty and staff FTE’s. The operating margin data reinforce the point, though with an important caveat. Average operating margin declined from 8.7% in 2020-21 to 5.7% in 2024-25, while the median declined from 8.7% to 4.9%. NBOA notes that margins were unusually high in 2020-21 and 2021-22 because some schools received federal pandemic assistance, so the decline partly represents a return to normal. Still, “normal” now includes higher recurring costs.

Admissions and Philanthropy Differ in Some Ways

Admissions and philanthropy offer mixed but useful signals. Application coverage, applications divided by new students, increased from 251% to 263% on average, and the median rose from 242% to 262%; 60% of schools improved this metric over four years, suggesting that many schools have deeper applicant pools than they did at the start of the period. At the same time, a better indicator may be calculating the applications coverage percentage for full pay students and for financial aid students. A much lower coverage ratio for full pay students could be indicative of a tighter market.

Fundraising, however, is becoming a differentiator. Average annual giving per student increased 11.2%, from $1,475 to $1,641, and the median rose 13.1%. But the bottom and top of the distribution moved in opposite directions. At the 10th percentile, annual giving per student fell from $485 to $312. At the 90th percentile, it rose 20.4%, from $4,037 to $4,862. Total contributions per student were relatively steady, peaking at $2,170 in 2021-22, while the median rose from $1,780 to $1,973.

That divergence aligns with broader philanthropy trends. Giving USA has reported declining donor retention, especially among small-dollar donors, while Blackbaud’s 2025 giving trends found that mid-level and major donors carried sector growth, with gifts of $1,000 or more increasing while sub-$1,000 giving declined. For independent schools, that means advancement capacity may increasingly determine which schools can close the operating gap without relying too heavily on tuition increases.

Costs of Operating and Maintaining Facilities

Facilities are the final piece of the post-pandemic picture. Average property, plant, and equipment, net of depreciation, rose from $42,104 to $49,556 per student, a 17.7% increase. Median PP&E per student increased 16.6%. Yet only 51% of schools increased PP&E per student, indicating uneven reinvestment.

More telling is the capital spending ratio, which compares capital expenditures with depreciation (capital expenditures divided by depreciation expense). NBOA reports an average ratio of 0.79 over five years; a ratio below 1.00 suggests facilities are being consumed faster than they are being renewed. The ratio improved from 0.52 in 2020-21 to 1.11 in 2024-25, but it was below 1.00 in four of the five years.

Debt does not appear to be the main driver of the facilities rebound. Average plant debt per student increased 7.3%, from $14,580 in 2020-21 to $15,645 in 2024-25, but only 35% of schools increased plant debt per student, while 53% decreased it. Among schools with plant debt, the ratio of PP&E to plant debt improved from 2.58 to 2.98, and 68% of those schools saw the ratio improve.

Key Takeaways from the Report

For school leaders and trustees, the report points to a clear planning agenda. First, enrollment strategy cannot be separated from tuition strategy. Second, financial aid must remain mission-centered but financially modeled over multiple years. Third, expense growth needs board-level visibility, especially compensation, benefits, program expansion, and facilities costs. Fourth, fundraising strategy must account for a donor base that may be narrowing even as dollars rise. Lastly, capital planning should move from deferred maintenance triage to disciplined reinvestment.

NBOA’s data do not suggest a sector in crisis. They suggest a sector with real strengths: resilient demand, improved balance sheets, stronger net tuition revenue, and renewed capital investment. But they also show why the next five years may be harder than the last five. The schools best positioned for the future will be those that use their improved financial position not simply to absorb higher costs, but to redesign the financial model around value, affordability, philanthropy, flexibility, and long-term stewardship.

While the report was generally favorable, I was skeptical about the section on tuition discounting and decided to dive a little deeper into the numbers disclosed in the report, as well as other reports that may provide deeper insights.

Tuition Discounting: The Aggregate Number May Hide Merit-Aid Pressure

One of the more reassuring findings in NBOA’s report is that tuition discounting appears to have moderated. Financial aid as a percentage of gross tuition and fees declined from 14.2% in 2020–21 to 13.3% in 2024–25, and the total discount rate, including tuition remission, declined from 16.5% to 15.7%. On the surface, that suggests schools have protected net tuition revenue even as affordability pressures have grown.

However, this is also one of the areas where trustees should read the report with caution. NBOA’s financial aid metric combines need-based aid and merit scholarships. It does not separately show whether schools are increasing merit awards to attract high-performing students, fill seats, improve yield, or compete with neighboring schools. That distinction matters because a school can reduce its overall discount rate while still increasing merit aid, especially if gross tuition is rising faster than aid dollars or if need-based aid is growing more slowly.

Outside sector data suggest that merit aid has been gaining ground over the longer term. NAIS reported that the average merit award increased by about 36% between 2010–11 and 2020–21, and that the number of schools offering merit aid grew by nearly 40%. The Enrollment Management Association’s (EMA) earlier NAIS/DASL-based analysis found that median merit aid grew faster than median need-based aid over a five-year period, even though merit aid remained a small share of total aid dollars. It’s possible because the schools included in this report were not part of a statistical sample that the averages in the report don’t match the reality nationwide, or especially in certain geographical markets. Based on my discussions with friends and colleagues who serve on independent school boards in various areas of the country, a discount rate below 20% does not represent the situation in their cities.

This makes trustee anecdotes about rising merit aid credible, even if they do not directly overturn NBOA’s aggregate finding. The more precise takeaway is that discounting is becoming more strategic and more uneven. Some schools may be successfully limiting discounting as a share of revenue. Others may be using merit awards more aggressively to generate applications, increase yield, or compete in crowded markets. NBOA’s own 2024 tuition discounting research points in that direction: the overall discount rate declined, but tuition remission and merit aid played a larger role in discounting.

For boards, the practical lesson is that the single “discount rate” is too blunt. Schools should separately track need-based aid, merit aid, tuition remission, average award size, new-student versus returning-student discounts, yield by award level, retention by award level, and net tuition revenue by cohort. A stable or declining total discount rate may still mask a recruiting strategy that is becoming more expensive, more competitive, and potentially more damaging to margins.

The fastest way to get NAIS, NBOA, or EMA to collect this data is to have Heads of School, Chief Financial Officers, and Board Officers ask for it. The Department of Education has collected and published college and university data like this for years. Even though private schools do not receive federal government financial aid funds, if Trustees want to avoid the quagmire of tuition discounting in which many private colleges are trapped, they should ask for more transparency.

The post NBOA’s Five-Year Snapshot: Independent Schools Are Stronger, but the Operating Model is Exposed appeared first on Wally Boston.

A report from the Center on Reinventing Public Education (CRPE) and Arizona State’s Mary Lou Fulton College for Teaching and Learning Innovation (CTLI) was issued this month. Titled Navigating the Math Wars: A Practical Guide to the Divides and Debates Influencing Math Instruction, the report provides an excellent perspective on initiatives to improve math instruction in public schools as well as the evidence and debates surrounding those initiatives.

After reading the report, I was preparing to write a blog article highlighting its various components. Then I thought I would see what my upgraded “Pro” version of ChatGPT would do if I asked it to read and analyze the article, then prepare a blog article for me. My exact prompt follows:

A report about the K-12 math initiatives in state public schools was recently published. Could you review it, summarize it and note the key findings in a 1200-1500 word article suitable for my blog? https://crpe.org/wp-content/uploads/Navigating-the-Math-Wars-2026.pdf

ChatGPT’s reply is appended below. It took 6 minutes and 32 seconds. It likely would have taken less time than that if I had not used the Pro version. The output was so good that I decided to conduct a comparison of “prompt responses” from several of the AI tools that I use. When I could, I used the exact wording of my original prompt. All those tools’ responses follow this section as does my usual Final Thoughts section.

ChatGPT Pro’s Response

Here’s a blog-ready draft, about 1,350 words.

The New Math Wars Are Already Shaping State Math Policy

A new report from the Center on Reinventing Public Education, Navigating the Math Wars: A Practical Guide to the Divides and Debates Influencing Math Instruction, offers a timely map of one of the most consequential debates in K-12 education: how public schools should teach mathematics. Published in April 2026, the report is not a simple pro-or-con brief for any one instructional camp. Its purpose is more useful than that. It explains the arguments, identifies where the evidence is contested, and shows how the latest wave of state math policy is already being shaped by ideas associated with the “Science of Math” movement.

The report’s central message is that the Math Wars are not over. They have simply evolved. Earlier battles often pitted traditional instruction against reform math: teacher-led practice, procedural fluency, and standardized accountability on one side; student-centered exploration, conceptual understanding, real-world applications, and less emphasis on timed testing on the other. CRPE traces this divide back to John Dewey’s early twentieth-century distinction between curriculum-centered and student-centered approaches to learning, arguing that later generations hardened what Dewey saw as a tension into rival camps.

What makes the current debate feel new is the rise of the Science of Math. Inspired partly by the Science of Reading, this movement argues that math instruction should be more strongly grounded in cognitive science, empirical evidence, explicit instruction, fluency practice, universal screening, intervention systems, and data-based decision-making. The report notes that math achievement has declined from its 2013 peak, with the pandemic accelerating those concerns, and that many policymakers are now looking for a math equivalent to the reading-policy reckoning. But CRPE is careful: unlike the Science of Reading, the Science of Math has not reached anything close to fieldwide consensus. Prominent math organizations have criticized it for using research too narrowly and for presenting too limited a view of good math teaching.

One of the report’s most helpful contributions is that it refuses to reduce the debate to caricatures. It emphasizes that most traditionalists do not reject conceptual understanding; they argue about when it should be sequenced relative to fluency. Likewise, most reformers do not reject explicit instruction; they worry that it can crowd out inquiry, discussion, and mathematical sensemaking. In other words, many disputes are not pure either/or choices. They are arguments about timing, default emphasis, and what kinds of evidence should guide policy and classroom practice.

The report organizes the “new” Math Wars around five classroom-level dichotomies. The first is conceptual understanding versus procedural fluency. Reform advocates often argue that students should understand mathematical ideas before formal procedures are emphasized. Traditionalists counter that fluency with facts and procedures frees up working memory for more complex problem-solving. The Science of Math position, as summarized by CRPE, is that educators should not assume one must always come first. Conceptual and procedural knowledge can develop together, strengthening each other when taught in an integrated way.

The second divide concerns standard algorithms versus invented strategies. Should students first invent and compare their own methods to build number sense, or should teachers introduce efficient standard algorithms earlier and then build flexibility around them? CRPE frames this as a timing debate. Reformers worry that early algorithms can encourage rule-following without understanding. Traditionalists worry that delaying reliable methods can leave students stuck using slow or error-prone approaches.

The third divide is inquiry-based instruction versus explicit instruction. This may be the most familiar battlefield. Inquiry advocates emphasize authentic problems, exploration, and teacher-guided sensemaking. Explicit-instruction advocates emphasize clear modeling, worked examples, guided practice, structured feedback, and attention to cognitive load, especially for novices and struggling learners. CRPE’s treatment again pushes past slogans: the relevant question is not whether students should ever explore or whether teachers should ever explain, but which approach works best for which learners, for which content, and at which point in the learning sequence.

The fourth debate is productive struggle versus cognitive load. Reform circles often emphasize the value of students wrestling with challenging problems before formal instruction. Science of Math advocates are more skeptical, warning that “productive” struggle can become destructive when students lack the prerequisite knowledge to make progress. The report notes that this disagreement is not merely philosophical; it reflects a deeper dispute over how to interpret the research on productive failure, scaffolding, and achievement outcomes.

The fifth debate centers on timed practice and math anxiety. Critics argue that timed drills can produce stress, damage confidence, and shape students’ negative identities around math. Traditionalists and Science of Math proponents respond that fluency practice, when low-stakes and individualized, can build automaticity and reduce later frustration. CRPE highlights the key nuance: harsh, public, punitive timed tests are not the same as carefully designed retrieval practice.

The report’s most policy-relevant finding is that state action has accelerated. CRPE’s scan covered enacted state math policies from 2022 through late 2025. It found a recurring pattern: many new policies emphasize priorities commonly associated with the Science of Math, including foundational numeracy, universal screening, systematic and explicit instruction, fluency, tiered intervention, data-based decision-making, high-quality instructional materials, and research-aligned professional learning. The authors are careful not to say states have formally adopted the Science of Math as a doctrine. Rather, they conclude that post-pandemic math policy has moved toward a more structured, intervention-oriented model, especially in early grades.

By late 2025, the report identified 18 states plus the District of Columbia with enacted math policies in this space. The list includes Alabama, New Mexico, West Virginia, Arkansas, Florida, Colorado, Louisiana, Texas, Nevada, Mississippi, Kentucky, Tennessee, Washington, D.C., Indiana, Iowa, Montana, Oklahoma, Virginia, and Maryland. Some policies are broad, such as Alabama’s Numeracy Act, Kentucky’s Numeracy Counts Act, Iowa’s Math Counts Act, and Oklahoma’s Math Achievement and Proficiency Act. Others are narrower, such as Texas and Virginia policies focused on access to advanced or accelerated math opportunities.

CRPE’s interpretation of this state activity is balanced. The most common policy themes are data-based decision-making, high-quality materials, and targeted intervention. Foundational numeracy, screening, and explicit instruction also appear frequently, especially in K-5 and K-8 reforms. But the report cautions that the landscape is not uniform. California, for example, moved in a more reform-oriented direction, emphasizing discovery learning, group work, and sociocultural responsiveness. Alabama’s recent fourth-grade NAEP gains are described as noteworthy but not causal proof that Science of Math-aligned policies produced the improvement.

Inside classrooms, the findings are sobering. Reviewing studies involving more than 5,300 math teachers, CRPE concludes that there is still a gap between recommended practice and what often happens day to day. Conventional routines such as worksheets, teacher-led explanation, and lectures remain common, while some unsupported practices, including “learning styles,” persist. At the same time, cognitively demanding tasks, meaningful student discourse, and specific evidence-based interventions such as concrete-representational-abstract sequences and schema-based instruction appear unevenly implemented. The report also makes an important distinction: ordinary lecture or worksheet use is not the same thing as systematic, explicit instruction.

The report’s recommendations are practical. Nationally, CRPE calls for renewed guidance, possibly through a new advisory effort that updates the 2008 National Mathematics Advisory Panel with nearly two decades of additional research from math education, special education, cognitive psychology, developmental science, and the learning sciences. Locally, it recommends math-specific instructional audits, clearer “stop, start, and strengthen” guidance, district math playbooks, and stronger Tier 1 supports such as multiple representations, scaffolded modeling, schema-based problem-solving, and structured opportunities for student response.

My biggest takeaway is that the report is less a call to pick a side than a warning against shallow certainty. The Math Wars persist because the disagreements are real: sequencing matters, evidence standards matter, and instructional defaults matter. But the report also shows that the most useful path for public schools is unlikely to come from treating traditional and reform approaches as mutually exclusive. The stronger question is: What works, for whom, at what point in learning, and with what supports?

For state leaders, that means passing a numeracy law is only the beginning. For districts, it means translating broad priorities into coherent classroom routines. For teachers, it means being supported with tools that are more specific than slogans. And for everyone watching the next phase of K-12 math reform, it means paying close attention not just to which side wins the debate, but to whether students actually get clearer instruction, richer mathematical thinking, and better access to advanced learning.

WEB note: I thought the ChatGPT Pro analysis and writeup was accurate and thorough. It included the major details that I would have included had I written the article. It appeared to be more “academic” than other products generated from my prompts. The output also included references to specific sections of the report. If I were considering submitting this for publication, I would have uploaded it to one of the plagiarism tools like Turnitin. I also would have compared the details in the article to the original report, making sure that ChatGPT captured the pertinent facts accurately.

Google Search’s AI-Enhanced Response

I visited www.google.com to sign into my Gemini account. However, the search bar had an “AI-Enhanced” pop-up next to it, so I opted to try it first. I copied the prompt from ChatGPT Pro and received the following response:

A new report from the Center on Reinventing Public Education (CRPE), “Navigating the Math Wars 2026,” highlights that national student proficiency has hit critical lows, with 2024 data showing only 39% of fourth graders and 28% of eighth graders are proficient in math. The analysis links this decline to pandemic learning loss, a 36% decrease in new math teachers, and ongoing disputes over instructional methods. To improve outcomes, the report urges adopting evidence-based practices that balance conceptual understanding with procedural fluency and the integration of AI tools for student support. For more insights from this report, visit The 74 Million.

I must hand it to Google. It found an article that wrote about the report and included a link to that article in its short reply. However, my intention was to see how various AI tools analyzed and wrote about the report. Clearly, using the AI enhancement feature in a regular Google search didn’t meet my expectations.

Gemini’s Response

I was able to copy and paste the prompt from ChatGPT into Gemini. Its response follows:

The New Math Frontier: Can We Move Beyond the Ideological Math Wars?

For decades, American K–12 mathematics has been a battlefield. On one side, “traditionalists” argue for rote memorization, math facts, and procedural fluency. On the other, “progressives” champion conceptual understanding, inquiry-based learning, and real-world problem-solving.

The Center on Reinventing Public Education’s (CRPE) latest 2026 report, Navigating the Math Wars, suggests that these ideological silos are the primary reason for the “quiet crisis” in our schools. While the rest of the world advances in STEM, American math scores have stagnated or declined, a trend exacerbated by the pandemic and a chronic shortage of qualified educators.

Here is a breakdown of the report’s most urgent findings and what they mean for the future of our public schools.

1. The False Dichotomy: Procedures vs. Concepts

The most significant finding of the CRPE report is that the “Math Wars” are based on a false premise. The report highlights that conceptual understanding and procedural fluency are not mutually exclusive; they are mutually reinforcing.

Research cited in the report shows that students who understand why a mathematical rule works are more likely to remember how to use it. Conversely, students who have “math facts” (like multiplication tables) at their fingertips have the “cognitive real estate” available to tackle complex word problems. The report calls for a “Middle Way” that integrates explicit instruction with deep conceptual exploration.

2. The Staffing Crisis: A 36% Decline in Math Teachers

Perhaps the most alarming statistic in the report is the state of the teaching pipeline. CRPE found that the number of graduates prepared to teach math declined by a staggering 36% between 2012 and 2020, and the numbers have not recovered post-pandemic.

This isn’t just a numbers game; it’s an equity issue. The report notes that under-resourced districts and schools in low-income neighborhoods are the hardest hit, often relying on long-term substitutes or teachers with “emergency” certifications who lack deep subject-matter expertise. Without a “qualified math teacher for every student,” the report argues, even the best curriculum in the world will fail to move the needle.

3. The “Instructional Coherence” Problem

CRPE identifies a lack of “instructional coherence” across state lines and even within districts.

• Curriculum Gaps: Teachers are often left to navigate the gap between research and practice on their own.
• Grade Inflation vs. Proficiency: The report notes a disturbing trend where student GPAs are rising while standardized test scores fall—a “grade inflation” that masks the reality that 57% of grades do not align with student knowledge.
• The Mobility Trap: Because math standards and teaching sequences vary so wildly, students who move schools mid-year often find themselves “missing” foundational concepts that were never taught in their previous school, leading to a cumulative disadvantage.

4. AI as the “New Frontier” in Math Instruction

For the first time in a major math initiative report, Artificial Intelligence takes center stage. CRPE 2026 explores how “Early Adopter” states are using AI to solve the “impossible task” of the modern classroom: teaching 30 students who are all at different math levels.

The report highlights:

• AI Tutoring: Using generative AI to provide “just-in-time” support for students struggling with specific steps in a problem.
• Diagnostic Tools: Using AI to help teachers instantly identify “precursor” skills that a student might have missed in a previous grade.
• Teacher Support: Recommending that states move beyond “AI guidance” and toward “AI implementation” policies that help teachers manage the cognitive load of differentiated instruction.

5. Key Recommendations for 2026 and Beyond

The report concludes with a roadmap for state leaders and school districts:

• Set Bold Goals: CRPE suggests a national goal: “All students prepared for Algebra I by eighth grade by 2030.”
• Prioritize Explicit Instruction: Moving away from “pure inquiry” models toward “Explicit Instruction”—where teachers directly explain concepts while supporting student engagement—especially for students who are already behind.
• Reform Teacher Prep: Encouraging states to raise the bar for math certification, ensuring elementary teachers receive more than just “minimal” math-specific training.
• Build a “Science of Math” Consensus: Similar to the Science of Reading, the report calls for a national advisory panel to cut through ideological debates and provide educators with a clear, evidence-based framework for what works.

Final Review: Why This Report Matters

Navigating the Math Wars 2026 is a wake-up call. It shifts the conversation away from “Who is right?” to “What works for students?” By focusing on the intersection of teacher shortages, curriculum coherence, and the potential of AI, CRPE has provided a pragmatic blueprint for recovery.

For parents and educators, the message is clear: the era of picking sides in the math wars must end. The “Middle Way”—grounded in evidence and supported by technology—is the only way to ensure that the next generation of American students isn’t left behind in a data-driven world.

WEB note: I thought the Gemini response was an accurate overview but did not include some of the details that the ChatGPT Pro response included. I also thought its organization including headers was punchier and oriented toward blog publications like Medium. It fell short of providing the depth of commentary and analysis that I try to include in one of my blog article reviews.

NotebookLM’s Response

I enjoy using NotebookLM as a repository for projects that involve reading and summarizing many different articles. The only difference in my prompt from ChatGPT Pro was that I had to upload the report’s PDF. I kept the prompt the same, except that I eliminated the link to the PDF. That made the prompt’s wording a little awkward, but for this exercise, I wanted it to be the same if possible. The response is appended below:

Decoding the Math Wars: Key Findings from the New K-12 Math Initiatives Report

For decades, educators, policymakers, and parents have been locked in a fierce, unresolved debate over the best way to teach mathematics in K-12 public schools. What began as a dispute between advocates of traditional, teacher-led instruction and proponents of reform-oriented, student-centered learning has evolved into a complex national conversation. A recently published comprehensive report by the Center on Reinventing Public Education (CRPE), titled Navigating the Math Wars: A Practical Guide to the Divides and Debates Influencing Math Instruction, sheds light on the history, current conflicts, and future of math education.

If you are a school leader, educator, or parent trying to make sense of shifting curricula and new legislative mandates, here is a detailed review and summary of the report’s key findings.

The Historical Roots of the “Math Wars”

To understand today’s debates, we must look back to the early twentieth century. Education reformer John Dewey identified a fundamental divide in pedagogy: should education begin with the accumulated knowledge of the teacher, organized into a formal curriculum, or should it center on the needs, interests, and abilities of the student? Over time, this nuanced observation hardened into polarized camps.

Traditionalists came to value explicit, sequenced, teacher-led instruction focusing on procedural fluency and the mastery of established mathematical structures. Reformers, conversely, championed inquiry-based, student-centered learning, emphasizing conceptual understanding, problem-solving, and real-world application.

This philosophical tug-of-war has played out across several eras. The post-Sputnik 1950s brought the abstract, highly rigorous “New Math,” which was eventually rejected in the 1970s for a “Back to Basics” approach. The 1980s and 1990s saw the rise of Reform Math, heavily guided by the National Council of Teachers of Mathematics (NCTM), which sought to de-emphasize rote memorization. In 2010, the Common Core State Standards attempted to broker a peace by integrating conceptual understanding, procedural fluency, and application, but implementation issues largely reignited old tensions.

The Emergence of the “Science of Math”

In recent years, stagnant student performance—exacerbated by the COVID-19 pandemic—has catalyzed a new movement known as the Science of Math. Officially formed in late 2020, this movement is largely led by researchers from special education and school psychology. Inspired by the successful “Science of Reading” movement, proponents argue that instructional decisions must be grounded in rigorous, empirical cognitive science.

The Science of Math pushes back against what it deems “pseudoscientific” practices. It strongly advocates for systematic, explicit instruction, regular practice to achieve automaticity, multi-tiered systems of support (MTSS), and data-based decision-making. While this movement aligns with many traditionalist priorities, it has drawn sharp criticism from prominent mathematics education organizations who argue it promotes too narrow a vision of math instruction and misapplies certain research.

The Five Frontlines of Modern Math Instruction

The CRPE report breaks down the contemporary “Math Wars” into five major dichotomies that directly impact classroom teaching:

1. Conceptual Understanding vs. Procedural Fluency Reform advocates argue that students must grasp underlying mathematical concepts before learning formal procedures, warning that rote memorization leads to “fragile knowledge”. Traditionalists contend that fast, accurate recall of arithmetic facts frees up limited working memory for complex problem-solving. The Science of Math movement claims research does not support a strict sequence; instead, conceptual and procedural knowledge should be taught simultaneously, as they develop bidirectionally and reinforce one another.
2. Standard Algorithms vs. Invented Strategies Should students invent their own math strategies, or should they be taught standard algorithms immediately? Reformers often delay standard algorithms to encourage children to build deep number sense through their own computational methods. Traditionalists argue that standard algorithms are essential, efficient tools, and withholding them leaves struggling students relying on slow, error-prone methods. The Science of Math strongly promotes explicitly teaching standard algorithms alongside their conceptual meaning to guarantee students know when and how to apply them.
3. Inquiry-Based Instruction vs. Explicit Instruction Reformers champion inquiry-based learning, where students construct their own knowledge through rich, open-ended tasks with the teacher acting as a facilitator. Traditionalists counter that unguided discovery is inefficient for novices and can lead to deep misconceptions. Drawing on cognitive load theory, Science of Math advocates call for systematic, explicit instruction as the default for foundational content, arguing that novices simply lack the working memory capacity to juggle the unknowns of unguided inquiry.
4. Productive Struggle vs. Scaffolded Support The reform camp popularized the concept of “productive struggle,” asserting that wrestling with difficult problems builds perseverance, resilience, and conceptual insight. Traditionalists caution that unsupported struggle quickly turns into frustration and anxiety, favoring heavily scaffolded instruction rooted in Vygotsky’s Zone of Proximal Development. Science of Math proponents warn against “destructive struggle,” arguing that having students attempt to apply skills they do not yet possess is an ineffective use of instructional time.
5. Timed Practice vs. Math Anxiety Reformers point to student self-reports to claim that timed math tests cause early-onset math anxiety, blocking working memory and performance. Therefore, they advocate for untimed assessments focusing on ease and accuracy. Conversely, traditionalists and Science of Math advocates argue there is no causal evidence linking low-stakes timed practice to lasting anxiety. They believe that timed drills serve as vital retrieval practice, preventing future anxiety by ensuring basic facts are automated before students face complex, higher-order math tasks.

The Shifting Tide of State Policies

One of the most revealing sections of the CRPE report is its analysis of recent state legislation. Policymakers are watching math achievement closely, with National Assessment of Educational Progress (NAEP) scores sliding since their 2013 peak. State legislatures are increasingly adopting frameworks mirroring the Science of Math’s priorities.

Remarkably, the report notes that math-related policies are currently accumulating more quickly than early reading laws did during the nascent stages of the Science of Reading movement. Between 2022 and late 2025, 18 states and the District of Columbia enacted significant math policies.

These new policies converge around seven key priorities:

• Foundational numeracy focus (K-5)
• Universal screening in mathematics
• Data-based decision making
• Systematic, explicit instruction
• High-quality instructional materials
• Targeted intervention for struggling students (MTSS)
• Professional learning aligned to math education research

For example, Alabama’s 2022 Numeracy Act mandates universal early numeracy screeners, evidence-based curricula, and K-5 math coaches. Florida’s recent laws emphasize screening and intervention for students with dyscalculia. Texas has focused on expanding equitable access by establishing automatic enrollment into advanced middle-school math courses for students meeting objective criteria. Louisiana now requires K-5 teachers to complete 50 hours of professional development explicitly focused on evidence-based instructional practices.

The Classroom Reality: Mind the Implementation Gap

Despite the fierce debates and shifting policies, what is actually happening in the classroom? The report reviews studies encompassing over 5,300 mathematics teachers, revealing a frustrating reality: a significant implementation gap exists between recommended practices and daily instruction.

Conventional routines, such as relying heavily on worksheets, remain widespread, while high-quality reform practices (like meaningful student discourse and cognitively demanding tasks) are rarely executed with consistency. Furthermore, while traditional teacher-led lectures are common, they frequently lack the rigorous, carefully structured components of true “systematic, explicit instruction”. Highly effective evidence-based interventions for struggling learners—such as the Concrete-Representational-Abstract (CRA) sequence and schema-based problem solving—are broadly underutilized by educators.

Moving Forward: How Do We End the “Forever War”?

The CRPE report makes it clear that evidence alone will not settle the Math Wars until the educational community can agree on how to interpret and apply that evidence. To move forward constructively, the report outlines several actionable steps for national leaders, districts, and educators:

1. Build an Interdisciplinary Evidence Base: Math instruction research spans special education, cognitive science, developmental psychology, and the learning sciences. Rather than existing in “epistemic bunkers,” these disciplines must collaborate to provide coherent, comprehensive guidance that addresses both effective scaffolding and meaningful opportunities for inquiry.
2. Leverage Cognitive Load Theory (CLT): The report strongly endorses CLT as a framework for designing instruction. Because human working memory is limited, instruction should be pared down to essentials. Teachers should use visual diagrams combined with audio narration to exploit working memory resources, provide external memory supports (like procedure guides), and gradually transition from heavily worked examples to independent problem solving.
3. Renew National Guidance: The field desperately needs an updated version of the 2008 National Mathematics Advisory Panel. A bipartisan, broadly representative national panel could synthesize two decades of new developments across various scientific fields, reducing the current fragmentation that leaves school districts guessing.
4. Conduct Local Instructional Audits: School districts do not have to wait for national consensus. Leaders should conduct math-specific instructional audits to identify gaps between intended and enacted instruction. Districts must bring high-leverage, evidence-based supports—like explicit modeling, multiple representations, and structured discourse—into general “Tier 1” classroom instruction, rather than reserving them solely for special education or intervention settings.

Conclusion

The CRPE’s 2026 report proves that the Math Wars are far from over, but the landscape is rapidly shifting. The introduction of the Science of Math, coupled with an aggressive wave of state legislation, is pulling K-12 math education toward a more structured, explicit, and data-driven paradigm.

Ultimately, finding success in the classroom will likely require drawing selectively from both the traditional and reform traditions. Students deserve math instruction that simultaneously develops automatic, usable skills and deep, meaningful conceptual understanding. By embracing interdisciplinary research, adopting principles of cognitive load theory, and actively supporting teachers with high-quality materials and coaching, we can finally begin to prioritize student outcomes over ideological battles.

WEB Note: NotebookLM’s response included as many details as ChatGPT Pro’s response and more than Gemini’s. The format was not as “punchy” as the Gemini response and was likely designed to flow like the study guide that NotebookLM was intended to be. Like ChatGPT Pro, if I planned to publish this article on my website or use it otherwise, I would want to submit it to a tool like Turnitin. I would also match the details with the original report for accuracy.

Claude’s Response

I was able to copy and paste my original prompt from ChatGPT Pro in Claude (Sonnet 4.6). Its response is appended below:

The Math Wars Are Back — And This Time, the Fight Is Over Science Itself

A new report from the Center on Reinventing Public Education maps the divides shaping K-12 math instruction — and why they matter more than ever.

If you’ve ever argued with another parent about whether your child’s school should be teaching “standard algorithms” or “invented strategies,” or found yourself baffled by a Common Core homework problem, you’ve stumbled into one of American education’s most persistent and heated controversies: the Math Wars.

A report published in April 2026 by the Center on Reinventing Public Education (CRPE) — Navigating the Math Wars: A Practical Guide to the Divides and Debates Influencing Math Instruction — offers the most comprehensive roadmap to date of these conflicts. It’s essential reading not just for policy wonks, but for any parent, teacher, or school leader trying to make sense of why something as ostensibly straightforward as teaching children mathematics has become so contentious for so long.

A Century-Old Fight in New Clothing

The report traces the Math Wars back further than most people realize — all the way to the early 20th century, when philosopher John Dewey identified a fundamental tension in education: should learning begin with organized curriculum knowledge, transmitted teacher-to-student, or with the student’s own interests and experiences as the starting point? That question, Dewey warned, should not become a war. It did anyway.

Over the following century, the battle swung back and forth. After the Soviet launch of Sputnik in 1957 prompted a push for rigorous “New Math” — introducing set theory and symbolic logic into K-12 classrooms — parent and educator frustration eventually produced a “Back to Basics” backlash by the early 1970s. Then, in 1989, the National Council of Teachers of Mathematics (NCTM) released landmark standards emphasizing problem-solving, reasoning, and real-world application over rote memorization. The resulting curricula — Everyday Mathematics, TERC Investigations, MathLand — sparked fierce opposition and gave the conflict its name.

The 2010 Common Core State Standards were supposed to end the hostilities by weaving together conceptual understanding, procedural fluency, and real-world application into a unified framework. Instead, they reignited them. Unfamiliar terminology and methods, amplified by social media outrage, entangled math education in the broader culture wars of the 2010s.

Then the pandemic hit, accelerating a decline in student achievement that had already been underway since 2013. Into this climate stepped the newest combatant: the Science of Math.

The New Front: Science of Math vs. the Field

Officially formed in December 2020 and modeled on the successful “Science of Reading” movement, the Science of Math argues that classroom instruction should be grounded in cognitive science and empirical research — and that a great deal of current math teaching relies on what it calls “pseudoscientific” practices. Led primarily by researchers from special education and school psychology, the movement advocates for systematic, explicit instruction, regular practice to build automaticity, multi-tiered systems of support, and data-based decision-making.

The Science of Reading parallel is worth examining carefully, and the CRPE report does so. That movement achieved something close to a professional consensus and drove sweeping legislative change — phonics-based reading instruction is now mandated in dozens of states. Could math be headed for a similar reckoning?

The report’s answer is nuanced: maybe, but not so fast. Unlike reading, where the evidence base eventually coalesced, the Science of Math has attracted sharp criticism from the field’s most prominent organizations, including the NCTM and the National Council of Supervisors of Mathematics (NCSM). Critics argue the movement applies research too narrowly, draws disproportionately from special education populations, and promotes too limited a vision of what rich mathematics teaching can look like. The debate has thus opened a genuinely new front in the Math Wars — not just about how to teach math, but about what kind of evidence should guide those decisions.

Five Fault Lines, Carefully Mapped

The heart of the CRPE report is its detailed examination of five major instructional dichotomies. A key insight running through all of them: these are rarely pure either/or choices. The real disputes are about sequence, default emphasis, and evidentiary standards — not absolute incompatibilities.

1. Conceptual Understanding vs. Procedural Fluency

Should students understand the why before they master the how, or does drilling procedures actually help build understanding? NCTM’s 2014 position — that procedural fluency should follow and build on conceptual understanding — reignited this debate. The Science of Math pushes back, citing research suggesting the relationship is bidirectional: conceptual and procedural knowledge reinforce each other when taught together. Most researchers on both sides now agree that students need both; the fight is over which comes first.

2. Standard Algorithms vs. Invented Strategies

Reform-oriented educators often favor letting students develop their own computational methods before introducing standard algorithms, arguing this deepens number sense. Traditionalists counter that withholding standard algorithms too long leaves struggling students — often those with less support at home — floundering with inefficient, error-prone methods. The equity dimension here is real: expecting all students to independently discover efficient strategies may inadvertently disadvantage the students most in need of clear instruction.

3. Inquiry-Based Instruction vs. Explicit Instruction

Perhaps the most divisive of all the dichotomies. Reform advocates argue that students learn more deeply when they investigate mathematical ideas actively, with a teacher as guide rather than lecturer. Traditionalists draw on cognitive load theory to argue that novice learners, whose working memories are easily overwhelmed, benefit most from structured, step-by-step instruction with immediate feedback. The research debate here is genuinely contested, but a growing number of scholars on both sides are calling for a more nuanced question: not “which is better,” but “which works best for whom, when, and for what kind of learning goal?”

4. Productive Struggle vs. Scaffolded Support

How much difficulty is actually good for learning? Reform math has popularized the concept of “productive struggle” — the idea that wrestling with hard problems, even failing at first, builds resilience and deeper understanding. NCTM explicitly added it to its list of effective teaching practices in 2014. But traditionalists warn that unproductive frustration is a real risk, particularly for students who already lack confidence, and that timely support and scaffolding are not signs of failure but essential ingredients of effective teaching.

5. Tracking and Acceleration

The report also addresses the contentious question of whether students should be grouped by ability or given access to accelerated content. Traditionalists tend to favor early algebra access for high-achieving students; reform advocates emphasize equity concerns and argue that tracking can entrench disadvantage. This debate has been especially charged in California, where a recent revision to the state’s math framework drew national attention and fierce pushback from both sides.

Why the Battlefield Stays Frozen

One of the report’s most valuable contributions is its analysis of why these wars never seem to end. Both sides have built what researchers call “epistemic bunkers” — insulated professional communities that generate, circulate, and validate evidence largely within their own networks, while discounting research from the other camp. Separate conferences, separate journals, separate professional organizations: the structural barriers to genuine dialogue are formidable.

The result, as the report puts it, is “a fragmentation of knowledge that produces incoherent or incomplete guidance for practice.” School leaders and teachers are left navigating competing recommendations with little help distinguishing strong evidence from advocacy dressed up as research.

What Happens in Actual Classrooms?

The CRPE report also surveys what’s actually happening in policy and practice since the pandemic. State-level Science of Math legislation has begun to appear, with some states moving to mandate or restrict certain instructional approaches. Meanwhile, surveys of math teachers suggest considerable variation in practice — many teachers draw eclectically from both traditions, adapting to their students’ needs in ways that don’t map neatly onto either camp’s preferred model.

What the Report Recommends

The CRPE researchers are careful not to adjudicate the underlying research debates — that’s explicitly outside the report’s scope. Instead, their recommendations are procedural and structural:

For local leaders, the report urges investment in professional development that helps teachers understand and navigate these debates themselves, rather than simply mandating one approach. It calls for selecting curriculum materials based on evidence of effectiveness, not ideological alignment, and for building school cultures where teachers can examine evidence and adapt practice collaboratively.

At the national level, the report calls for greater investment in research that bridges the epistemological divide — studies designed to examine not just whether explicit instruction or inquiry works, but how, for whom, and under what conditions. It also calls for professional organizations on both sides to model the kind of engagement across difference that they’re currently failing to practice.

The Bottom Line

The Math Wars are not going away. But the CRPE report makes a compelling case that they don’t have to remain as destructive as they’ve been. The underlying disagreements are real — about pedagogy, about evidence, about what mathematics education is fundamentally for. But the polarization has been amplified by institutional incentives, tribal epistemology, and a media environment that rewards conflict over nuance.

For parents, teachers, and school leaders trying to find their footing, this report is an unusually honest and balanced guide to terrain that has long been needlessly treacherous. The Math Wars may be a century old. Understanding them clearly is the first step toward fighting them more wisely — or better yet, transcending them entirely.

“Navigating the Math Wars: A Practical Guide to the Divides and Debates Influencing Math Instruction” was published in April 2026 by the Center on Reinventing Public Education (CRPE). The full report is available at crpe.org.

WEB note: I liked Claude’s response. It was detailed, accurate based on my memory from reading the original report, and delivered the analysis in a punchier article more oriented to my blog. If I was looking for a quick and dirty blog post, I’d consider using Claude to generate the article. Of course, I’d also submit the article to a tool like Turnitin to verify that sections of the report were original writings and not a verbatim reprint of language from the report (please note that I am not endorsing Turnitin as an accurate tool for determining whether an article was written by AI – I leave that to others with more experience using plagiarism detection tools).

A Few Final Thoughts

Approximately a year and a half ago, I utilized several AI tools to analyze the results of the presidential election. Perhaps because this article’s analysis was more focused (i.e., reviewing a specific report), the tools’ output has improved. At the same time, the nature of AI tools is that they improve over time as their algorithms learn from the datasets that they ingest. Other than my AI-enabled Google search output, the responses from ChatGPT Pro, Gemini, NotebookLM, and Claude were impressive. I could have used any of them for the basis of a blog article about the report.

NotebookLM produced the most detailed report. Upon reflection, that’s not surprising given that the tool was developed for academics/students to use to take notes, summarize those notes, and prepare for tests or write papers.

ChatGPT Pro produced a report that sounded a lot like my personal blog voice. I suspect that it has familiarized itself with my writing style, given that I have asked the earlier versions to analyze some of my previous posts or to produce infographics that I can use for my posts.

Gemini produced a punchy, fact-laden blog article. It read to me like something that I might publish on Medium. It was a little shorter than Claude’s but included most of the pertinent facts.

Claude’s response was more than adequate. It was more detailed than Gemini’s article. The output was not only detailed but also delivered in a punchy format much akin to the style many bloggers use on their websites or a publishing site like Medium..

All in all, these tools have improved. All but one provided sources for various data points in their outputs. Having just read the report before I utilized these tools, I can say that they captured the main points of the report and summarized it well. If I had a deadline to meet or were a blogger who posted daily, I could have posted one of these as my work product without significantly modifying it.

Fortunately, I still enjoy reading, analyzing, and writing. I can see, however, why economists predict that professionals’ productivity is increasing. I can also understand why companies are eliminating some professional positions or reducing the number of new hires in those fields. The employment picture is going to remain messy, and a major factor is the ability of these tools to replace many hours of reading, analysis, and writing.

The post AI Analysis and AI Reporting Tools Are Getting Way Better appeared first on Wally Boston.

The introductory comments from the co-chairs note that mass adoption of AI has occurred faster than the adoption of the personal computer or the internet. They also report that leading AI companies are reaching revenue scale faster than most technology companies have, and that global corporate investment in AI more than doubled in 2025.

Unlike some of my previously published summaries, I chose to focus on a single chapter, Chapter 7 – Education. That’s not to say that I skipped over the other chapters. It’s just a lengthy and weighty publication, and most of my readers are interested in education.

The education material in the 2026 Stanford AI Index is one of the report’s strongest sections. Many of us have heard that education’s AI problem is mainly about cheating, productivity, or national competitiveness. The report’s chapter about education deals with all three as consequences of institutional lag.

That fits the report’s broader picture: AI capability and adoption are accelerating quickly. Organizational adoption of AI has reached 88 percent, and Stanford’s overview says four in five university students now use generative AI, even as responsible AI measurement and governance trail behind deployment. Education is where the report’s main theme becomes solid. The AI tools are already here, but the systems meant to govern them are not.

AI Realignment of Education Choices

One of the chapter’s biggest strengths is conceptual clarity. The authors separate “AI in education” from “AI literacy” and from “AI education.” Those phrases are frequently blurred together, yet they point to three different goals:

1. using AI tools in classrooms
2. giving everyone a grounded understanding of AI and its risks
3. teaching the technical skills required to build AI systems

The report is also candid about its own evidence base. The authors note that data on AI education is fragmented, lagging, and incomplete, and in the past, the report relied on computer science education as a proxy. This year, the authors added AI-relevant majors as determined by the January 2025 White House AI Talent Report. The authors write that education systems are being pushed to redesign around AI before they can even measure access, curriculum quality, or outcomes very well.

At the postsecondary level, the chapter captures a real realignment. U.S. undergraduate computer science enrollment at four-year universities fell 11 percent between 2024 and 2025, but AI-related software degrees kept growing, especially at the master’s level. The report notes an 82 percent increase in AI-software-related master’s graduates between 2022 and 2024, including 17 percent growth from 2023 to 2024, while AI-hardware-related degrees were flat or declining.

It’s important to note that most AI-related graduate program students are male, non-U.S. resident students. The report’s authors write that the Trump Administration’s crackdown on student visas may impact these enrollments in the near term.

The change in program majors suggests students are not abandoning computing so much as moving toward narrower, more explicitly AI-linked credentials. Stanford links this to the wider labor-market story that entry-level software work appears to be under pressure even as overall AI hiring grows. For colleges, that implies a shift from offering generic technical training toward helping students combine AI capability with domain knowledge and judgment.

The talent-pipeline data points in the same direction. The number of new AI PhDs in the United States and Canada rose 22 percent from 2022 to 2024, and the report notes that the additional growth went to academia rather than industry. Industry still employed the largest share of new AI PhDs in 2024, at 62.75 percent, but that was down from a 77 percent peak in 2022, while academia rose to 31.59 percent.

For several years, the dominant worry was that universities were becoming mere feeders for frontier labs. This year’s data suggests something more mixed: academia may be regaining some pull, even if industry remains the main destination. That matters for who teaches future cohorts and where public-interest expertise in AI still lives.

AI Adoption is Student-Led

The heart of the chapter is student use. The report illustrates how normalized AI already is. Among university students surveyed across 15 countries, 80 percent said they had used generative AI to support learning in 2025, double the 40 percent reported in 2023. Usage ranged from 95 percent in Indonesia to 67 percent in both the United States and the United Kingdom, and 56 percent of users said they asked AI questions at least once a day.

In U.S. middle and high school populations, the report cites estimates ranging from 50 percent to 84 percent using AI for school-related tasks. High school students report using it most for research, essay revision, and brainstorming. University students are more likely to use it to understand a concept, which the report identifies as their top use. That is different than the claims that student AI use is mostly about evasion. Much of it appears closer to study support and academic scaffolding.

What Happens When Institutions Lag

Where the chapter becomes alarming is governance. Only about half of U.S. middle and high schools have policies on AI use. Of those, 28 percent allow AI in some circumstances, and 22 percent prohibit it. Yet policy presence is not policy clarity. Only 36 percent of students described their school’s AI policy as extremely clear, and 47 percent said they wanted to use AI for schoolwork but were unsure whether it was allowed. Teachers were even more negative: just 6 percent said their schools had clear, comprehensive policies.

The report’s most important educational insight may be that ambiguity is itself a system design choice. In practice, unclear rules shift judgment to individual teachers and students, making the actual policy informal and unequal. Since schools with AI policies are more likely to be wealthier and more urban, that ambiguity also becomes an equity issue.

The equity problem appears again when the chapter uses computer science (CS) access as a proxy for AI readiness. In 2025, 91 percent of large U.S. high schools offered foundational CS, compared with 77 percent of medium-sized schools and just 44 percent of small ones. Suburban schools were more likely to offer it than rural or urban schools, and non-Title I schools slightly more than Title I schools.

Access also varied by race and ethnicity. Asian students had the highest access to foundational CS at 91 percent, while Native American students had the lowest at 70 percent. At the same time, access should not be confused with participation. Based on data from 42 states, only 6.1 percent of students were enrolled in CS in 2024-25. This is the chapter’s clearest warning against simple “teach AI everywhere” rhetoric. Standards matter, but access, staffing, scheduling, and actual student uptake matter more.

The policy section is both encouraging and sobering. As of January 2026, 30 U.S. states had issued guidance on AI in education. Seventeen states had clarified that computer science is foundational to AI, and five had allocated specific professional-development funding for AI education. Forty-five states have adopted K-12 CS standards, but most include AI only minimally, usually at the high school level. 10 states make no specific mention of AI.

New Architecture is Emerging

Revised CSTA (Computer Science Teachers Association) standards are due in summer 2026, and the April 2025 executive order on AI education created a federal task force. It pushed agencies to prioritize AI in grants, teacher preparation, apprenticeships, and workforce pathways. The report’s authors are persuasive when they argue that the real problem is implementation. State guidance is largely nonbinding and decentralized; teacher training lacks state-level standards or stable funding; and AP Computer Science still lacks AI-specific content. The United States government is developing policy language faster than its capacity to implement.

The global picture sharpens that point. The report estimates that 93 percent of countries taught CS in 2025, but only 30 percent mandated it, while 63 percent made it available only in at least some schools. China and the UAE stand out because they moved beyond rhetoric and mandated AI education for the 2025-26 school year, with grade-level curricula that begin with literacy and move toward system design and ethics.

Elsewhere, the pattern is more tentative. South Korea introduced AI textbooks, then reversed course amid pushback. Greece partnered with OpenAI to train teachers, and Estonia launched a pilot serving 20,000 students and 3,000 teachers. The chapter does not confuse pilots or procurement with real curriculum integration, and it repeatedly reminds the reader that global education data is messy. Computer Science (CS) and Information Communications and Technology (ICT) are often conflated with digital literacy, and major countries are missing from some datasets. That makes the chapter’s claims more credible, not less. Its picture is directional, careful, and useful.

My Perspective

This is one of the best chapters in the 2026 report. The authors demonstrate that education has entered the AI era not through orderly reform but through mass student adoption, patchy policy, uneven teacher preparation, and widening pressure to upskill across the lifespan.

The closing section on skill acquisition reinforces that point. AI skills are increasingly being built outside formal education, through certificates, online learning, and work itself, and those skills remain unevenly distributed across countries and genders.

For schools and universities, the decision to “allow AI” has already been forced on them. The question that needs to be answered now is how to offer what informal AI learning cannot: shared standards, expert guidance, equitable access, stronger assessment, and a civic framework for using powerful tools well. That, to me, is the central educational insight of this year’s AI Index.

Schools and universities that work to establish best practices in AI policy and, at the same time, encourage faculty to incorporate lessons and experiences that use AI will forge ahead of those waiting on the sidelines “to see where this settles out.” The evidence is conclusive. Use of AI tools will remain on an upward trajectory. Employers will seek college graduates who can effectively use AI tools. Parents will expect their children to be prepared for the AI-enabled workplace of the future. It’s time to narrow the utilization gap between faculty and students because AI advancements are not slowing down.

The post AI in Academia: What Stanford’s 2026 AI Index Tells Us appeared first on Wally Boston.

That distinction between formation and function is not new. According to McCord, it traces back to Wilhelm von Humboldt, a Prussian diplomat and the founding visionary of the modern research university. But as McCord observes, while the architecture of that vision still stands, its animating purpose has largely been hollowed out. Credentialism has overtaken formation.

What makes the essay timely is not simply its critique of universities. It is the context in which that critique is delivered: the rise of artificial intelligence. McCord’s premise is that AI doesn’t just change what we know. It changes how we think, and over time, whether we think at all.

The Quiet Dependency Problem

One of the most compelling observations in the essay is that many people who use AI today are drawing on intellectual habits formed before AI became ubiquitous. They know when to trust a result and when to challenge it because they learned how to read carefully, argue rigorously, and sustain attention without technological scaffolding.

That point deserves more emphasis than the article gives it. We are now entering a period in which those habits may no longer be formed at scale. If that happens, the consequence is not simply weaker graduates. It is a subtle but profound shift in human capability:

• from judgment to prompting
• from understanding to summarization
• from reasoning to selection among outputs

The danger is that students will never develop the cognitive independence required to use AI well. That is where the article’s critique of career-oriented education becomes more than philosophical; it becomes structural.

The False Tradeoff Between Utility and Formation

McCord is right to push back against the growing chorus, from technologists and investors alike, that higher education should focus narrowly on employable skills. In a recent article, I argued for the merits of a liberal arts degree in the era of AI. Others often frame the argument incorrectly as a binary choice between preparing students for careers or developing them as whole human beings.

The historical insight behind Humboldt’s Bildung, as used to develop the University of Berlin, is that these are not competing goals. A broad intellectual formation produces better professionals who are more adaptable, more resilient, and better able to operate in uncertain environments.

The issue is that modern institutions have redefined “career preparation” in overly narrow terms, such as:

• mastering current tools instead of underlying concepts
• optimizing for first job placement instead of long-term adaptability
• rewarding correctness instead of intellectual rigor

In an AI-mediated world, those choices become liabilities. Why? Because the half-life of technical skills is shrinking, while the value of judgment is rising. I used ChatGPT to generate the infographic below illustrating the differences between outsourcing cognition and developing it.

AI as a Force Multiplier, For Better or Worse

Far from eliminating the need for critical thinking, Artificial Intelligence amplifies the consequences of having or lacking it. A student with strong analytical habits uses AI to:

• test hypotheses
• explore alternative framings
• accelerate iteration

A student without those habits uses AI to:

• generate answers that they cannot evaluate
• accept outputs they do not fully understand
• outsource thinking entirely

The same tool produces radically different outcomes depending on the user’s intellectual formation. This is the central point McCord emphasizes that higher education has not yet fully absorbed. We cannot simply integrate a new technology into the curriculum without redefining the baseline assumptions about cognition itself.

What Educators Should Do Differently

If we take the argument seriously, the response cannot be cosmetic, merely adding AI modules or updating course content. It requires rethinking how we design learning environments.

Three shifts seem particularly important.

1. Design for Cognitive Resistance, Not Efficiency

McCord emphasizes that formation occurs through “encounter with what resists us,” namely, difficult texts, complex problems, and sustained efforts. AI removes the friction students encounter. A good curriculum must reintroduce it intentionally. That means including:

• assignments that cannot be completed through simple prompting
• problems that require iteration, failure, and revision
• evaluation based on reasoning process, not just final output

In other words, we must preserve the conditions under which thinking develops.

2. Separate “AI-Assisted Work” from “AI-Free Work”

I suspect that most institutions are drifting toward an unexamined middle ground on which the inclusion of AI in courses broadly without redefining expectations is encouraged. That is a mistake. To build the same educational foundation as many of us using AI today, students will need:

• AI-free environments to build foundational reasoning skills
• AI-integrated environments to learn how to use tools responsibly

Without that separation, we risk short-circuiting the very capabilities the technology claims to enhance.

3. Assess Judgment, Not Just Knowledge

Traditional assessment models reward recall, procedural accuracy, and compliance with instructions. AI can now outperform students on all three. What AI cannot replicate is judgment:

• knowing when an answer is wrong
• recognizing ambiguity
• framing better questions

Curriculum should shift toward:

• oral defenses
• iterative projects
• open-ended problem framing

If we continue to assess what machines do well, we will train students to become moderators of the output from AI. Our future college graduates will know when an answer is wrong, will recognize ambiguity, and will reframe questions for better AI-generated output.

The Institutional Constraint Few Want to Address

There is a deeper tension underlying all of this, one that McCord’s essay hints at but does not fully explore. Higher education institutions are not insulated environments. They are shaped by incentives:

• enrollment pressures
• student satisfaction metrics
• job placement statistics
• cost structures tied to scale

Those incentives push toward efficiency, standardization, and measurable outcomes, all of which favor function over formation. Institutions tend to optimize for what their economic structures reward, even when that diverges from their stated mission.

What should educators do? is too simple a question for this moment. We need to be asking, What are institutions structurally able to sustain over time? That is a governance issue as much as a pedagogical one.

A Closing Thought for Presidents, Provosts, and Trustees

McCord’s essay asks a deceptively simple question:

What is a student, and how should education serve them?

In the age of AI, that question becomes operational. If higher education continues down its current path of optimizing for short-term employability while outsourcing cognition to machines, it risks producing graduates who are less capable of independent thought than the generations before them.

And if that happens at scale, the implications extend well beyond the labor market. The implications will impact the foundations of civic life, leadership, and human autonomy itself. The irony is that the more powerful our tools become, the more we depend on the quality of the minds using them.

Higher education cannot afford to forget that.

The post You Are Not a Function – Is Higher Education Acting Like One? appeared first on Wally Boston.

The authors of that 2016 working paper, Jaison Abel and Richard Deitz, examined why many recent college graduates were underemployed. Among their key findings were:

• Underemployment rose sharply, even as the economy recovered in the years following the Great Recession
• Underemployed does not mean that graduates are working in service jobs
• The likelihood of underemployment varies significantly by program major
• Underemployment is often temporary, not a permanent phase
• Initial job placement can play a critical role in long-term career trajectory

It’s not surprising that the New York Fed’s report organizes its data by program major based on the findings of Abel and Dietz. With data from recent graduates of 73 programs, the Fed Report provides the following five columns of data points for each program:

• Unemployment rate
• Underemployment rate
• Median wage early career
• Median wage mid-career
• Share with graduate degree

Table 1 below lists the program majors in the order in which the report downloaded them. Tables 2 through 6 that follow will look at the top 15 (20%) and bottom 15 (20%) for each of the five categories.

Unemployment Rates by Program Major

Table 2A provides the 15 lowest (best) program major unemployment rates. It’s not surprising, given the national teacher shortage, that four of the five education majors are among the programs with the lowest unemployment rates. The exception is the early childhood education major. Engineering, Construction Services, Nursing, Agriculture, and Social Services are programs tailored to specific occupations as well.

Table 2B provides the 15 highest (worst) program major unemployment rates. It did not surprise me that Anthropology majors had the highest unemployment rate, but it surprised me that six of the fifteen majors in this grouping were STEM programs. One explanation for the STEM programs in this table is that graduates of these programs may be more inclined to enroll in graduate school.

Underemployment Rates by Program Major

Table 3A provides the 15 lowest (best) program major underemployment rates. One key observation is that all but one of these programs (Miscellaneous Physical Sciences) match specific careers.

Table 3B provides the 15 highest (worst) underemployment program major underemployment rates. Notably, only one STEM program (Biology) was included in this table. Sadly, all 15 of these programs selected had underemployment rates greater than 50 percent, with Criminal Justice and Performing Arts the two highest at 65.8 percent and 63.9 percent. An explanation for Criminal Justice majors being in this grouping is that many police officers are hired without college degrees and are admitted into their respective city or state police academies assuming that they pass the physical fitness test. Criminal Justice program majors may be overqualified for those entry-level policy academy jobs.

Median Wage Early Career

Table 4A provides the 15 lowest (worst) median wage early-career program majors. Pharmacy and Biology were the only STEM programs included, and both require graduate degrees to achieve higher earnings. It was not surprising to see that education program majors were included in this group, as were social services and theology programs.

Table 4B provides the highest (best) 15 median wages by program major. All but two of these programs (Business Analytics and Economics) are STEM programs. All of the engineering programs, as well as nursing and construction services, are included.

Median Wage Mid-Career

Table 5A provides the lowest (worst) top 15 mid-career wages by program major. Six of the seven lowest programs are education programs (the other is social services). Many of these were included in the median wage early career low 15 as well. It’s no wonder that the average tenure of a new teacher is only five years. With early career and mid-career earnings in the bottom 20 percent of all program majors, a large percentage of education majors are opting for another career.

Table 5B provides the highest (best) Top 15 median wages for mid-career by program major. Only a couple of programs moved up into the Top 15. Almost all the engineering programs that were included in the top 15 for early-career remained in the top 15 for mid-career. Computer Science and Construction Services remained, and nursing dropped out of the top 15.

Percentage of Graduates Earning Graduate Degrees

Table 6A provides the 15 lowest program majors and the percentage of their graduates who go on to earn graduate degrees. At a glance, the inclusion of these 15 programs in this list is not surprising. Almost all of the careers that graduates of these specific programs match up with are careers that do not require a graduate degree to get ahead.

Table 6B provides a list of the top 15 programs with postgraduate degrees. None of these was a surprise. Some of these programs with a postgraduate degree lead to careers in medicine and law. Some professions may require a master’s degree to be licensed (Social Services).

Analysis of Overall Findings

The annual report of the New York Fed is oriented toward data presentation and not oriented toward detailed analysis and commentary. As I reviewed the data, I thought about ways in which I might be able to comment on the overall ranking of program majors using these five metrics.

I debated weighting the metrics equally versus differently. I decided to go with equal weighting because of the difficulty of accounting for the time value of money (early-career earnings versus mid-career earnings, as well as length of time unemployed or underemployed).

I opted to weight the highest/best outcome to 73 (there were 73 unique program majors in the dataset), sort the program majors from highest to lowest, and scale the numbers from 73 to 1.

I debated how to rank “Share with graduate degree”. I saw the merits of graduates pursuing graduate degrees from an academic and educated populace perspective. I opted to rank the programs with the lowest share of graduate degrees as the highest because those graduates did not have to spend time in graduate school or pay graduate school tuition.

Table 7 below shows the rankings of each category for all 73 program majors. The program majors are listed in order from the highest to the lowest aggregate points. Degrees oriented toward specific careers ranked highest. Economics was the only “traditional” liberal arts program included in the top 15 (top 20 percent), ranking 15th.

I highlighted traditional liberal arts programs in yellow. Most of them are below the median total points score of 177. If you look at the three lowest-scoring non-liberal arts programs (pharmacy, miscellaneous biological science, and biology), their percentages with graduate degrees are in the single digits (meaning that a large percentage have earned graduate degrees), which negatively affected their overall scores.

My Final Thoughts

The evidence that program major influences college graduates’ lifetime earnings is not new. The New York Fed’s report confirms that. At the same time, the rapid adoption of artificial intelligence (AI) into the workplace will impact recent college graduates. While the data looks gloomy now for most liberal arts graduates, the critical thinking skills they acquire during their education may be the ones employers actively seek.

Supply and demand also influence metrics like unemployment and underemployment. I can remember years when there were more newly minted engineers than relevant job openings for engineers. If all first-year college students consulted reports like the New York Fed’s and had the interest and academic ability to complete a STEM degree, the supply would change and likely outstrip the demand.

The World Economic Forum has predicted that tens of millions of jobs worldwide will be impacted by AI. That impact includes job incumbents and recent college graduates seeking employment. More and more, I’m convinced that schools and colleges should step up their career counseling and advising services. The more clarity they can provide students, the better prepared they will be as they navigate their program major choices. Colleges and universities should also consider tweaking their liberal arts programs, providing evidence of graduates acquiring critical thinking skills as well as evidence of graduates acquiring AI-utilization skills.

The post Earnings, Underemployment, and Unemployment Rates for Recent College Graduates appeared first on Wally Boston.

The feedback that I received was positive. As a liberal arts graduate, that shouldn’t have been surprising. I have a few friends with engineering backgrounds, but my liberal arts colleagues outnumber them. Many of them were also positive about the six recently published articles that I cited.

I deliberately used recent sources to support my liberal arts argument. I didn’t have to. For example, for more than 40 years, MIT computer science professor (and AI pioneer) Patrick Winston delivered a lecture titled How to Speak. In his introductory remarks, Professor Winston stated that a person’s success in life will be determined largely by their ability to speak, their ability to write, and the quality of their ideas, in that order. As Professor Winston formulated and refined his lecture over the years, AI was available only to a select few.

From Knowledge Scarcity to Knowledge Abundance

For most of modern higher education, the system rewarded those who knew more. Content was scarce, and access to information was limited. Faculty served as gatekeepers of knowledge. Thanks to AI, that world is gone. Those of us who have used AI know that it can produce essays in seconds, passable research summaries, working computer code, and presentation outlines.

The constraint is no longer access to knowledge. It is the ability to make sense of it. Individuals who will stand out are not those who can generate the most content, but those who can clarify, interpret, and communicate it effectively.

The Empowerment Promise

Professor Winston introduced a practical concept for speakers called the empowerment promise. The promise is a tactic where the speaker tells the audience what they will be able to do or understand by the end of the talk. Interestingly, this is a discipline of thinking, not just speaking.

Artificial intelligence can generate answers to almost any question posed to it. However, it is not good at answering philosophical questions such as “Why does this matter?” Thankfully, humans can answer questions like that. Winston emphasized simple but powerful techniques in his advice about speaking. The key ones were:

• Clear structure
• Repetition for emphasis
• Signaling transitions

These tactics outline a stylistic approach to crafting a speech or lecture and can be powerful cognitive tools.

In the age of AI-generated content, structure becomes even more valuable. Machines can generate information quickly, but they do not consistently prioritize what matters most, nor do they organize ideas for maximum clarity. People who can turn complexity into coherence become indispensable. The role of the educated individual is shifting from generator of knowledge to interpreter of knowledge.

Attention Has Become the Scarcest Resource

Professor Winston was known for insisting that his audiences avoid distractions. He asked that his students close their laptops and turn off their cell phones. His reasoning was straightforward: people have a limited capacity to process language, and divided attention diminishes understanding. Not only that, he pointed out that the presence of laptops also distracted him as the speaker and reduced his effectiveness.

Winston’s observations have become more relevant. In fact, many high schools across the country have recently banned cell phones in classrooms to improve learning quality. Today, we are not just competing with other speakers or writers. We are competing with:

• Algorithmically curated feeds
• Constant notifications
• An endless stream of AI-generated content

In today’s environment, delivering information is not enough. Communicators must find ways to earn attention and hold it long enough to make an impact.

The AI Paradox: More Content, Less Clarity

Artificial intelligence has created a paradox. We generate more content than ever before. But comprehension does not scale at the same rate. As volume increases:

• Attention fragments
• Depth decreases
• Meaningful understanding becomes harder to achieve

As many people know, Large Language Models (LLMs) are the technology behind AI tools like ChatGPT, Claude, Gemini, etc. The models work by ingesting billions of pieces of content, then training the model to search for an answer to the prompt and write its response using probabilistic algorithms. But what if the content generated is written in an outdated style?

The Misunderstanding at the Heart of Writing

Winston’s lecture reminded me of a lecture from Larry McEnerney, the Director of the University of Chicago Writing Center, that has been circulating online for years. Professor McEnerney maintained that most of us were taught, explicitly or implicitly, that writing is about expressing what we know. As students (K-20), we were rewarded for things like demonstrating knowledge, organizing ideas, and writing clearly structured essays.

Those of us who followed instructions well received good grades and achieved academic success. However, Dr. McEnerney posited in his lecture that there is a problem with this style of writing.

The traditional model of writing that most of us were taught only works in an environment where someone is required to read what you write.

In school, your teacher or your professor was required to read your paper. In the real world, no one is required to read your paper.

McEnerney made a powerful claim. Writing is not about expressing your ideas. It is about creating value for a reader. If your writing does not solve a problem, improve understanding, or help someone make a decision, then it has no value to the reader. If your writing has no value, it will not be read.

Smart People Often Struggle

McEnerney was direct in his observation that highly educated people are often the worst writers outside their domain. This isn’t because they lack intelligence, but because they were trained in the wrong model. Highly educated people learned to write for instructors, critical evaluation, and practical demonstration.

When these highly educated people inevitably encounter the real world, they often find themselves writing for starkly different and more diverse audiences, like busy professionals, decision-makers, and users who have no obligation to read anything.

McEnerney stated that this is not a gap between writing models but is a structural flaw in how colleges prepare students for the workforce.

Effective Writers Operate From a Different Starting Point

Effective writers do not ask “What do I want to say?” They ask, “What problem does my reader need solved?” This question leads to a totally different writing structure.

Traditional academic writing demands rigor, but structurally, it can be perfunctory. We expect background on the topic, detailed research of the relevant literature, and a polished thesis.

In contrast, real-world writing that’s bogged down by research and “inside baseball” language, you’re likely to lose attention quickly. Effective real-world writing demands problem-solving, an understanding of the stakes involved, and an explanation of how your ideas lead to a solution.

The shift from topic-centered writing to problem-centered writing is the difference between being read and being ignored.

Why This Matters More in the Age of AI

Until recently, poor writing could still find an audience. That is no longer the case. We live in a world where millions of pieces of content are generated every year. Last week, I reported that more than four million books were published in the United States in 2025. The marginal cost of producing “acceptable” writing is approaching zero. Artificial intelligence can already:

• Summarize information
• Organize ideas
• Produce grammatically correct prose

But AI struggles to:

• Identify a meaningful problem for a specific audience
• Establish urgency or stakes
• Deliver insight that changes how a reader thinks

In other words, AI can replicate the academic writing model remarkably well. AI struggles with the value-creation model. The primary reason for this is that Large Language Models (LLMs) create their probability-derived responses based on billions of ingested articles and books. Naturally, most of those articles and books were written in the old style.

The Emerging Divide

This creates a new divide, one that I believe will become increasingly visible in the workforce. There are two sides. On one side are individuals who can produce traditional content. On the other side are individuals who can create value through writing.

The first group will be augmented and, in many cases, replaced by AI because AI is trained to write like that group. The second group will become more valuable.

A Higher Education Problem We Can No Longer Ignore

From my perspective, this has direct implications for higher education. For decades, higher education has focused on teaching students what to know and evaluating their ability to explain it.

We have spent far less time teaching students how to communicate in environments where attention is scarce, how to write for audiences who are not obligated to listen, and how to create value through communication. More than a writing issue, this is a looming workforce readiness issue.

A Different Model for Teaching Writing

If we were to redesign writing instruction for today’s world, it might look like this:

1. Start with the reader, not the writer
2. Define a real problem that matters to that reader
3. Make the stakes explicit
4. Deliver insight that improves the reader’s position
5. Measure success by impact, not completion

This is not how most writing is taught today. As McEnerney pointed out, it is how most effective writing works in practice. Ironically, it is also not how most LLMs generate their articles about various topics.

A Personal Reflection

Over the course of my career as a healthcare executive, a university president, and now a board member and investor, I have read thousands of reports, proposals, articles, and books. The ones that stood out were not the most polished. They were the ones that:

• Framed a problem clearly
• Made me see something differently
• Helped me make a better decision

That is what effective writing does. Effective writing is becoming increasingly important in the age of AI. Given the vast proliferation of content that no individual can possibly keep up with, effectively written content will be more valued by humans who serve as expert moderators of knowledge.

Parting Advice

Patrick Winston’s lecture, which I used to lead off this article, was primarily about speaking. However, in his introductory remarks for his How to Speak lecture, he noted that a person’s success will be defined by their ability to speak, their ability to write, and the quality of their ideas in that order.

Professor Winston’s Empowerment Promise, which advises the speaker to tell the audience what they will be able to do or understand by the end of the talk, is eerily similar to Professor McEnerney’s maxim that if your writing does not solve a problem, improve understanding, or help someone make a decision, then it has no value to the reader.

Winston was a Computer Scientist at MIT who directed the MIT Artificial Intelligence Laboratory from 1972 to 1997. While LLMs hadn’t been invented at that point, it’s likely that Winston saw the wave of content and knowledge generation coming and wanted to emphasize the importance of individuals mastering human skills like effective speaking and writing.

Larry McEnerney earned a PhD in English but didn’t want to spend his career writing about literature. He joined the University of Chicago’s writing program and, five years later, became its director, a position he held for nearly three decades. While McEnerney had no background in computer science or AI, his focus on effective writing is excellent advice for anyone using and adapting AI-generated content in the workplace.

We are entering a period where the ability to generate content and ideas is no longer scarce. The ability to create value through speaking and through writing is. Having the critical thinking skills to evaluate content, especially AI-generated content, and to communicate about that content to your bosses, colleagues, and team members will be crucial to your career success. Education leaders should make sure that their students learn how to evaluate AI content and output, edit, and communicate the ideas generated effectively. McEnerney advocates for effective writing. Winston advocates for telling the audience what they will understand at the end of your lecture or speech. Both tactics lead to value, and that’s what matters the most if you are working to establish your brand, your authority as an expert, and your professional career.

The post Is Communication the Last Mile of Human Intelligence? appeared first on Wally Boston.