Was this forwarded to you by a friend? Sign up for the On Student Success newsletter and get your own copy of the news that matters sent to your inbox every week.

Before I jump into this week’s post, three things to be aware of.

Tomorrow I will be participating in a webinar sponsored by Mentor Collective that, as it happens touches on some of the themes I am writing about today. Check it out and maybe join us.

Next week I will be attending most of the Open edX conference happening in my backyard of Salt Lake City. Last year it was in Paris; this year it’s in the Paris of the Wasatch Valley. If you will be there, let me know. I also have a highly opinionated and non-comprehensive food guide to Salt Lake City, which I have linked to at the bottom of this post for anyone who is interested.

I will also be heading to Norway the following week and, again, the offer to meet and chat stands. I will alas be missing a catch-up with my regular and favorite Norwegian co-conspirator. But Roger won’t be able to escape me forever. On to the post.

Preparing students to thrive in the job market is an increasingly critical part of student success. Not only because it is the right thing to do, or because students increasingly judge the value of higher education in terms of its ability to prepare them for what comes later, but also because universities are increasingly being held accountable for students’ post-graduation earnings.

And the task facing students in the labor market is increasingly daunting.

Navigating entry into and flourishing within the job market is often framed as an information challenge. But in many ways students are awash in career information: labor market data, career centers and counselors, advice networks, and increasingly AI tools that help them identify opportunities and prepare for them.

Yet students still struggle to navigate careers and make decisions about their futures. One report after another this week circled the same underlying issue from different directions. Together, they convinced me that students increasingly suffer not from an information deficit, but from a translation deficit.

Career navigation as translation

A new report from Harvard’s Kennedy School Project on Workforce provides some fascinating insights into how students and workers think about and navigate careers. Its central argument is that careers are no longer linear. Instead, most people’s working lives are characterized by a constant need to adapt, pivot, and potentially exit and re-enter the workforce. The report treats career navigation as a problem of interpreting fragmented labor-market signals in an unstable economy.

What I found particularly interesting were the findings about how students and workers navigate this increasingly complex environment. The results reveal a growing divide between institutional authority and practical utility.

Traditional institutional sources remain trusted, but students increasingly turn elsewhere for guidance that feels more immediate, actionable, and personalized.

The contrast between traditional sources of career information — teachers, counselors, and career service centers — and newer, less formal sources was particularly striking. Traditional sources ranked higher on reliability than helpfulness, while networks, online tools, and social media ranked surprisingly high on helpfulness despite being viewed as less reliable.

What this chart suggests to me is that students and workers are finding value in:

• low-friction,

• ambient,

• conversational,

• always-on systems.

Students and workers increasingly appear willing to trade some degree of reliability for information that feels more immediate, personalized, conversational, and actionable.

The challenge is that the kinds of translation students increasingly prefer — low-friction, always-on, AI-mediated — may not be the kinds of translation best equipped to help students navigate uncertainty, identity, and long-term career development.

Many low-friction guidance systems are optimized for immediate decisions. Relational translation, by contrast, is often about helping students navigate uncertainty over time.

Which partly explains why students are still holding onto people and institutions that provide guidance. Students appear to be constructing hybrid guidance ecosystems made up of both traditional and non-traditional sources. But what is still missing is what the authors refer to as “interpretive infrastructure” to help people make sense of the labor market and navigate pivots, exits, and re-entries.

The role of interpretation, or translation is critical. As they explain:

Career readiness as curricular translation

While the Harvard report focuses on labor-market navigation, a report from Every Learner Everywhere (ELE), The Career Readiness Imperative in Gateway Courses, treats career readiness primarily as a curricular translation problem.

The ELE report is based on the premise that too few students engage with traditional career supports such as career centers or co-curricular activities (for example, career-oriented clubs). Institutions already teach students many skills and competencies that are enormously valuable in the workforce, including in the liberal arts. But what they have traditionally failed to do is help students recognize, narrate, articulate, and connect their learning to what employers need.

To address this, the report proposes embedding career readiness throughout the curriculum. Drawing on frameworks such as those from NACE and AAC&U, it emphasizes integrating career discussions and career-oriented content across courses, especially gateway courses such as required quantitative, writing, or introductory disciplinary courses.

Including career content in these kinds of courses means that students are exposed to career translation early and often, particularly when they may need it most. Faculty help provide students with ways of translating between the curriculum and the workplace.

I like this approach, especially in how it moves away from the narrow vocationalism that characterizes many efforts to incorporate career support into the curriculum. But it is only a partial solution.

Even apart from its deeply institutional focus, the model relies heavily on faculty to do the translation work. We have seen this become a problem in many other “across the curriculum” initiatives, where responsibility often gets pushed downward and informally assigned to the most junior faculty member, or to the one person in the department willing or interested enough to take it on.

But the Harvard report suggests the challenge is messier and far more ongoing. Translation is not simply a one-shot matter of helping students connect coursework to careers inside a class. Labor markets themselves are fragmented, unstable, and difficult to interpret. Students pivot, stop out, re-enter, relocate, and continually reassess opportunities and risks.

Translation in that context is not an episodic curricular intervention but an ongoing process of interpretation, navigation, and adaptation. In unstable labor markets, translation increasingly becomes lifelong rather than episodic work.

Mentorship as relational translation

What both reports never fully solve for me is who actually performs this translation over time, especially as students pivot, reassess, stop out, re-enter, and encounter changing labor markets. A third report, this time from Mentor Collective, partially answers that question by reframing translation as relational rather than merely informational. (In the interests of full disclosure, I am participating in a Mentor Collective webinar tomorrow — you should come — but my opinions, as always, are my own.)

Unlike many institutional supports, mentoring relationships are adaptive. Mentors can contextualize advice, personalize guidance, revise interpretations over time, and help students navigate ambiguity rather than simply deliver information. That matters in labor markets characterized by uncertainty, pivots, and unstable pathways.

The report describes this relational translation in several ways.

What is striking about these examples is that mentors are not simply providing career information. They are helping students interpret unfamiliar systems, decode professional expectations, imagine possible futures, and convert aspiration into action. Much of this work involves navigating tacit knowledge and hidden norms that are difficult to fully codify into websites, career modules, or competency frameworks.

Mentors are not only interpreters. They are also cultural brokers, navigators, and translators of institutional and professional norms. And they do this on an ongoing basis. In institutions where support systems are often fragmented and siloed, mentors can also function as continuity figures, helping students connect experiences and navigate across institutional boundaries.

Taken together, these reports suggest that translation is doing much more than simply conveying information about jobs. It is about explaining worlds, navigating systems, and helping students construct identities and possible futures.

Translation does more than convey information

Across these reports, it becomes clear that translation in career success — and the translation being done by mentors, faculty, advisors, and institutions — is performing at least six different kinds of work:

Cognitive work: Helping students make sense of complexity.

Interpretive work: Helping students connect signals to decisions.

Narrative work: Helping students tell coherent stories about themselves.

Emotional work: Reducing intimidation, uncertainty, and paralysis.

Social capital work: Helping students access norms, networks, and hidden rules.

Institutional work: Bridging the gap between formal institutional systems and lived student experience.

Higher education institutions have traditionally been organized around content delivery and information provision. But increasingly, student success depends not just on helping students acquire information, but on helping them construct coherent professional identities and navigate uncertain systems.

Increasingly, student success — especially in career terms — depends on interpretation, navigation, identity construction, contextualization, sense-making, and relationship-building.

What these reports collectively suggest is that translation itself is increasingly becoming institutional infrastructure. Not a supplemental service or optional co-curricular support, but a core institutional function.

Universities historically organized themselves around content delivery, credentialing, and information provision. But in labor markets characterized by uncertainty, pivots, and fragmented pathways, students increasingly need systems that help them interpret, contextualize, and navigate complexity over time. That is a very different institutional challenge.

It is also a huge shift, and AI complicates this further because it dramatically lowers the cost of information access while potentially increasing the need for interpretation and judgment. Higher education increasingly suffers not from an information deficit, but from a translation deficit.

But once translation becomes infrastructure, questions about power, norms, and inequality become unavoidable.

The challenges of translation

Translation may be increasingly critical, but it is important to recognize that it is not always an unalloyed good. Translation in any context is complex, contested, unequal, partial, and political. In student success and career navigation, translators inevitably carry assumptions about what success looks like, what kinds of careers matter, and how professionalism should be performed.

Translators carry their own biases and blind spots

Mentors, faculty, advisors, networks, and AI systems all interpret the world through partial experience. And social capital has a tendency to reproduce itself. The Harvard report hints at this when discussing networks that are narrow and self-reinforcing.

Mentors as translators may steer students toward familiar paths. Faculty may overvalue academic trajectories. Employers may privilege dominant norms. Networks often reproduce class and cultural assumptions.

Translation is never neutral.

Translation risks becoming assimilation

A great deal of translation work in career readiness implicitly involves:

• teaching students how to speak professional-managerial language,

• teaching them institutional norms,

• teaching them how to perform employability.

But this raises difficult questions: whose norms, whose professionalism, and whose communication styles?

This especially matters for:

• first-generation students,

• neurodivergent students,

• working-class students,

• students of color,

• international students.

Sometimes “translation” quietly becomes less about helping students navigate systems and more about helping them adapt to unequal systems rather than changing them.

When organizations outsource translation

Organizations increasingly rely on mentors to provide interpretive and developmental labor that formal systems either cannot or will not provide directly.

When I worked at Gartner, we were assigned mentors — usually from outside our immediate unit — to help us find our feet as analysts. My mentor Kristin was the best mentor imaginable, and she remains so to this day. In her words, “there is no endpoint to mentoring.”

At first, I thought the system was amazing — and to be fair, I got enormous value from it. But over time I also realized it partly allowed the company to absolve itself of the responsibility to provide more formal professional development and institutional support.

Gartner-style mentoring systems can be enormously valuable. But they should not become replacements for broader investments in training, development, and institutional support.

Translation as infrastructure

Higher education has spent years building information systems. But increasingly, student success may depend on building interpretive systems instead.

Those systems need to be intentionally developed and supported. They include:

• mentoring networks of all kinds,

• embedded advising,

• peer interpretation,

• labor market contextualization,

• alumni guidance,

• career translation infrastructure,

• adaptive coaching.

But building interpretive systems also requires understanding how translation actually functions within student success work. If translation is increasingly becoming infrastructure for student success, then institutions need to think much more carefully about who does that translation, whose norms shape it, and who gets left out.

A Highly Opinionated Guide to Eating in Salt Lake City

Morgan’s completely quirky, personal and incomplete food guide to Salt Lake City, for those attending Open edX or just visiting.

The main On Student Success newsletter is free to share in part or in whole. All we ask is attribution.

Thanks for being a subscriber.

Was this forwarded to you by a friend? Sign up for the On Student Success newsletter and get your own copy of the news that matters sent to your inbox every week.

Somehow it’s the end of April already. Time for a few charts that have stuck with me this month, and what they might mean for student success.

Humor is harder than it looks

I use humor a lot in presentations (which I am pretty great at, and available for). So I know how hard it is.

So I wasn’t surprised to read that in a study done on humor in scientific conference presentations a small minority of efforts at humor landed well enough to get a laugh.

The cost of college, measured in time

We’ve all been saying this for years. Now we have the numbers. Current students would have to work so much longer to cover the costs of annual tuition. At a public institution in the US, for in-state tuition, they would need to work 26 hours per week. At a private institution, it is 102 hours a week.

Employment while at university has morphed into a structural barrier to completion rather than a pathway to affordability.

From cost pressures to research quality

How much of this should we trust?

The reproducibility of research results is a problem in many fields. But I was more than a little dismayed to see the poor showing of education on this chart on the reproducibility of research in the social sciences.

We should probably do fewer randomized controlled trials (RCTs) in education anyway, which would at least reduce the number of fragile findings. But if we can’t reliably reproduce results in education research, we should be cautious about building entire student success interventions on them.

Florence Nightingale & W.E.B. Du Bois, data nerds

I love a good data visualization (though I limit myself to admiring them rather than producing them). I recently came across a post sharing some great historical visualizations I hadn’t ever seen before.

I knew that Florence Nightingale was a big data geek and was reminded of this by her depiction of causes of death among British soldiers in the Crimean War.

But I didn’t know that W.E.B. Du Bois was also a superb visualizer of data. Du Bois had studied under William James at Harvard.

Part of me is cringing at the concept of an “American Negro Exhibit.” But the visualizations are delightful and more of them are available at the Library of Congress.

This is what great data visualization does: it doesn’t just show data, it makes an argument.

Students, as always

First, a disclosure: my spouse works at the University of Utah Library.

I have been having a rough time with all the national news and negativity. Unsurprisingly, students provide a great antidote.

Yesterday this Instagram reel from the University of Utah library won an award from the Association for Research Libraries. It’s called Whisper Something Good.

Instagram post

The internship bottleneck

Brandon Busteed shared data from the National Association of Colleges and Employers (NACE) showing the rate at which interns convert to full-time employees. Internships are hugely important and there aren’t enough of them for the number of students seeking them.

We’ve quietly made internships a gatekeeper to opportunity, but without building enough capacity to support that expectation. It has also become an equity problem.

When the metric misses the point

Hannah Ritchie (whose writing on data I adore) recently wrote about national rates of train electrification, and why they didn’t quite make sense.

The numbers didn’t feel right to her, so she had a closer look. It turns out the data used in this and other similar charts are based on the proportion of the railway lines that are electrified, with no reference to how busy the lines may be.

She reran the numbers looking at train traffic, but not passenger numbers, and got some quite different results.

This is exactly the kind of measurement problem we run into constantly in student success.

AI and the new divide

It wouldn’t be a student success charts post without a John Burn-Murdoch chart. This one has been doing the rounds this week, understandably so given the sharp relationships it shows between earnings and AI use.

I keep using AI more and more in the hope it shows up in my earnings. So far, no luck. I guess the causality runs the other way. But if that ever happens, one of the things I would do is pay for an actual subscription to the Financial Times so that I could get more John Burn-Murdoch charts.

But this chart is a useful reminder: AI adoption is not evenly distributed. It’s clustering among higher earners, which risks widening gaps rather than closing them.

Musical coda

Great songwriting and great data visualization do the same thing: they convey a lot with very little.

“Busted flat in Baton Rouge, waiting for a train.”

Nine words. Character, situation, mood.

That’s the bar.

H/T Steven Mintz

The main On Student Success newsletter is free to share in part or in whole. All we ask is attribution.

Thanks for being a subscriber.

Was this forwarded to you by a friend? Sign up for the On Student Success newsletter and get your own copy of the news that matters sent to your inbox every week.

It has been a hellish week of allergies, exhaustion, and frustration with datasets, balanced by a birthday celebration in the OSS household. But apart from that, what is the news in student success?

This week’s readings all point to a familiar pattern: we keep building better ways to pull students out of the river, but we are far less willing to redesign the system that keeps pushing them in.

It’s a design problem, again

The Trellis Strategies Some College, No Credential (SCNC) Survey Report describes a landscape familiar to many of us. There are 43.1 million people with some college but no credential, including 37.6 million working-age adults. Trellis’s large survey of former undergraduate students produces a set of insights that are both encouraging and frustrating.

On the encouraging side, a majority believe that higher education is a good investment, and nearly 73% say that returning to earn a credential would improve their career and earning potential. Sixty-three percent of respondents report intending to return to finish their studies.

On the frustrating side, the report presents a familiar list of reasons why students stop out: family responsibilities, finances, health issues, and employment demands.

The report reinforces something we have seen repeatedly: students are not leaving because they cannot succeed academically. They are leaving because higher education is structured around assumptions of time and financial stability that no longer hold.

I found some of the differences between two- and four-year schools striking. On one level, the greater reported financial stress at four-year institutions makes some sense, given the often lower cost of two-year schools. Differences in the role of campus life also make sense, since more students at two-year institutions are commuters and may have less time for campus engagement. But the divergence in other areas—particularly the much higher share of students at four-year institutions citing health and academic reasons for leaving—is cause for concern.

But I was mostly frustrated by the implicit (and sometimes explicit) assumption that institutions can solve these issues simply by identifying which barriers affect their students and layering on more services. That assumption becomes even clearer when you look at the toolkit.

The recommended actions are largely reactive. For example, the report states that 71% of SCNC students never spoke to anyone before stopping out. The toolkit responds with recommendations around referral channels and scripts for advisors. Similarly, it recommends instituting exit surveys, tracking students who leave, and monitoring payment plans. These are symptom-detection mechanisms, not system redesign. Institutions are not redesigning the system to accommodate student reality; they are building increasingly complex processes to manage the consequences of a system that does not fit that reality.

This is less a critique of Trellis than a reflection of how the sector as a whole approaches student success. It is less a student problem than a system design problem. The result is a system that becomes increasingly complex without becoming more effective.

Managing around constraints

Reading the Trellis report, I kept thinking about this video of a man going to extraordinary lengths to park in an impossibly narrow garage.

Students are optimizing for reality

That mismatch between system design and student reality doesn’t just show up in retention. It may also explain another trend. The National Student Clearinghouse Research Center recently released its Undergraduate Degree Earners Report. Some coverage focused on the fact that, for the first time, students ages 18–20 comprise the largest share of first-time associate degree earners. But what caught my eye was the increase in students earning certificates rather than associate (two-year) or bachelor’s degrees.

Certificate completers are up 5.7% year over year, compared to more modest gains for bachelor’s degrees (2.8%) and associate degrees (2.6%).

This year’s increase in certificate completers is less dramatic than last year’s—5.6% compared to 16.0%. But over the two-year period, certificates have grown by more than 21%, while bachelor’s and associate degrees have seen much more modest growth in the 2%–4% range. This may be less about increased demand for certificates and more about students seeking faster, more flexible pathways that better fit their lives.

Same problem, same playbook

That same pattern shows up in institutional responses as well. A news article in The Philadelphia Inquirer paints a picture of significant student success challenges at Temple. Based on an internal university report obtained by the newspaper (and not shared with readers), the university is experiencing notable declines in retention—not only from freshman to sophomore year, but also from fall to spring. The consequences are clear in terms of students not staying enrolled, but the report also frames the issue as a substantial loss of revenue.

The news report does not give us a lot of information about why the drop in retention happened. Officials attribute the decline in enrollment to financial pressure as well as some academic challenges. In seeking to address the challenges Temple is following a tried and true playbook, according to the President John Fry.

The response is familiar: more tools, more tracking, more interventions. But note what is missing—any indication that the underlying structure might be part of the problem.

The story also surfaced something else that is easy to overlook but deeply related. The report included a study by NACUBO showing that the majority of graduates are concentrated in a small number of programs (the Pareto principle at work). At first glance, this might appear to have little to do with the retention challenge.

This raises an uncomfortable question: how much of the curriculum actually supports student pathways, and how much exists because of internal structures and incentives? Again, the response is to manage symptoms rather than redesign the system.

It’s a design problem—assessment edition

This question about what we are actually designed to deliver shows up in another domain as well: assessment. There has been no shortage of bad takes on higher ed this week, but there was at least one strong piece from Mark Bassett, a faculty member at Charles Sturt University (which I always think of as the Johns Hopkins of Australia—always one letter away from predictability).

Bassett writes about the impact of AI on assessment. Many of us have been stumbling toward the conclusions he reaches for some time, but he frames the issue in an especially cogent way, using Stephen Covey’s circles—control, influence, and concern.

I am going to quote from the piece more than I usually would, because he frames the issue so well. He begins by arguing that GenAI has thrown the question of assessment into sharp relief, laying bare a set of assumptions and weaknesses that were always there. The emphases are mine.

But GenAI also raises questions about professional identity that are embedded in those practices and the role assessment plays in higher education.

All faculty face these challenges in assessment from GenAI, but how they respond depends on whether they operate within the circle of concern or the circle of control. Those in the concern camp see the impact of GenAI as something whose root causes they cannot influence. They respond by emphasizing detection and punishment.

For faculty or institutions in the control camp, Gen AI is seen as something where the root causes of impact can be understood and action taken.

It is not a perfect piece. It does not address how to solve these problems. But as an analysis, it surfaces issues I haven’t seen framed this clearly before and will, I believe, ultimately help us address one of the core challenges facing higher ed today.

What GenAI exposes in assessment is the same problem we see everywhere else: we have built systems that assume stability, and we are trying to patch them in a world where that assumption no longer holds.

I started with a quote from Desmond Tutu, whom I deeply admired. I have a funny story about the time I got to meet him. For the cost of a fake beer or a real coffee, I will share it with you the next time we meet.

Musical coda

To the firecracker, on the occasion of her birthday.

If you know someone who spends a lot of time pulling students out of the river, feel free to send this newsletter along to them. Or better yet, someone who might be willing to walk upstream.

Thanks for being a subscriber.

Was this forwarded to you by a friend? Sign up for the On Student Success newsletter and get your own copy of the news that matters sent to your inbox every week.

EdTech has a recurring habit of announcing the next transformative breakthrough with enormous fanfare and little introspection. A new technology. A new model. A new price point that promises to reshape higher education, reduce costs, and finally achieve scale.

Often, these innovations are not entirely new. They are familiar ideas repackaged with updated language and some contemporary technology. But they tend to rest on a persistent set of misunderstandings — about how students actually engage, how learning unfolds, and how difficult scale really is.

The recent sequence of events involving Sal Khan offers a particularly instructive example. First came the quiet recalibration around Khanmigo, the AI tutoring tool that was initially framed as transformative but appears to have seen limited sustained use. Almost immediately afterward came the announcement of a new joint venture with ETS and TED to launch a low-cost online degree. The new venture was touted as.

Individually, neither development is remarkable. But together, they illustrate something more revealing: the rapid pivot from one ambitious claim to another, without fully reckoning with the structural constraints that limited the first.

The 5% problem

The first signal of this latest cycle was the Chalkbeat article revisiting the performance of Khanmigo, Khan Academy’s AI tutoring tool. In it, Sal Khan appeared to temper some of his earlier claims about how transformative AI tutoring would be. The article noted that student uptake had been limited and that sustained usage was concentrated among a relatively small subset of students.

The piece quickly spawned a familiar secondary wave of commentary. But the most interesting detail in the article was a familiar one. Only a small fraction of students regularly used Khanmigo; roughly 5 percent in some implementations.

Khan and other company representatives expressed frustration at students’ lack of engagement. The students weren’t using it “correctly.” They weren’t self-initiating.

This is where the pattern reasserts itself.

When only 5 percent of students engage with a voluntary educational tool, that is not primarily a student problem. It is a design and integration problem. The hardest challenge in EdTech is not building a capable tool. It is embedding that tool into systems that support motivation, accountability, and sustained use — especially for students who are juggling many demands and uneven academic preparation.

Teaching the already motivated 5 percent is not transformation. It is amplification. Which makes the rapid pivot — from limited engagement with an AI tutor to the announcement of a fully AI-mediated, $10,000 degree — especially striking.

From tutor to degree

Within a week of tempering expectations about Khanmigo, Sal Khan pivoted to something new, this time in higher education. He announced a collaboration with testing giant ETS and TED (of TED Talks fame) to create an accredited university offering low-cost degrees.

The rhetoric is expansive. The ambition is unmistakable.

The details are still emerging. But the scale aspirations are clear: tens of thousands of students, rivaling flagship state universities.

Blue-chip corporate names lend credibility, and many of these same companies are already involved in creating and offering certificates and content on other platforms as well as their own.

None of this is unreasonable. But none of it is new. And none of it resolves the core challenges that have faced other “revolutionary” online low-cost degrees: sustained engagement, student persistence, and the economics of large-scale asynchronous delivery.

If voluntary AI tutoring struggled to engage more than a small fraction of students, what exactly changes when the same logic is applied to a fully asynchronous degree? To understand why this matters, it helps to remember what happened the last time we declared scale solved.

Reinventing what already exists

The Khan TED Institute is presented as a reimagining of higher education for the AI age. Strip away the rhetoric, however, and the underlying structure is familiar: asynchronous lessons, simulations, peer dialogue, remote faculty guidance.

Fully online higher education is not an untested frontier. Nearly 30 percent of U.S. higher education enrollments now include online coursework. Entire institutions operate at scale in fully asynchronous formats.

And yet the proposed design is framed as if it marks a departure:

There is nothing radical here. Virtual lessons, asynchronous assignments, group projects, simulations, peer dialogue sessions— essentially discussion boards or tutorials —have been core elements of online pedagogy for two decades. They are not breakthroughs. They are baseline design choices.

This pattern is familiar. A decade and a half ago, MOOCs promised to democratize higher education with nearly identical rhetoric: lower cost, global reach, scalable delivery, brand-name partners.

The MOOC model evolved into online degrees, many priced below traditional programs and delivered through platforms like Coursera — the most serious and well-resourced attempt to operationalize this approach at scale. Coursera combines brand-name university partners, a massive global learner base, and what Phil Hill has called an enrollment “flywheel.” And yet, even with those advantages, the degree business has proven harder than early projections suggested, as recent earnings reports make clear.

Recent data show that degree enrollments have plateaued and revenue per degree has declined. Coursera has increasingly shifted emphasis toward shorter credentials and enterprise offerings rather than continued degree expansion. This is the best-resourced version of the model. If scale were easy, it would be visible here.

The lesson is not that online degrees cannot succeed. It is that scale requires sustained marketing investment, institutional credibility, student support infrastructure, and retention strategies that go well beyond content delivery.

The proposed Khan TED Institute would enter a more crowded and mature market than Coursera did in 2017. It would do so without an existing institutional brand, and with an undergraduate target population that is typically more brand-sensitive and retention-challenged than graduate learners. A $10,000 price point is rhetorically powerful. But price alone does not solve acquisition costs, persistence challenges, or the economics of sustained student support.

What makes this different from earlier Khan Academy expansions is that the degree market is not a philanthropic distribution problem. It is a competitive acquisition problem. In the past, Khan Academy and Khanmigo benefited from large institutional adoptions, foundation support, and government partnerships. They did not have to fight for individual tuition-paying undergraduates in a saturated, brand-sensitive market with high marketing costs and low retention margins.

Competing in the undergraduate degree market requires sustained marketing investment, enrollment operations, student support infrastructure, and regulatory compliance — not simply compelling rhetoric and strong corporate brand associations.

Scale is not frictionless

The proposed curriculum for the new Khan TED Institute is fairly predictable and not unreasonable, if a bit STEM and jargon heavy.

On paper, this looks coherent. It blends some liberal arts with technical fluency, and applied collaboration. But what is striking is how frictionless the model sounds.

Content can be delivered asynchronously. Skills can be practiced in simulations. Dialogue sessions can be scheduled. But learning — especially for under-prepared or time-constrained students — is not a smooth pipeline from exposure to mastery. Learning requires feedback loops, accountability, structured practice, and often sustained human intervention. AI can assist with parts of this. It does not replace the systems and practices that make engagement durable.

If a voluntary tutoring tool struggled to move beyond a small, self-motivated minority, it is reasonable to ask what mechanisms will ensure persistence in a largely asynchronous degree, especially when the underlying logic still seems to be that access to a tool or content is enough and that students themselves bear responsibility for engagement.

This fact comes out in the recent Chalkbeat article from Sal Khan.

And from Kristen DiCerbo the Chief Learning Officer.

And yet the need for deliberate and designed engagement is even more important in an online degree than in a tutoring app. The students most drawn to low-cost, asynchronous degrees are often balancing work, care-giving, and uneven academic preparation. Designing for their success requires more than access to content and AI tools. It requires structured support systems, good pedagogy and instructional design, integration into institutional processes, and sustained human accountability. And nothing in the current framing suggests that these structural engagement challenges are being treated as primary design requirements rather than downstream engineering problems.

The interval is shrinking

The debate over Khanmigo also prompted a sharp observation from one of the most perceptive critics of technology-mediated learning at scale. Before the announcement of the Khan TED Institute, Justin Reich proposed what he called the “time-to-TED-talk-renunciation” metric — the interval between a bold claim about technological transformation and the subsequent retreat to a more modest position.

Reich’s framing is humorous. But it captures something real. Each new technological promise arrives with expansive rhetoric. Then implementation collides with student behavior, institutional inertia, and economic reality. The recalibration follows.

Reich even sketched a trend line.

The interval appears to be shrinking.

And then, within days of Reich posting that chart, Sal Khan announced the Khan TED Institute — from the stage of a TED talk.

The humor works because the pattern is familiar. But the stakes are not trivial. Each cycle absorbs institutional attention, philanthropic dollars, public imagination — and some students who will enroll in the new experiment and struggle.

The deeper issue is not ambition. It is repetition. We keep repackaging familiar models without grappling with the structural constraints that limited them in the first place. In this case, that constraint is engagement — how to design for sustained participation among students who are balancing work, caregiving, uneven preparation, and financial risk.

The interval between declaration and re-calibration may be shrinking. The underlying mistakes remain unchanged — and they will produce the same outcomes unless the design logic changes.

If this was helpful or interesting to you, forward this post to others who may be interested. All we ask is attribution.

Thanks for being a subscriber.

 

Was this forwarded to you by a friend? Sign up for the On Student Success newsletter and get your own copy of the news that matters sent to your inbox every week.

So I may have spent most of yesterday trying to construct a student success version of the NCAA women’s basketball tournament after being annoyed by an alternative bracket someone shared on LinkedIn (ok, if you insist, it was this one). I’m not going to share my version with you because I ultimately learned that coming up with good measures is really hard—something you would think I would know by now.

The things we count (and the things we don’t)

I love Sketchplanations, but this week I really love Sketchplanations. Jono Hey has a wonderful graphic on the McNamara Fallacy.

He goes on to cover some related measurement challenges, such as Goodhart’s Law (which I thought about on an hourly basis during my last couple of years at Gartner), Campbell’s Law, and the Cobra Effect.

We wouldn’t know anything about that in student success, now would we? No, never.

And as it turns out, this week offered several reminders of what happens when we measure what’s easy rather than what matters. One place this shows up very clearly is in how we define completion.

The students who almost finished—and why that matters

One of the things I tried to use to construct my scores for the alternative bracket was completion. Oddly enough, one of the things I read this week was a fascinating report from About Texas 2036 on the need to grant recognition to students for credits they have already earned.

The central thrust of the report is that by 2031, 61% of jobs in Texas will require some form of post-secondary education, but the state is falling well short of that number. Yet a hidden opportunity exists: between 2012 and 2022, more than 54,000 Texans earned 60+ credit hours but left higher education without a credential. The report dubs these students “Potential Completers.”

But I found two things about the report striking. The first is the scale at which we are losing students from the post-secondary system—and how late we are losing them. Almost half (47%) of those who entered the post-secondary system in Texas completed either a bachelor’s degree, an associate degree, or a certificate, which is less than ideal but not as bad as I might have predicted. But tens of thousands left after earning 60+ credits—i.e., really close to completion. As the report argues:

This does beg the question of why these students are falling through the cracks—and whether the factors causing them to stop out are different from those affecting the almost half of students who leave earlier.

The second thing that struck me in the report was the shift in how we talk about these students. Instead of framing them as students who didn’t finish, or even as failures, there is a move toward thinking of them as near-completers whose progress has gone unrecognized.

That is a subtle but powerful move, and one I hope we see more of. It’s also a reminder that sometimes the problem isn’t that students aren’t succeeding. It’s that we’re not very good at recognizing the ways they already have. But measurement isn’t just about credentials. It’s also about how we understand students’ day-to-day lives.

This is not what “affording college” looks like

Trellis Strategies just released their Student Financial Wellness Survey report, and it makes for some interesting—and at times uncomfortable—reading. On one level, the report can be read as a description of the financial hardships students face. But it quickly becomes clear that this isn’t about students being stretched. It’s about structural instability and students living on the edge.

• 54% can’t access $500

• 65% ran out of money

• 54% face basic needs insecurity

This is not exactly the financial cushion most of us would hope for.

On one level, this reads like financial stress. But it quickly becomes clear that this isn’t about students being stretched. It’s about structural instability.

Financial instability is the baseline environment in which college happens, and it affects a majority of students in both sectors. And it’s clear that this financial stress is directly interfering with learning. Financial and financially related disruptions are widespread among students responding to this survey, with around a fifth reporting that they have had to miss class because of a problem.

Of the students experiencing financial difficulties, nearly half said their financial stress made it hard to concentrate on their coursework. This isn’t just about whether students can enroll. It’s about whether they can actually learn once they get there. None of these are unsolvable problems. But they do require us to design around students’ actual lives rather than the ones we imagine they have.

What’s also striking is that students are financing college with consumer financial tools, including credit cards, buy-now-pay-later services, and payday loans.

Which, when you think about it, makes a certain kind of sense. If your financial aid doesn’t cover your rent or your groceries, you don’t have a lot of options. As coping strategies go, it’s a pretty rational one.

What is also clear is that the mental health challenges we keep talking about are not separate from these issues, but are embedded in the financial and structural challenges students face.

Using validated scales to measure student well-being, the survey found that:

Increasingly, we see a chain of conditions: financial stress → time pressure → isolation → mental health → academic disruption. And it is a chain, not a set of separate issues.

But the most surprising thing in the report is that, despite all of this, students remain remarkably optimistic about higher education. A large majority still believe it is a good investment. Even more believe it will lead to a better quality of life. That matters. It means the core value proposition hasn’t collapsed. We haven’t lost them.

But it also raises a harder question: students’ continuing belief in higher education, alongside their debt anxiety, instability, missed classes, and depression, shows how decoupled their beliefs have become from their actual experience of college.

The report doesn’t offer easy solutions. But it does point in a direction. Financial wellbeing needs to be a bigger part of student success, and if we are serious about lowering the cost of college, we also need to pay attention to the ancillary costs that shape students’ day-to-day lives. It also suggests that improving student success may require spending a little less time refining our measures and a little more time paying attention to how students are actually living.

We spend a lot of time trying to measure and manage student success. But weeks like this are a useful reminder that students are navigating systems that don’t always line up with their lives. And when that happens, they adapt.

What happens when systems don’t fit

I’m going to leave you as I started, talking about the NCAA college basketball championships. There is a growing issue with the transfer portal in U.S. college sports, and no, I don’t think an executive order forcing students to stay put is either serious or likely to succeed.

But I am struck by the impact of the urge to transfer to a new program at the end of a season on college athletes. For example, just two days after the end of the season, 557 women’s Division I basketball players had entered the transfer portal, and most will be playing for new teams next year. That number will only increase over the next few days.

For a not inconsiderable number of players, this is not the first time they have entered the portal, as this example shows.

The vast majority of these players will never play professionally. What havoc is all this movement having on their careers and intellectual development, especially given how hard we make transfer when they do try to settle somewhere? It does make you wonder what this level of churn is doing to students’ academic lives.

And I will get off my soapbox by sharing the buzzer-beater that ended LSU’s season.

Musical coda

And thinking about measurement—and the growing distance between B1G campuses—here’s a very old rendition of Peter, Paul and Mary’s “500 Miles.”

If you found this useful, consider forwarding it to your teammates. If not, I will assume I am measuring the wrong thing.

Thanks for being a subscriber.

Was this forwarded to you by a friend? Sign up for the On Student Success newsletter and get your own copy of the news that matters sent to your inbox every week.

This Very Good Boy turns 13 today.

But what happened this week in student success?

A number of things I read this week point to a similar underlying issue. Across very different domains—and in quite different ways—AI and workforce credentials, as well as career services and the use of adjuncts, reveal the same pattern: systems that optimize for short-term performance, efficiency, or provision rather than long-term learning, mobility, and effectiveness.

When performance isn’t learning

A new research report from Jason Lodge and Leslie Loble, part of the Australian Network for Quality Digital Education project, throws into stark relief the dangers of unwise AI use. Although the report is technically aimed at K–12 education, it is highly applicable to higher education. Lodge and Loble describe how AI, when used poorly, can undermine student learning.

This outsourcing of knowledge, or cognitive offloading, is increasingly being identified as a risk in the use of AI in learning, but this report advances our understanding of the issue.

The authors usefully distinguish between beneficial offloading—where AI is used to help with something peripheral to a task (for example, checking grammar)—and detrimental offloading, “when a learner uses AI to bypass this intrinsic cognitive effort (the desirable difficulties) required to build long-term knowledge schemas.”

Their argument relies heavily on the idea of a performance paradox, which I think is central to the challenge AI poses to learning.

This is a key observation. AI can be used to improve short-term performance while harming long-term learning. A core problem in higher education—especially in conventional methods of assessment—is that we are optimized to measure and reward performance rather than learning. This problem predates AI, but AI exposes it and makes it something we can no longer ignore. AI doesn’t create this problem. It exposes it.

A second insight from the report is that one consequence of cognitive offloading is that students never fully engage with domain knowledge. The authors usefully push back against the idea that skills like critical thinking can be decoupled from domain knowledge.

They further argue that in the absence of domain knowledge, problems arising from cognitive offloading such as the performance paradox become that much worse.

This is a compelling argument against the claim—sometimes made (and which I won’t link to)—that the best use of AI is in teaching Gen Ed courses.

Finally, their identification of the real “cheating” risk is spot on. The issue is not AI doing the homework, but cognitive offloading itself—something far more difficult to detect and address.

Early wins, long-term tradeoffs

An important report from the Fordham Institute examines the impact of industry credentials earned by high school students in Ohio. These credentials have grown immensely in popularity in recent years.

In the report, the author examines the longer-term effects—seven years after high school—of earning an industry credential, drawing on data from 1.3 million students who entered high school in Ohio between 2011–12 and 2019–20.

The top-line findings are striking.

Students who earn industry credentials are more likely to finish high school (87% versus 81%) but far less likely to enroll in formal post-secondary education (40.6% versus 49.3%).

Credential earners also earn substantially more—about 21%—than non-credential earners immediately after high school, but this advantage fades over time. By their mid-twenties (seven years after graduation), the premium has shrunk to just 5%. As the report notes, this points to a very short-term benefit.

Not all credentials are created equal. There are vast differences in the wage return of credentials in different sectors.

The wage gains not only depend on sectors, but are highly dependent on gender as well, though this is likely a function of the sector in which the credential was obtained.

Even setting aside those gender differences, the data point to the risks of focusing too heavily on getting students into jobs and optimizing for earnings too early. Short-term gains may be short-lived. We are systematically steering students toward decisions that look successful early—and may limit their long-term options. These credentials are less about building pathways than replacing them.

The research also leaves a number of questions unanswered. The finding that earning a high school credential is associated with lower college or university attendance is critical. But it remains an argument about correlation. The author controls for demographics, prior achievement, and course-taking patterns, comparing students within the same schools and cohorts.

Students who pursue credentials may have always been less interested in—and less likely to attend—college, and the author explicitly acknowledges this. He also suggests that credential pathways themselves may “nudge” students toward immediate work, but stops short of unpacking what that might mean or how it operates. Even here, we see the same issue: we are quick to interpret outcomes as effects when they may reflect underlying choices and structures. Students are responding rationally to the signals we give them. The problem is that those signals are often misaligned with long-term success.

Seen together, these examples point to a deeper pattern. We are not just measuring the wrong things—we are structuring systems around those measurements.

Is your university financially sustainable?

This is a bit of an inside joke, as it relies on some knowledge of what happened at the University of Dundee last year. I wrote about the report authored by Pamela Gillies, which dissected what happened there, over at On EdTech.

The report makes for fascinating, if grim, reading. I would strongly recommend it to anyone as obsessed with interested in university politics as I am. It outlines how, although there were some external stressors (such as declining international enrollments), the real cause of the problems lay in the university’s culture, including a lack of tolerance for dissent or differing opinions.

What stands out is not just the financial mismanagement, but the absence of challenge. Systems optimized for smooth operation and consensus often suppress the very signals that would reveal underlying problems.

Against this backdrop, someone writing under the nom de plume Prof Serious has created a quiz to help people determine whether their institution is financially sustainable. We could probably come up with an equivalent scoring system for the U.S.—I’m open to suggestions.

A design problem, not a motivation problem

Although students increasingly identify getting a job as a key goal, yet few of them use the career center.

Students want career outcomes, but they don’t use career services. That’s not a contradiction—it’s a design failure. We are offering support at the wrong time, in the wrong format, and often in the wrong place.

The systems we don’t talk about

A growing share of teaching in U.S. higher education is done by adjuncts. Much of this is wonderful: inspired instruction from people with a real passion for their subject, many of whom work in the field or bring a valuable real-world perspective to their teaching.

But it is also a real challenge. Many adjuncts are low-paid and must string together multiple teaching roles to make ends meet. They can be only loosely connected to the institution and, as a result, less likely to be integrated into student success structures and practices.

The proportion of adjuncts teaching students also varies significantly by institution type, as a new CUPA-HR report makes clear. Adjuncts teach at two-year institutions at nearly twice the rate as at doctoral institutions.

It also varies a great deal by topic. Most adjuncts are concentrated in the humanities, security/protective services, multidisciplinary studies and education.

This both intensifies the need for—and complicates—the delivery of student supports in these institutions and in the areas they serve. We design student success systems as if institutions are stable, cohesive environments, when in reality large portions of instruction are delivered by people structurally disconnected from those systems.

Across all of these examples, the pattern is similar. We focus on what is easiest to measure—performance, early earnings, initial placement. But these are not the same as learning, long-term mobility, or meaningful outcomes. We are confusing what is visible now with what matters over time.

Musical coda

I love this video (and this song).

In keeping with this week’s theme, this post is optimized for short-term performance—specifically, how widely it gets forwarded. Please don’t disappoint.

Thanks for being a subscriber.

 

As always, a pattern runs through this month’s charts. Student behaviors and needs are changing. Markets are changing. But the way we design and measure student success is not keeping up.

Students Are Not Who We Think They Are

I keep turning this chart from John Burn-Murdoch over in my head. It would be easy to dismiss this as a story about youth. But even at the upper end of that range, these are not just “young people.”

The shift is clear: younger adults report lower levels of persistence and follow-through, and higher levels of distraction and carelessness. That has direct implications for how students navigate college and what it takes to succeed.

According to Brandon Busteed, demand for internships is rising sharply even as supply is falling.

Adapted from Brandon Busteed

These are not small shifts. They fundamentally change what it takes for students to succeed and quietly invalidate many of the assumptions built into our advising models, course structures, and support systems.

And even if students themselves were not changing, the environment they are entering would still make success harder.

The Ground Is Shifting Beneath Them

The world our students are entering is becoming more competitive, more constrained, and more uneven. Some of this is good, much of it is not. But all of these changes create challenges.

Ranking institutions is, at best, an imperfect science. But the improvement in STEM rankings by Asian institutions over the past decade is striking.

Another great graphic from John Burn-Murdoch (anyone else want to join me in the fan club?).

If these numbers hold, this does not augur well for the affordability of higher education institutions in London in the future.

Seen this way, the outlook for UK higher education is increasingly fragile.

In the US, some institutions are doing better than others at enrollment. For many this is not news, but this graphic from Stevens Strategy is striking

In the day to day it can be hard to remember that US college costs have actually been decreasing. And, at least at highly selective private institutions, that decrease varies quite a lot by parents’ income.

And yet, despite all of this, our systems are not adapting in response to either of these shifts.

Designed for the “Average” Student

According to the New York Times.

The reason: car safety features are designed around crash test dummies modeled on young to middle age men weighing 170 lbs.

Adapted from the New York Times

This is what happens when you design for a “default” user that doesn’t reflect reality. We do the same thing in student success. We often establish baselines around a specific type of student. These averages don’t just “fail to work” for many students. They actively mislead us about where the real problems are and who is most at risk.

And when we measure, we often confuse precision with accuracy. Both are important concepts in data analytics and student success, but in many cases we are becoming more precise about the wrong things. From the incomparable Jono Hey.

Put these together, and the picture is clear. Students are changing. The environment is changing. But we are still designing systems and measures for a world that no longer exists.

We are often precise about the wrong things and less accurate about what matters. And then we call for more data.

The main On Student Success newsletter is free to share in part or in whole. All we ask is attribution.

Thanks for being a subscriber.

Was this forwarded to you by a friend? Sign up for the On Student Success newsletter and get your own copy of the news that matters sent to your inbox every week.

Happy St. Patrick’s Day! May your bacon and cabbage be delicious.

What is new in student success? A theme ran through several of the things I read this week: how easy it is to talk about student success while quietly ignoring the structural forces shaping student outcomes.

When “college quality” really means “student wealth”

Scrolling through LinkedIn this weekend, a trenchant comment on a post by the American Enterprise Institute’s (AEI) Preston Cooper caught my eye.

It turns out that this was hardly the spiciest comment on the post. Although DRTC (Don’t Read the Comments) is generally great advice, in this case I encourage you—in fact—to go read the comments. Do not pass go, do not collect $200. Go read the comments.

But what caused all this criticism and spicy consternation? Preston Cooper’s post, on the surface, seems perfectly reasonable and something that we, as proponents of student success, would favor.

He used the post to reiterate arguments he had made previously: that enrollment declines at what he calls “low quality” institutions were a market correction. But in this post he laments the fact that they are showing signs of a rebound.

He starts off with the correction.

And then laments the fact that enrollments at “low quality” institutions are growing again

Cooper defines quality in terms of a composite index made up of four factors: the share of students who complete within 150 percent of normal time, the rate at which borrowers pay down principal, the loan default rate, and former students’ median earnings after leaving school. He weights each factor equally. At first glance, this seems like a reasonable way to define institutional quality. And who could possibly be opposed to quality? But the measures Cooper uses are not really measures of institutional quality at all. They are mostly measures of who institutions enroll.

The measures have little to say about the quality of education students receive. Not only are they more about who institutions enroll, but they are also strongly indexed to economic factors, making them even more reflective of the conditions shaping students’ lives before they enter college and after they leave.

Let’s look again at the four variables he uses: completion rate, loan repayment rate, loan default rate, and earnings. Each of these is strongly correlated with who the students are. For example, completion rates are heavily affected by:

• part-time enrollment

• transfer behavior

• adult learners

• academic preparation

Community colleges therefore appear “low quality” not because they fail their mission but because they serve students whose educational pathways often include part-time enrollment, transfer, and stop-outs—patterns that the metric treats as failure.

Loan repayment and default rates depend heavily on:

• family income

• student debt levels

• post-college earnings

Institutions serving low-income students will almost always have worse repayment metrics, even if the institution is performing well.

Earnings are strongly influenced by:

• major

• regional labor markets

• student demographics

• occupational sorting

A nursing program will always produce higher earnings than early childhood education, regardless of institutional quality.

What Cooper’s index largely measures is not institutional quality but the socioeconomic composition of the student body. Institutions that enroll wealthier, better-prepared students will almost always look “high quality” by these metrics, regardless of what actually happens inside the classroom.

Hence Kemi Jonas’s comment.

But wait, it gets worse

Even though Preston Cooper’s argument is flawed and revealing, and was rightly roasted by the commenters, it is nevertheless an example of two related things I see far too often in EdTech and student success research and commentary.

The first is the simple ignoring of externalities. In the student success and online learning literature that I follow most closely, I repeatedly see research or analysis that does not acknowledge several basic realities:

• Some students are far less well prepared than others because they attended less well-resourced schools.

• Many of these same students have far less to draw on in terms of family wealth and social networks.

• Some students have to attend part-time, work, and take care of family while attending college or university.

• And the kinds of returns students receive from their educations are heavily shaped by gender, race, and whether they enter the workforce in a high-cost-of-living area or a region where wages are lower.

Instead, student outcomes are often treated as a function simply of whether the institution was “low” or “high” quality, or of the sector or modality of instruction.

Should we hold institutions to account for poor outcomes? Absolutely. I do that frequently in this newsletter, and I try to highlight and broadcast other research and analysis that does the same. But if that analysis does not take externalities into account, it is not being honest—either with itself or with its readers.

A second and related tendency I see is what I refer to as the “caveat-and-proceed maneuver.” Analysts and researchers acknowledge a limitation, signal methodological awareness, and then continue as if the limitation does not materially affect the conclusion.

In Cooper’s original post he acknowledges that student preparedness may play a role. The emphasis is mine.

This is a very small and rote admission. But he waves off even this fleeting and partial acknowledgement.

Admitting that important factors are being left out and then using that admission as a get-out-of-jail-free card to ignore those factors while drawing bold and sweeping conclusions has become far too common in student success, online learning, and EdTech research.

These two tendencies are insidious, and should be called out every time.

A reminder to widen the lens

Related to this larger point, I was reminded recently of the need to check our assumptions.

I grew up in a pretty lower–lower-middle-class household; neither of my parents attended school beyond the age of 14. So I think of myself as fairly attuned to noticing economic and class assumptions in higher education. But a recent conversation with a friend and her husband has made me rethink that.

She and her husband both grew up in trailer parks in different parts of the Midwest, and over coffee with me recently they started talking about how there were always a couple of teachers in each school who “didn’t like poor kids.”

I asked what this meant, and they said that, for example, if you missed a day and asked the teacher if there was work to catch up on, the teachers who didn’t like poor kids would answer no. But there was indeed work, and so you would get a zero on that assignment.

Both of them ended up going to college and working in higher education. But I was shaken by this story.

Stories like this are a reminder that the forces shaping educational outcomes begin long before students arrive on campus. They are also a reminder to all of us that we need to constantly widen the lens through which we look at these issues.

The five unwritten rules of student success product names

After all that seriousness, let’s take a brief detour into something much lighter. Normally, by this point in the newsletter I’ve covered so many serious—and honestly often depressing—topics that I like to include a palate-cleansing funny post. Today, in that slot, I want to talk about student success product names.

I spend a lot of time these days looking at student success tools (and there are a lot of them). I’ve noticed that they tend to follow five key—but unwritten—rules when it comes to naming. I’ve left out actual examples, but I’m sure you can all think of some.

This is all meant in fun. I feel vendors’ pain in trying to come up with a name that hasn’t already been snagged and that will resonate with the intended audience.

Belonging vs connectedness

One place where the structural nature of student success shows up very clearly is in students’ social lives. A new survey by the College Student Mental Wellness Advocacy Coalition and the Hi, How Are You Project, highlights the importance of belonging, but the more striking signal may actually be connectedness. Thriving students report dramatically higher levels of social interaction with friends and family (81% vs. 34%), suggesting that belonging may be less a psychological state and more the result of being embedded in real social networks.

The report also reveals a large gap between awareness and help-seeking. Students overwhelmingly recognize the importance of mental health and know where to find support, yet roughly four in ten say they are uncomfortable discussing it, and nearly half fear being judged for doing so. The barrier appears to be social rather than informational.

That pattern shows up across many areas of student success. Students often avoid tutoring, advising, and other support services not because they do not know about them, but because using them signals struggle and invites stigma. If that is the case, the challenge may not simply be expanding services but normalizing help-seeking behavior.

When student success disappears from the conversation

And maybe we also need to normalize college and university presidents thinking about student success. The new Inside Higher Ed 2026 Survey of College and University Presidents is out, and it is well worth a read. But something that struck me is how oddly absent the concept of student success is. Presidents talk at length about demographic pressures, enrollment challenges, changes to DEI, institutional finances, public perceptions of higher education, and the potential of AI.

But student success itself—retention, progression, completion—rarely appears as a central concern. This is striking because many of the issues presidents highlight are tightly connected to whether students succeed once they enroll. Financial health, enrollment stability, and public confidence in higher education all depend, in part, on student outcomes. Student success shows up only indirectly—in discussions of AI tools that might support learning or in concerns about student wellbeing—but it is rarely framed as the underlying issue tying many of these concerns together.

In other words, we spend a lot of time talking about student success in higher education. But when leaders talk about institutional risk, student success itself often disappears from the conversation.

Taken together, these pieces highlight a recurring problem in higher education discourse. We talk about student success constantly, but we often define it using measures that obscure the structural realities students face, the social networks that support them, and the institutional practices that actually shape outcomes.

Until we grapple with those underlying forces, much of the student success conversation risks becoming a debate about metrics rather than about students.

Musical coda

In honor of St Patrick’s day, some Sharon Shannon, of whom I am a giant fan. The Galway Girl.

Also in honor of St. Patrick’s Day, imagine this newsletter is a snake and your job is to drive it out of Ireland.

Or at least out of your inbox, and into the inboxes of anyone who might enjoy it.

Thanks for being a subscriber.

Was this forwarded to you by a friend? Sign up for the On Student Success newsletter and get your own copy of the news that matters sent to your inbox every week.

The Modern Learner 2.0 report from Education Dynamics is primarily framed as an enrollment and marketing study. But one of its most interesting implications lies elsewhere. If the report is correct, the enrollment decision may not actually end when students enroll. For a significant share of students, it continues well into their time at the institution.

The report argues that students choose institutions through a dynamic risk analysis involving three factors: cost, convenience, and career. But the more interesting implication is that these same factors may continue shaping decisions after enrollment, making them not just enrollment variables but student success variables.

If the model described in the Modern Learner 2.0 report is accurate—both in the kinds of things students are considering and in the way they consider them—there are important implications for our understanding of student success and how we need to support students.

The three factors shaping student decisions

The report rests on the argument that the three factors shaping student choice—and for some students the decision to remain at a particular college—are cost, convenience, and career, defined in the following ways.

Cost is operationalized as affordability relative to financial risk. This includes consideration of whether tuition, fees, and aid options make the program financially feasible without unacceptable debt or economic uncertainty. In discussing cost, the report includes issues such as tuition, scholarships, and, importantly, transparency about costs.

Convenience, sometimes described as flexibility in the report, refers to whether a student perceives that a program can realistically fit into their life without disruption. In practice, the report often reduces this to questions of modality and whether preferred modalities are available.

Career refers to the degree to which a program clearly leads to specific employment outcomes and economic advancement. It includes factors such as job placement rates, earnings potential, employer partnerships, and industry-relevant curriculum.

In the report’s framing, these are not simply preferences that students weigh against each other. They function more like filters in a decision system: institutions that fail on any one of them are eliminated from consideration. Importantly, the report argues that these factors do not operate simultaneously or equally. Each governs a different stage of the decision process. Together they function as a decision system that students use to evaluate whether a program is possible, workable, and worthwhile.

Students are therefore not simply comparing institutions. They are trying to reduce uncertainty across these three dimensions simultaneously.

A nonlinear decision process

The three-factor model is complex enough on its own. But the report also argues that the decision process itself is non-linear.

The report describes how prospective students repeatedly move closer to and then further from institutions while evaluating their options, rather than progressing linearly through a funnel. Learners continually gather new information, compare alternatives, and reassess fit, with commitment stabilizing only when enough signals—around cost, convenience, and career outcomes—build sufficient confidence over time.

The report describes this style of decision-making as orbiting—a wonderfully descriptive metaphor that captures how students, in choosing an institution, draw near as they consider one aspect, may be pulled further away as they consider another, and then are pulled closer again.

Rather than moving step by step toward a final decision, students repeatedly reassess institutions as new information emerges. The key issue is the dynamism and rational risk weighting of the decision-making process.

The decision doesn’t end at enrollment

But the most striking statistic in the report is this: 28% of students—and 31% of traditional undergraduates—continue researching other institutions even after enrolling.

In the traditional enrollment model, the decision process ends at enrollment. In the Modern Learner 2.0 model, enrollment is only a provisional commitment. For the roughly one-third of students who continue reevaluating their options after enrolling, the image is less orbiting than something closer to Brownian motion—students constantly moving, evaluating, and adjusting their position as new information appears. If students continue evaluating these factors after enrolling, then cost, convenience, and career cease to be purely enrollment concerns. They become student success variables.

Researching other institutions does not necessarily mean a student will transfer. But it does create instability. Switching institutions carries real risks: credit loss, delayed progress toward a degree, and, in some cases, stopping out entirely.

In order to avoid losing students—whether to other institutions or to stopping out—institutions have tended to focus their interventions on mitigating or addressing the academic or personal barriers students are assumed to face. These interventions include supports such as tutoring, advising, early-alert systems, belonging initiatives, mental health supports, and, more recently, career guidance. The underlying assumption is that persistence problems arise primarily from academic difficulty or personal barriers. The Modern Learner 2.0 model suggests another possibility: that some students leave because they are re-evaluating whether the institution is delivering what they expected.

From enrollment decision to student success risk

If this model holds, even partially, student success work may need to expand in three areas.

Cost is not simply a financial-aid issue that appears when students run out of money and require emergency loans or micro-grants. It is a continuous risk calculation that students make throughout enrollment. In this model, affordability and cost transparency extend beyond enrollment and become part of a retention strategy.

Convenience—and the flexibility it relies on—therefore becomes not just an access feature but a persistence mechanism. When students’ lives change, programs and courses must be able to accommodate those changes through multiple delivery options and the ability to move, for example, from online to on-campus formats or vice versa. Without that flexibility, institutions risk losing students. Flexibility therefore becomes structural infrastructure for persistence.

If career alignment stabilizes enrollment decisions, then weak signals about career outcomes—especially as students learn more about a field or as labor market conditions change—may destabilize those decisions. Career visibility becomes a stabilizing signal that helps sustain student confidence.

The model also implies something important about how students make decisions. Rather than leaving because they are struggling academically, many students may be making a rational evaluation of whether the institution is delivering what they expected.

In that sense, the model turns the traditional student deficit model of student success on its head. Instead of assuming that students leave because they lack the preparation or resilience to succeed, the model suggests that students may be making rational judgments about whether the institution is delivering what they expected.

What this means for student success

If students continue evaluating cost, convenience, and career after enrolling, the implications for student success are significant.

First, cost transparency and affordability need to become a more explicit part of student success strategy. The Modern Learner 2.0 model treats cost as a threshold condition shaping whether students enter and remain confident in their decision. Yet institutions often address financial issues only when they become crises, through mechanisms such as emergency grants or stopgap financial aid. If students are continually weighing financial risk, then pricing clarity, financial planning, and transparency about costs need to be treated as ongoing supports for persistence rather than as one-time enrollment information.

Second, flexibility must be treated as a persistence mechanism rather than simply an access feature. Students evaluate whether programs can realistically fit into their lives, and that evaluation does not end once they enroll. Institutions built around rigid modality structures or fixed delivery formats may therefore create instability for students whose work, caregiving, or life circumstances change. Real flexibility—allowing movement between modalities, mixing formats within programs, and accommodating changing schedules—may become increasingly important for keeping students enrolled.

Finally, career alignment and return on investment must be visible and credible throughout the student experience. If career outcomes act as a stabilizing force in students’ decision-making, then institutions need to make the connections between curriculum, skills, and post-graduation pathways clearer and more consistent. This means embedding career relevance more deeply into the curriculum, strengthening industry connections, and ensuring that students can see evidence that their education is leading toward meaningful employment outcomes.

Taken together, these shifts suggest that student success may depend not only on academic support but also on how well institutions sustain students’ confidence that enrolling was the right decision.

A model worth taking seriously, but cautiously

I don’t believe that the Modern Learner 2.0 model fully captures how all types of students make decisions for all types of institutions. The report raises a number of methodological questions, and some of its claims feel broader than the data presented can comfortably support.

That said, I still find the model compelling. It resonates with many patterns that practitioners see in the field, particularly the idea that students are continually reassessing their choices rather than making a single definitive decision at enrollment. Even if the model is only partially correct, it suggests an important shift in how we think about student success.

It also has the virtue of subverting the deficit framing that has long shaped much of the student success conversation. The deficit model tends to locate problems primarily in students themselves—whether in their preparation, resilience, or ability to navigate the institution. In doing so, it can subtly shift attention away from the ways institutions may be failing to deliver on what they promise. The Modern Learner 2.0 model points in a different direction: students may be leaving not because they lack the capacity to succeed, but because they are continually evaluating whether the institution is meeting their expectations around cost, flexibility, and career outcomes.

If students are continually evaluating whether their institution is delivering on these factors, then supporting student success may require more than academic interventions. Institutions may also need to stabilize the conditions that sustain students’ confidence that enrolling was the right decision—because for many students, the enrollment decision doesn’t actually end at enrollment.

The main On Student Success newsletter is free to share in part or in whole. All we ask is attribution.

Thanks for being a subscriber.

Was this forwarded to you by a friend? Sign up for the On Student Success newsletter and get your own copy of the news that matters sent to your inbox every week.

I am writing most of this on March 1st, so Happy St David’s Day to you all, or Dydd Gŵyl Dewi hapus, as the Welsh might say.

But what happened this week in student success?

What if the biggest problem in student success isn’t financial aid, academic preparation, or advising capacity, but isolation?

This week I read two studies from very different contexts: HSIs in California and a large national online university. They have little in common structurally, but they point to the same uncomfortable conclusion: students are making consequential decisions alone.

They are not seeking help from the infrastructure we have spent a decade building. At the very moment when institutions have constructed the largest student success apparatus in higher education history — advising centers, early alerts, dashboards, one-stop shops, CRM nudges — students are quietly navigating without us.

That should stop us in our tracks.

The silent exit

A deceptively short research brief from the Community College Research Center (CCRC) surfaces a troubling pattern in how students leave college. Understanding the Needs of First-Generation College Students Who Stop Out examines a small sample of 55 students, with 10 in-depth interviews, drawn from a broader study of first-generation students at four public HSIs in California. The sample is modest but the implications are not.

The reasons students stopped out will not surprise anyone: financial strain, family responsibilities, uncertainty about academic pathways, feelings of isolation and disengagement, academic challenges in online courses.

But the more important finding is not why students stopped out. It is how. Students largely made the decision to stop out alone.

To the extent that students did discuss stopping out with anyone, it tended not to be with advisors or faculty. From the interviews, help-seeking was focused on immediate family members. No students consulted faculty.

Adapted from Understanding the Needs of
First-Generation College Students Who Stop Out

This should trouble us deeply. We spend enormous energy building advising systems, dashboards, early-alert flags, and one-stop shops. Yet when students face one of the most consequential decisions of their academic careers, they are not consulting those structures. That is not merely a service delivery gap; it is a relational problem.

Students Blame Themselves

Something that is often overlooked, but that appears again and again in qualitative studies of student success, is how frequently students blame themselves for having failed. The researchers note:

No matter what student support structures we create, no matter how many early alerts, one-stop shops, or dashboards we build, if students frame the difficulties they face as personal failure rather than structural friction, they won’t raise their hands. That’s not a resource problem; it’s a narrative problem. If students interpret structural friction as personal deficiency, no amount of infrastructure will fix it.

If the CCRC study shows us what happens at the moment of stop-out, silence, self-blame, decisions made in isolation, the WGU study I am about to discuss helps explain the broader ecosystem that produces that silence. It is not just that students fail to seek help at the point of crisis. They are structurally disconnected long before crisis arrives.

Belonging isn’t enough

If the CCRC brief shows us how students leave, a new report from Western Governors University (WGU) Labs helps explain why. In Degrees Without Doors: Why Peer and Professional Networks Still Elude Online Learners, WGU surveyed 545 students about peer connections and professional networks. The results are sobering.

Nearly two-thirds (64%) reported making zero connections with fellow students outside coursework. More than 70% said they wanted more meaningful connections.

Students report belonging — 77% feel they belong at WGU — but only 28% feel connected to other students. Belonging is not the same as connection.

We talk endlessly about belonging in student success circles. We rarely measure connectedness.

The gaps extend beyond peers. Nearly one in five WGU students reported knowing no one working in their desired field. More than half reported knowing three or fewer.

For students from households earning under $45,000 annually, nearly a quarter reported knowing no one in their intended field compared to just 15.3% of their higher-income peers.

Social capital is unevenly distributed. Higher education is not compensating for that.

And unsurprisingly, when networks are weak, students default to self-reliance. Eighty-one percent of WGU students said they prefer to handle things on their own. Half feel uncomfortable asking for help. Nearly half feel they should already know the answer before reaching out.

Isolation reinforces silence, and silence reinforces isolation.

The hidden curriculum

WGU Labs describes this as a failure of the “hidden curriculum” — the informal, unwritten lessons students absorb outside formal coursework: how to network, seek mentorship, ask for help, and navigate professional norms. On traditional campuses, these skills are presumed to develop organically, through study groups, campus jobs, extracurricular activities, and casual interactions. Online learners lack those ambient opportunities.

I love this framing. But I also think we give traditional institutions too much credit. We assume students absorb this social operating system simply by proximity. My hunch is that, if we looked closely, we would find that many on-campus students are no better connected than their online peers. WGU simply had the backbone (I originally referenced a different body part in this sentence) to explore the issue and publish the results.

I would love to see similar studies of on-campus students and institutions. My suspicion is that the findings would not differ dramatically from WGU’s. If student success is about progression and completion, it is also about connection, to peers, to professionals, to mentors, and to systems that normalize struggle and make help-seeking ordinary rather than exceptional.

We are very good at building infrastructure but far less intentional about building networks.

Books outside of higher education

I read a great deal of higher education research. I also read a great deal outside it. More often than not, the most useful insights into student success come from books that are not about higher education at all.

A few from the past year that sharpened my thinking:

The Friction Project: How Smart Leaders Make the Right Things Easier and the Wrong Things Harder by Robert I. Sutton and Huggy Rao

Sutton and Rao examine how unnecessary obstacles and inefficiencies slow organizations down. They describe some leaders as “friction fixers” or people who remove harmful friction and, at times, introduce constructive friction where it improves decision-making.

Student success is a minefield of normalized friction: opaque processes, unspoken expectations, stigma around help-seeking, confusing financial systems, invisible networking norms. We have mistaken adding resources for reducing friction. They are not the same thing.

How Big Things Get Done: The Surprising Factors That Determine the Fate of Every Project, from Home Renovations to Space Exploration and Everything In Between
by Bent Flyvbjerg and Dan Gardner

Bent Flyvbjerg and Dan Gardner analyze why large projects so often go over budget and underdeliver, and what distinguishes successful ones. Their prescription is deceptively simple: think slow and act fast, define scope clearly, break work into modular pieces, and plan with realistic odds.

Student success initiatives may not involve billion-dollar infrastructure, but they are enormously complex. Too often, student success reform is treated as a series of initiatives rather than as a disciplined execution challenge. We rarely conduct serious postmortems when efforts fall short. We underestimate complexity and overestimate implementation capacity. We would benefit from treating student success as rigorous project management rather than as a sequence of perpetual pilot programs — and from building in the kinds of postmortems Flyvbjerg and Gardner describe.

Recoding America: Why Government Is Failing in the Digital Age and How We Can Do Better by Jennifer Pahlka.

This book is fascinating on many levels. Pahlka was part of a team brought in to fix the digital infrastructure problems that bedeviled early versions of the Obama health care initiative. She later played a similar role in repairing California’s unemployment claims system during the pandemic. Her account of fixing government digital systems is a masterclass in how technology-mediated, people-facing services go wrong, and how they can be made humane.

Her diagnosis applies directly to the digital infrastructure common in student success: advising platforms, student portals, and early-alert systems. The problem is rarely intent. It is misalignment among policy, technology, and lived experience. Our systems are optimized for compliance and reporting. Students experience them as obstacle courses.

Blackett’s War: The Men Who Defeated the Nazi U-Boats and Brought Science to the Art of Warfare by Stephen Budiansky.

I read a great deal of history, especially about World War II. I read so much of it that my spouse accuses me of having run out of options and being reduced to reading histories of dentistry during the Nazi occupation of Luxembourg. That is not quite true.

But this book offers a highly particular take on the Second World War. It examines the role of natural scientists — physicists, chemists, and others — who were brought in to advance the war effort, for example by devising new ways of tracking submarines. These scientists struggled to persuade military and policy leaders to admit them into decision-making processes and to take their expertise seriously.

On the surface, this history of scientists collaborating with military leaders is far removed from higher education. But it illustrates something deeply relevant: progress requires combining expertise from radically different domains. In student success, we need faculty, advisors, technologists, institutional researchers, and administrators working in concert. Historically, those groups have operated independently. Integration creates friction. It also creates possibility.

Breakneck: China’s Quest to Engineer the Future by Dan Wang.

I found this book compelling and fascinating, and I keep trying to get people to talk about it with me. One reader who knows China well did suggest that the arguments and examples may be a bit too neat. That is very likely the case — I am a China newbie (even though this guy helped teach me how to drive). But I was struck by Wang’s distinction between the United States as a nation of lawyers and China as a nation of engineers — and by how that distinction plays out in policy.

I certainly see this dynamic in AI policy in education in both the United States and China. But I have been wondering whether there is a similar — and China-unrelated — way to structure an analysis of student success (or EdTech) initiatives: less focus on policy and rules, and more about implementation and experimentation.

Where This Leaves Us

Two different studies, two very different institutional contexts, one common pattern: students navigating alone.

We have built advising centers, predictive analytics, CRM campaigns, early-alert systems, tutoring platforms, success coaches, and emergency aid programs. And yet, when students contemplate leaving, they do not call us. When they struggle, they blame themselves. When they need professional networks, they often have none.

This is not primarily a funding problem, a staffing problem, or even a data problem. It is a connection problem.

The infrastructure of student success has grown dramatically over the past decade. But infrastructure without connection is architecture without inhabitants. If students continue to experience higher education as something they must navigate alone, then we are not redesigning the student experience; we are decorating it.

Student success is not only about removing barriers. It is also about designing connection. And connection does not emerge accidentally. It must be built as intentionally as any advising platform or dashboard.

Musical Coda

In honor of St David’s Day, traditional Welsh lullaby Suo Gan. This version is also pretty awesome.

You may share this newsletter freely, though preference will be given to anyone named Jones, Evans, Thomas or some other similarly awesome Welsh name. I will also extend grace to those with Celtic roots, anyone who has admired a daffodil in earnest, anyone who voluntarily consumes leeks, anyone who cannot remember how many d’s are in Eisteddfod, and anyone who has ever risen from their seat during a Six Nations rugby match convinced they could improve the referee’s decision-making.

Thanks for being a subscriber.

Was this forwarded to you by a friend? Sign up for the On Student Success newsletter and get your own copy of the news that matters sent to your inbox every week.

This is the second in my monthly (roughly) Student Success Charts posts where I share charts and graphics that have caught my eye in the past month.

Unsurprisingly, the charts this month are not cheerful. They tell a story about labor markets tightening unevenly, degrees decoupling from earnings, platforms drifting away from universities, and systems aging in place. I’m not going to over-interpret them. But I can’t stop thinking about them.

Returns, Risks, and Relative Decline

A striking John Burn-Murdoch chart showed that the UK is one of the rare cases where the earnings premium has declined as higher education has expanded.

Oversupply? Or economic under-performance in creating high-paying jobs?

Either way, the relationship between participation and payoff is not automatic.

The Economic Policy Institute showed employment losses for under-represented groups.

Black women with bachelor’s degrees were especially hard hit.

The EPI ties this directly to federal layoffs — a sector where Black workers are disproportionately represented among degree holders.

Even a degree does not fully insulate against sectoral shocks.

And the US is falling further behind China in terms of first science and engineering degrees awarded.

Expansion at home. Acceleration elsewhere.

Quiet Realignments

In 2025, 84% of new Coursera courses came from non-university providers.

That is not a rounding error. That is a platform redefining its center of gravity.

Universities built the early MOOC ecosystem, but they are no longer the dominant suppliers inside it.

Meanwhile, the K–12 teacher workforce continues to age.

Source: Michael Marder

The bulge has moved steadily to the right. Fewer early-career teachers, more late-career teachers.

Pipelines tighten and replacement gets harder.

And Britain has become markedly less attractive to educated young movers.

Relative positions may shift again, but for now, the direction is clear.

Institutions age.
Platforms pivot.
Talent flows elsewhere.

Ideas That Refuse to Die

Why do weak ideas persist?

Source: Dave M

Because they are psychologically sticky, socially reassuring, and institutionally convenient.

Data alone rarely defeats incentives.

I had forgotten about the Spurious Correlations site. There are quite a few education-related spurious correlations.

Correlation is cheap. Causation is expensive. And education debates are not immune.

From the Awful Chart Collection (ACC)

And then there are charts that simply defy comprehension.

Not all data clarifies. Some of it obscures.

Parting Thought

More next month.

The main On Student Success newsletter is free to share in part or in whole. All we ask is attribution.

Thanks for being a subscriber.

Was this forwarded to you by a friend? Sign up for the On Student Success newsletter and get your own copy of the news that matters sent to your inbox every week.

This week we finally got much-needed snow here in Salt Lake City. The already stunning surroundings are even more so right now.

But what happened this week in student success?

The False Binary of Degree vs. No Degree

In a short article in Forbes, Jamie Merisotis powerfully lays out the false dichotomy between careers that require higher education and those that do not. These days we hear so much about how higher education is not worth it and that young people would be better off learning a trade. The benefits of trade-based careers are often overblown in these stories. As Merisotis argues, the distinction between the two career paths is far less stark than many would like to admit.

Merisotis describes the kind of framing to which we have become accustomed.

The reality is that earning that level of income without a bachelor’s degree is the exception rather than the rule.

Much of the discourse about jobs that do not require degrees glosses over the training required to do them, especially at a level that commands strong wages. Merisotis returns to the example of elevator repair and installation professionals introduced at the start of his article.

Much of this training is classroom-based, often at community and technical colleges. The discourse suggesting that higher education is no longer worth it disguises the fact that career preparation is not a binary choice between higher education and no higher education. Instead, as Merisotis argues, it exists on a continuum.

This needs to become a core message from higher education. Everything involves learning — just in different forms. But it should not be merely an external message; we also need to internalize it and build that understanding into how we design programs and prepare students.

Higher education has a habit of collapsing complex continuums into tidy stories. The trades myth is one example. The current AI discourse is another. In both cases, the interesting questions lie beneath the headline.

Cyborgs, Centaurs, and Cognitive Offloading

Indulge me in a few minutes of unsolicited fangirling, but I love DEC survey reports. They ask interesting questions of impressive numbers of people and present the results in a compelling way. The most recent survey on AI adoption in higher education in Latin America is a case in point.

This is the largest regional AI-in-higher-ed survey to date in Latin America, with responses from 22,941 students and 7,319 faculty. Overall, attitudes toward AI are positive, and it is widely used.

• 92% of students are using AI

• 79% of faculty are using AI

• 94% of faculty expect to use AI in the future

Given this breadth of use and the generally positive attitudes (68% of students report feeling positive or very positive about AI), the nature of that use, and the concerns about that use that students express, are somewhat surprising. By far the most common student use of AI is for basic information seeking. More formal learning applications appear much lower on the list.

Given the prominence of information seeking, I was initially surprised that so many students reported concern about AI making learning more shallow — in other words, encouraging cognitive offloading. Fully 65% of students said they worry that AI will make learning too shallow and discourage critical thinking.

But as I looked back at the list of activities, those concerns began to make more sense. A great deal depends on how students are using AI. If they are relying on it to:

• Generate ideas

• Do much of the writing — despite the close relationship between writing and thinking

• Summarize documents they do not read, wrestle with, and revisit a second or third time

then their use of AI is very likely making learning more shallow, and it shows a remarkable level of prescience on their part to be concerned. AI literacy in action!

The problem is that most frameworks for understanding AI literacy do not get down to that level. They tend to focus on risk assessment, ethics, and application. All of those are important, but I would argue they are not enough. DEC’s own AI literacy framework, which I think is a strong one, illustrates the point.

AI literacy is not just about risk, questioning, or evaluation; it is about how the student designs and manages the relationship between themselves and AI. The distinction between cyborgs and centaurs in the jagged frontier literature gestures toward this, though it does not fully capture it.

In centaur models, humans and AI split the work: the AI does certain tasks, the human does others. In cyborg models, human and AI thinking are intertwined in real time. The AI is embedded in the workflow so tightly that the boundary between the “human step” and the “AI step” begins to blur.

But just as it is not always straightforward which model produces better outputs, I do not think one model is inherently better for literacy than the other. It depends on how much cognitive work the human — in this case, the student — is still doing. If we care about student success, we cannot simply ask whether students are using AI. We have to ask how they are structuring the partnership, and how much of the intellectual labor they are retaining.

Field Guide to the Higher Ed Species

After spending a week reading about institutional collapse, cognitive offloading, and retention strain, I needed a palate cleanser. Inspired by this post, ChatGPT and I collaborated — cyborg-like — to come up with some collective nouns for the higher education, EdTech, and student success space.

The original post had some good ones, including.

We came up with the following.

• A pivot of EdTech startups

• A demo of vendors

• A deck of consultants

• A hallucination of AI strategists

• A firewall of CISOs

• A rollout of change managers

• A glitch of beta testers

• A dashboard of student success leaders

• An alert of advisors

• A nudge of behavioral economists

• A backlog of software developers

Jokes aside, and switching topics, there is nothing funny about burnout, yet it remains a perennial problem in student success.

Burnout Is a Structural Design Problem

Just as students struggle when expectations and effort fall out of alignment, so too do the staff tasked with supporting them. The EAB Office Hours podcast recently featured an episode examining the work of student success professionals and what institutions can do to support them more effectively.

One of the guests, Brooke Paradise, Associate Dean of Student Affairs for Inclusion and Engagement at Skidmore College, offered excellent insights into the factors that contribute to burnout among student success staff. As she makes clear, burnout is rarely about individual weakness. It is usually about structural design and operations.

She also had some great advice on the steps that senior college and university leaders could take to better support student success staff so as to maximize their impact and effectiveness.

Building on these comments, and framing them slightly differently, here are some ways senior institutional leaders can better support student success staff — and, by extension, institutional success goals.

Align Resources with Rhetoric

If retention is mission critical, fund it like it is.

• Keep caseloads manageable.

• Reduce reliance on short-term, grant-funded roles for core work.

• Ensure compensation reflects the complexity and impact of student success roles.

Stated priorities gain credibility when fiscal decisions reinforce them.

Prioritize Focus Over Initiative Overload

Protect staff capacity by being disciplined about change.

• Sunset older initiatives when launching new ones.

• Pilot strategically, not continuously.

• Evaluate whether a new tool actually reduces workload before adding it.

Every new priority consumes bandwidth. Focus is a critical leadership decision.

Pair Accountability with Authority

If you expect outcomes, provide decision-making power.

• Give student success leaders meaningful authority over the processes for which they are accountable.

• Clarify decision rights.

• Ensure operational voices shape policies that affect day-to-day work.

Responsibility without authority creates frustration. Responsibility with authority creates ownership.

Include Student Success Leaders in Strategic Conversations

Make advising and retention expertise visible at the highest levels.

• Invite student success leaders into institutional planning discussions.

• Seek input before decisions that affect frontline operations.

• Treat advising as a strategic function, not merely a transactional service.

Inclusion signals value and improves decision quality.

Musical Coda

This video of a recording of Huw Montague Rendall & Elisabeth Boudreault singing Pa-Pa-Pa-Papagena from Mozart’s The Magic Flute is absolutely wonderful.

Someone please let me spouse know that we are soon to be headed to Paris, or maybe Quebec, or both.

Forward this post to a dashboard of student success leaders, a pivot of edtech startups, or a full-blown hallucination of AI strategists. Or all three. All we ask is attribution.

Thanks for being a subscriber.

Was this forwarded to you by a friend? Sign up for the On Student Success newsletter and get your own copy of the news that matters sent to your inbox every week.

Over the past several years the discussion on non-degree credentials has been a story of two extremes. On the one hand, there was a lot of hype. Across multiple contexts, they have been seen as solutions to all kinds of problems. They are touted as a remedy to the high cost of higher education and a way to get students into the labor force more quickly as well as an important source of new revenue for higher education institutions.

On the other hand they have been demonized as low quality (particularly in the online versions) and a way that students would be short-changed by low quality, overly quick credentials that would bear little fruit in terms of payoff.

Finally we have a correction to these skewed ways of thinking about non-degree credentials in two excellent recent reports — Brookings’ Market Value of Non-Degree Credentials and the Burning Glass Institute’s Measuring What Matters. Both are empirically serious and methodologically impressive. Both complicate the hype and the gloom, and both deserve credit for grounding the debate in data. They mark an important maturation of the debate. But maturation is not the same as completeness.

Both reports improve the measurement of credentials. But they still measure the wrong thing. Even in their sophistication and rigor, they share a common assumption: that the value of a credential can ultimately be captured through relatively short-term economic return. That assumption is the problem. It treats higher education like a vending machine: you obtain a credential and receive a wage boost. This means they don’t capture the reality of how jobs work but, worse than that, our measurement system reshapes the ecosystem. Everything becomes about the wage premium, and we never fully understand how non-degree credentials work or how to improve them.

The Correction Was Necessary

Brookings provides a careful analysis of wage outcomes associated with non-degree credentials (NDCs). The headline is not that credentials are worthless. It is that returns are uneven.

The strongest wage gains appear among individuals without bachelor’s degrees. Returns are higher when credentials are directly aligned with a worker’s occupation (or job-relevant as they phrase it). For degree holders and experienced employees, the marginal wage benefit of additional, job-irrelevant credentials is limited.

This is an important corrective. It pushes back against the idea that simply stacking credentials will reliably produce upward mobility. It also complicates marketplace strategies built around the assumption that every additional badge meaningfully moves the needle.

Burning Glass takes the debate further. Rather than focusing only on immediate wage gains, they broaden the definition of value to include career mobility. 6% of NDCs payoff immediately (what they call a Launchpad), about 8% allow employees to move to better positions in the same field (Promotion Catalyst) or in a new field for 17% of them (Lateral Move). But the majority of NDC holders, 69%, are unfortunately still stuck in the Dead End category with no wage payoff. In other words, only about one-third of credentials meaningfully change a worker’s trajectory, while the majority function as economic dead ends.

They also extend the analysis to longer term advancement, looking five years down the road to see which of these types of moves pays off. Things shift a little but not much. 65.5% of NDC holders are still in the Dead End category, with low wages, and low onward mobility.

Even with a broader time horizon, the underlying logic remains the same: credential → economic payoff. The frame has widened. The assumption has not.

The Signaling Story We Don’t Quite Name

One of the most revealing findings in the Brookings analysis is that returns are strongest for workers without degrees. While the Brookings report talks about signaling, it uses that designation for NDCs that are job-irrelevant. But their research shows that all NDCs have a signaling function and that should fundamentally change how we think about what these credentials are doing.

If non-degree credentials deliver their largest wage gains to those without bachelor’s degrees, that suggests that an important part of their value lies in signaling. For some workers, credentials function as threshold markers — signals that help them cross initial labor market barriers. They make a candidate legible to employers. They open doors. But once inside the labor market, additional credentials deliver diminishing returns unless tightly aligned with specific roles.

That has implications for how institutions design and market credentials. It also raises questions about the long-term viability of models built on the idea of continuous stacking as a universal strategy. Credentials may be most powerful at points of entry or transition. That is different from being universally transformative.

The Problem With Transactional Models

But my biggest problem with Brookings and the Burning Glass research is that as good as it is, it still hugely oversimplifies the way that the workplace functions. They are both ultimately based on a model where an investment in a credential pays off or doesn’t pay off in a wage gain, depending on the nature of the credential itself. But the workplace doesn’t work like this. Labor markets are messy, path-dependent, and shaped by timing, recognition, and context in ways that wage data alone cannot capture.

• A worker may complete a credential just as their firm freezes promotions. The value of that credential does not disappear, but its wage signal is delayed. A short-term earnings study may record “no return,” when the issue is timing, not irrelevance.

• Another worker may use a credential not to earn more, but to avoid earning less. In volatile industries, maintaining employability can be the return. Preventing displacement does not show up as a wage bump, but it may be economically decisive.

• In some cases, credentials function less as skill accelerators and more as signals. Brookings finds stronger returns for those without degrees. As noted earlier, part of this pattern reflects signaling dynamics.

• Few workplaces exhibit unrelenting rationality or instrumentalism. Networks and connections matter as does who gets noticed and why. Wage raises and promotions depend on these factors as on qualifications and credentials.

• Employer recognition also varies. A credential’s value depends not only on what it teaches, but on whether hiring managers understand it, trust it, and know how to interpret it. In one labor market, a certificate may open doors; in another, it may be invisible.

• And then there is simple human bias. Promotion decisions are not purely skills-based. They are shaped by relationships, networks, and institutional politics. Credentials operate within those systems, not above them.

None of this means that credentials lack value. It means that value unfolds over time, and often indirectly. It is cumulative, contextual, and contingent. The model used in so much of the debate about the ROI of credentials, including the Brookings and Burning Glass work, is appealing because it is measurable. But it is a poor description of how careers actually work.

And the problem is not just how we measure the payoff of credentials. It is how those measurements shape design and narrative. Measurement influences design. Design shapes perception. And perception ultimately drives funding and policy.

The Wage-First Feedback Loop

So it is not just conceptual. It’s structural. Our measurement system reshapes the ecosystem. This happens because there is an un-virtuous cycle or feedback loop at work. Policy increasingly demands measurable economic return. Workforce funding proposals, state ROI dashboards, and federal accountability frameworks rely heavily on short-term earnings thresholds. In response, researchers produce wage-based analyses. Those analyses reinforce the policy logic. Institutions then design credentials to optimize for those metrics. The debate narrows and becomes increasingly sterile.

If the vending-machine model describes how we think credentials work, the wage-first feedback loop explains why that model persists. The more we define value in short-term wage terms, the more we encourage a system that treats credentials as instruments for immediate income boosts rather than components of longer-term career development.
The cycle is understandable. Wage data are clean. Policymakers want clarity. But clarity is not completeness. The result is a policy environment that rewards short-term wage optimization rather than thoughtful pathway design.

We see this feedback loop in action across multiple states. In Alabama, wage-matched outcomes guide scholarship targeting and pathway development. Arkansas and Colorado have embedded ROI and employment metrics into funding formulas and reporting requirements, meaning colleges essentially compete on wage-linked outcomes. Louisiana mandates annual ROI analyses even for high school credentials, extending the wage frame earlier in the pipeline. Texas law now classifies credentials based on positive ROI and labor-market demand, steering program portfolios toward wage-linked definitions of value. Virginia’s FastForward evaluations embed continuous ROI assessment into short-term occupational training. These policies don’t just measure outcomes, they reshape how states define, fund, and prioritize credentials around wage-centric metrics.

Why This Matters

If we care about designing credentials that actually change trajectories rather than simply clear accountability thresholds, these are the questions that matter.
• For whom?
• Under what labor market conditions?
• At what career stage?
• With what employer recognition?
• Over what time horizon?

We should pay attention to whether credentials help workers enter fields, maintain employability, or navigate transitions, even if those outcomes do not show up immediately in wage data.

If two-thirds of credentials deliver limited economic mobility, that is not simply a verdict on credentials. It is a design challenge. It suggests that alignment, signaling clarity, employer engagement, and pathway sequencing matter far more than volume.
Until we answer the questions about for whom and under what labor conditions credentials work, higher education will struggle to design programs that deliver real value. Instead, we risk remaining stuck in a cycle of chasing wage metrics that will inevitably disappoint. We need to expand the argument and the measurements in order to design better credentials.

A Concluding Thought

There are clearly challenges with NDCs and they need to be improved. The problem is not that non-degree credentials fail to deliver value. The problem is that we keep trying to measure that value in ways that assume careers unfold in neat, linear increments, which unfortunately they do not.

Policy operates on annual cycles. Research often focuses on short-term wage shifts. But careers are shaped by timing, signaling, employer recognition, and economic cycles that unfold over years.

If we continue to evaluate credentials as though they were vending machines — put one in, get a wage bump out — we risk misjudging their role, mis-designing programs, and misleading students. Brookings and Burning Glass have made the debate more rigorous. The next step is to make it more complete.

ROI in terms of wage gain is not the wrong question. It is simply too small a lens for understanding how careers actually unfold.

As always, please share this with anyone who may find it interesting. All we ask is attribution.

Thanks for being a subscriber.

Was this forwarded to you by a friend? Sign up for the On Student Success newsletter and get your own copy of the news that matters sent to your inbox every week.

Everything I read this week returned to a theme I keep emphasizing: student success is not a program; it is infrastructure. Economic growth reflects educational attainment. Sector competition reflects institutional capacity. Staff well-being reflects organizational design. Analytics only matter if they reshape process.

How Is Educational Attainment Linked to State Income Growth?

A new report from the Urban Institute maps changes in real median household income from 1970–2023 across all 50 states. While national median income rose nearly $19,000 over the period, state trajectories diverged sharply. Western, New England, and Mountain West states saw the strongest growth (Utah was the highest, up 78% - none of which was due to me btw), while much of the Midwest stagnated; West Virginia was the only state with a slight decline.

And what were the two factors most strongly correlated with above-average income growth? Increases in educational attainment and growth in the foreign-born population. As strongly as I feel about the positive role immigrants play in economic vitality, for the purposes of this post I am going to focus on the education angle.

The Urban Institute is careful to note that these findings reflect correlation, not causation. The authors acknowledge several plausible alternative explanations. Higher-paying jobs may attract well-educated workers (the reverse of education driving income growth). Strong higher education institutions may draw students who remain in the state after graduation.

There are also important limitations. The education measure is narrow, focusing only on bachelor’s degree attainment. Median income, while useful, does not account for cost-of-living differences, inequality within states, or changes in household composition. Even so, the findings challenge prevailing assumptions that low taxes, warm weather, and population growth drive income growth. In fact, colder states and those with higher property taxes showed slight positive associations.

If educational attainment predicts long-term income growth at the state level, then higher education policy is not just about access or affordability; it is macroeconomic strategy.

Does Public Investment in Community Colleges Shift Students Away from For-Profits?

New research from Sophie McGuinness examines whether federal programs that funneled money into workforce development at community colleges had the effect of pulling students away from for-profit institutions.

This is an important question for student success, as she argues.

To this I would add that for-profit institutions tend to perform less well than nonprofit institutions on a range of student success metrics, including retention and graduation rates.

McGuinness uses FAFSA data and a sophisticated identification strategy to examine where students intended to apply and where they ultimately enrolled. Because FAFSA forms allow students to list multiple institutions, the data capture revealed preference before enrollment, including whether applicants considered both community colleges and for-profits.

She then tests whether students who expressed interest in both sectors shifted toward community colleges when those colleges received funding through the federal Trade Adjustment Assistance Community College and Career Training (TAACCCT) program. TAACCCT awarded roughly $2 billion to more than half of U.S. community colleges between 2011 and 2014, supporting the creation of approximately 2,000 workforce-aligned programs.

As much as I would like to geek out on the methodology, I will spare you. What she finds is a modest but meaningful shift:

• Community colleges that received TAACCCT funding saw roughly a 3 percentage point enrollment increase during the first two funding rounds.

• Among students who enrolled, those connected to funded colleges were about 5 percentage points more likely to choose a community college over a for-profit.

The methods are strong, but there are limitations. It is difficult to rule out alternative explanations entirely — for example, whether funded community colleges also improved marketing or visibility during this period.

Even so, the study highlights two important realities.

First, the boundary between sectors is far more porous than we often assume. Between one-fifth and one-quarter of for-profit enrollees also applied to community colleges.

Second, student loyalty to sectors is often overstated. Design and institutional capacity matter more. Investment alone does not guarantee outcomes, but capacity determines who captures demand.

Storytelling Masterclass

In honor of Valentine’s Day I.

Instagram post

The Staff Are Not OK: How Does Staff Well-Being Affect Student Success in Higher Education?

In a recent conversation with a senior leader in higher education, we discussed a familiar paradox. Universities are often criticized for resisting change, and from the inside it can sometimes feel as though very little shifts. And yet, it also feels as though change is constant — and occasionally gratuitous. Reorganizations. Changes in senior leadership. New technologies to implement. New strategic plans. Major initiatives reassigned to different divisions. The list goes on.

All of this change takes its toll.

A recent Australian report presents findings from a 2025 census of 11,477 staff across 42 universities. The census is part of a broader, ongoing study tracking workplace well-being over time. In this report, higher education data are compared with national workforce benchmarks as well as findings from related research on digital stress in the sector.

Using validated instruments, the authors examined psychosocial safety climate (PSC), which they define as.

The authors measure this in terms of factors such as.

The results are disconcerting, particularly the high levels of emotional exhaustion and the persistent restructuring and organizational churn.

Obviously, this is not a healthy situation, and I doubt these figures are unique to Australian higher education. Among the downstream consequences of sustained exhaustion and low psychosocial safety are students. Burnout affects the quality of advising, the responsiveness of support services, and the capacity of instructors to engage deeply with their classes.

And yes, as I was reading this, I did find myself muttering that they might try working in U.S. higher education over the past thirteen months.

Note: In Australia, the term “staff” typically includes faculty and instructors. I have assumed that is the case in this dataset.

Each of the previous three pieces points to the same conclusion: outcomes follow infrastructure.

What Is Missing from Higher Education’s Process–Data–Analytics Model?

A fascinating post by Willis Jensen got me thinking about how data actually flows through student success work. He describes a simple model he finds most useful for connecting analytics concepts: the process–data–analytics framework.

Jensen spells out the implication of using this model which I want to quote in full in part to keep myself honest.

I find the model useful, in part, as a starting point for thinking about what a data-informed student success framework should look like. But I also have some concerns.

The model does not emphasize action enough — especially not visually. It risks reinforcing a familiar higher education tendency: collecting student success data without acting on it. The framework implies that analytics automatically “improves” process. Improvement appears embedded in the loop, but action and decision-making are not explicitly represented.

In higher education, the break almost always happens here:

Data exist.
Analytics exist.
Dashboards exist.
Nothing changes.

The missing element is action. If nothing changes, you are not doing student success; you are doing reporting.

I am also uneasy with technology placed at the center, even with the caveat that it is merely an enabler. What sits at the center of a diagram inevitably becomes the pivot point. Over the next few months, I want to keep working on a model that captures the strengths of Jensen’s framework but better reflects the realities of student success work.

The Australian data show what happens when process overwhelms people. The Urban Institute data show what happens when states invest in educational infrastructure. McGuinness shows that targeted investment shifts student choice. Jensen reminds us that data alone change nothing, nfrastructure, process, and action do — but they are more work than dashboards.

Musical coda

In honor of Valentine’s Day, Part II: This is one of my favorite love songs — in its original Magnetic Fields version.

The Book of Love may be long and boring and written very long ago, but this week’s post appears also to be full of facts, figures, and charts. If you know someone who enjoys that sort of romance, feel free to forward this along and encourage them to subscribe.

Thanks for being a subscriber.

Was this forwarded to you by a friend? Sign up for the On Student Success newsletter and get your own copy of the news that matters sent to your inbox every week.

Few topics in online learning provoke stronger reactions than shortened or compressed courses. Also called accelerated or condensed courses, condensed course formats are used in both online and in-person learning. Compressed courses deliver the same credit hours and curricular content as a traditional 15-week semester course, but in roughly half the time, typically seven or eight weeks. To some, they are an obvious lever for improving completion and flexibility. To others, they are a shortcut that must inevitably erode quality.

Recent research suggests both camps are partly right, and partly wrong. Compressed courses show measurable gains in grades and withdrawals. But those gains are fragile and highly contingent. Fast isn’t the same as simple.

Compressed courses can be an important tool in both online learning and student success, but they need to be pursued as part of a larger strategy, not opportunistically adopted as a schedule hack. That larger strategy must include significant investment in faculty training and course redesign, as well as careful support for students. In this post I examine two important new research publications on compressed courses and what they tell us about how these are being used.

Compressed courses show measurable gains in grades and withdrawals. But those gains are fragile and highly contingent.

• Higher pass rates

• Lower withdrawals

• Slight GPA gains

The question is not whether compressed courses can work. The question is under what conditions they do.

The upside, and some nuance

Compressed courses are not new. As far back as 2014, Texas’s Odessa College shortened roughly 80 percent of its courses. At some institutions today, for example, Chattanooga State Community College, short-format courses now make up about 70 percent of enrollments. Yet despite their growing prominence, we have had surprisingly little rigorous evidence about how compressed courses perform at scale.

Much of what we know about compressed courses comes from single-institution studies. But two recent large-scale analyses, one examining all public community colleges in Tennessee using administrative, survey, and interview data, and another analyzing Virginia community colleges using statewide administrative records, have significantly improved our understanding of how compressed formats function in practice. Together, they provide some of the most comprehensive evidence to date on the effects of condensed courses in open-access institutions.

Both studies point to meaningful benefits, though with important nuance. The Virginia study explicitly frames this tension as one of efficiency versus burnout, asking whether compressed formats allow students to progress more efficiently without increasing longer-term academic strain. Using longitudinal administrative data, the authors examine not only immediate pass rates but also downstream enrollment and performance in subsequent courses.

Flexibility, not speed, drives student demand

We have known for a while that compressed courses are popular with nontraditional students, and the research confirms this. In the Tennessee study, for example, nontraditional learners were over-represented in condensed courses by as much as 3.5 percentage points. This makes sense: condensed formats give working adults, caregivers, and part-time students greater scheduling control.

Importantly, however, students do not appear to be choosing these courses primarily because they are shorter.

When Tennessee researchers asked students directly how they selected courses, length rarely entered the conversation. Instead, students emphasized program requirements, scheduling constraints, and modality. As the authors report.

Some students even framed efficient course selection as a form of financial stewardship and a way to stay on track for timely completion.

Compressed courses are attractive less because students want speed and more because they need scheduling control. For students balancing work, care-giving, and unpredictable lives, shorter terms function as a form of structural flexibility.

Improved short-term academic outcomes

Both major studies find that compressed courses are associated with stronger short-term academic outcomes.

In Tennessee, enrollment in condensed courses was linked to:

• Higher pass rates (approximately +4.9 percentage points) and lower failure rates (1.7 percentage points lower)

• Substantially lower withdrawal rates (49 percent lower—3.2 percent for compressed courses relative to a control mean of 6.5 percent)

• Higher average course grades (about +0.24 GPA points)

These results were robust across multiple statistical models and remained significant after accounting for student characteristics, course types, and institutional variation.

The Virginia study reported similar patterns, finding higher pass rates and improved retention among students who took condensed courses, particularly in their early coursework.

Qualitative evidence from interviews helps explain these effects. Students frequently described feeling more focused in shorter terms, with fewer competing courses and clearer near-term goals. Faculty reported that compressed schedules encouraged more consistent engagement and reduced opportunities for procrastination. Shorter time horizons, in many cases, appeared to support better study habits and sustained attention.

For students who are able to maintain the pace, the structure itself can function as a behavioral support.

Evidence of strong downstream benefits, with caveats

Where the two studies diverge is in their assessment of longer-term effects. The Virginia analysis paints a largely optimistic picture. The authors find that condensed courses not only improve immediate performance but also increase enrollment and success in subsequent courses. These downstream benefits are especially pronounced for adult learners and underrepresented minority students, who are also disproportionately enrolled in condensed formats.

From this perspective, compressed courses appear to function as on-ramps, helping students build early momentum that carries forward into later coursework.

The Tennessee study, by contrast, finds a more mixed pattern. While students who took multiple compressed courses early in their programs showed positive associations with persistence and GPA, students who took only a single condensed course in their first semester were less likely to persist. The benefits, in other words, were not uniform and appear to depend on how compressed courses are integrated into students’ overall pathways.

Taken together, these findings suggest that compressed courses can support longer-term success, particularly for adult and underrepresented students, but that these benefits are contingent rather than automatic.

A qualified success story

These wins are compelling, and it is easy to see why institutions, especially community colleges and online programs, might be tempted to expand compressed offerings as a way to increase entry points and provide greater flexibility. But looking beyond the headlines of “improved grades and mostly good persistence,” the Tennessee and Virginia studies, along with a number of smaller single-institution analyses, point to the need for caution.

Compressed courses do not create institutional strengths or weaknesses. They magnify them. In institutions with coherent advising, clear communication, and serious investment in course design, they amplify momentum and flexibility. In institutions where those systems are thin or fragmented, they amplify stress, confusion, and risk.

Stress is real — and it’s not trivial

In the Tennessee study, students who had taken both condensed and traditional courses consistently reported that the condensed format was more stressful:

For some students, especially those who are focused and whose lives are stable, the increased intensity may enhance engagement. For others, particularly those managing work shifts, childcare, health issues, or transportation challenges, the compressed structure magnifies disruption.

A 15-week course compressed into seven or eight weeks doubles the pace. A single missed week becomes proportionally more consequential. Acceleration sharpens focus, but it also sharpens risk.

Beyond the transcript: what GPA doesn’t capture

The Tennessee study’s mixed-method approach surfaces a deeper question: what are we actually measuring when we celebrate higher grades?

Both faculty and students raised questions about whether students were mastering material as deeply in condensed formats. Faculty also reported “adapting grading” in compressed courses, raising the possibility that some of the measured gains reflect institutional adaptation to time pressure rather than deeper mastery.

Another theme in the Tennessee interviews was reduced contact with peers and instructors. Students described fewer opportunities to build relationships, collaborate meaningfully, or form academic bonds, findings echoed in several recent single-institution studies.

In traditional discussions of student success, we emphasize belonging and connection. Compressed formats, by definition, reduce the time available for those relationships to develop. If student success is not only about passing courses but also about building durable academic identity and peer networks, then course length may matter in ways that GPA does not capture.

The flexibility paradox

One of the most interesting tensions in the Tennessee study is this:

Students value compressed courses for their flexibility, but the compressed nature of those courses simultaneously reduces flexibility.

Flexibility at the structural level can translate into rigidity at the week-to-week level. This does not invalidate the model. But it complicates the narrative.

Doing it right, and doing it wrong

If the research shows anything clearly, it is that compressed courses are not a plug-and-play innovation. Their effects depend heavily on how institutions implement them.

Over time, a distinction emerges between two types of institutional behavior.

Timetable hackers vs. system builders

Reading across the Tennessee and Virginia studies, along with a raft of recent single-institution analyses, it becomes clear that there are fundamental differences in how institutions approach compressed courses. Some treat them as calendar modifications: they shorten the term, keep most structures intact, and assume outcomes will follow. Others treat compressed formats as system redesign, rethinking advising, registration, course development, faculty workload, communication systems, and pathway sequencing.

The difference is not cosmetic. It determines whether compressed courses function as a support mechanism or a stress multiplier.

Warning sign #1: Students don’t know what they enrolled in

One of the most troubling findings from the Tennessee study was that some students did not realize they had enrolled in a condensed course. As the researchers report, course length was not clearly flagged in the registration system.

This is not a minor administrative oversight. It is a structural failure. Clear designation, on-boarding communication, and advising alignment are not optional details. They are prerequisites.

Warning sign #2: Ad hoc adoption

One of the most nuanced findings in the Tennessee study is that students who took multiple compressed courses in their first semester showed positive associations with persistence—but students who took only a single condensed course were less likely to persist. That is a remarkable finding. It suggests that compressed courses are not neutral units that can be sprinkled into a schedule. Their effectiveness depends on pathway coherence—on being integrated deliberately into a student’s overall program rather than scattered randomly among traditional courses.

When institutions dabble, offering a few isolated sections without broader pathway alignment, they may inadvertently increase complexity rather than reduce it.

What doing it right looks like

The large-scale studies in Tennessee and Virginia provide valuable evidence of both the strengths and the challenges of compressed courses. Taken together with insights from recent single-institution studies, they point toward the components of an effective compressed-course strategy. These include:

• Clear course designation in registration systems

• Integrated academic calendars

• Limits on overload (policies governing how many compressed courses students take at one time)

• Structured advising aligned with pacing

• Faculty development and intentional instructional redesign

• Explicit on-boarding about workload expectations

• Coordinated student support services

In a future post, I will explore what these elements look like in practice.

Parting thoughts

Compressed courses are not inherently good or bad. They are amplifiers. In institutions with coherent advising, transparent communication, and serious investment in course redesign, they can accelerate momentum and provide meaningful flexibility for adult learners. By contrast, in institutions where they are treated as timetable hacks, they magnify stress, confusion, and fragmentation.

The emerging research suggests that compressed courses can provide needed flexibility while keeping top-line metrics stable. But it tells us something more important: institutional design still matters.

As institutions face demographic pressure and increasing scrutiny over outcomes, the temptation to treat compressed courses as a quick lever will only grow. But fast is not the same as simple.

The main On Student Success newsletter is free to share in part or in whole. All we ask is attribution.

Thanks for being a subscriber.

Was this forwarded to you by a friend? Sign up for the On Student Success newsletter and get your own copy of the news that matters sent to your inbox every week.

Zora Neale Hurston once wrote.

We are only a month in, but I am beginning to suspect that we are in an answering year: about what is important, where our values lie, and even about AI. I fear that I am not going to like all the answers, but even so, there is something to be said for clarity.

So what is new this week in student success?

What I read this week reinforced something I’ve been noticing for a while. We put enormous explanatory weight on the choices that students make: whether to go to university, what to major in, and which supports to use. But student success, and the choices themselves, are deeply shaped by the larger ecosystems in which those choices are embedded.

Student success is being judged more and more by outcomes. Yet those outcomes are shaped less by individual decisions and more by institutional systems that students often can’t access, don’t trust, or don’t benefit from equally.

It’s not the major, its the market

In explaining differential earnings by gender and race, we are accustomed to hearing explanations that attribute these gaps to choice of college major. Women and underrepresented minorities (URMs), we are often told, do not choose potentially high-paying majors, especially in STEM, and therefore tend to earn less than men and white students.

Some new research published by the NBER puts a serious dent in that argument. The researchers used administrative data from the state of Texas to track earnings over twenty years for graduates of Texas public schools working in STEM-related occupations. At some point, we do need to talk about the fact that so many of our insights come from research based on Texas data, and what that might mean. But that day is not today.

What they find is that if you look only at white or Asian graduates, major choice has some explanatory value. For everyone else, it has little. Within the same major, women and URMs are paid less, and this gap increases over time.

A similar pattern plays out in terms of race.

What they do find is that occupational choices within specific majors explain some of the differences. Using complex methods, the authors show that much of the payoff from a major depends on which jobs graduates end up in, not on the major itself. Women and minority students are much more likely to end up in lower-paying occupations within the same field. When the authors simulate what earnings would look like if occupational sorting were equalized, most of the gaps disappear.

In other words, the problem is not mainly what students study. It is where they are able to go afterward, and who helps them get there.

This matters for how we think about “return on investment” in higher education.

If we are concerned about earnings, it is not sufficient simply to advise students to major in STEM and enter STEM fields. We also need to do a better job of helping students understand within-occupation choices and their implications.

We should also be skeptical of accountability data used to evaluate the ROI of degrees, such as that coming out of the recent OBBB in the US, that assumes similar payoffs for all students or relies on averages that obscure in-major differences in earnings. Failing to account for this risks penalizing institutions for labor-market dynamics they do not control.

Why students don’t use services

If outcomes depend on what happens after graduation, they also depend on whether students can navigate support while they are still enrolled. A while back, I wrote about a Tyton Partners report that seemed to point to a fairly straightforward problem: a mismatch between the kinds of support institutions provide and the kinds of support students say they want. The implicit story was that students weren’t using services because they didn’t know about them, couldn’t access them easily, or didn’t see how they were relevant.

A more recent report from Tyton Partners complicates that story.

Its updated visualizations again show large gaps between institutional availability, student awareness, and actual use of support services. On the surface, this looks like a familiar problem: institutions build services, advertise them imperfectly, and students fail to take advantage of them.

But when you look more closely at why students say they don’t use available supports, a different pattern emerges.

The top three reasons students give have nothing to do with logistics. They are:

• “I want to do things on my own.”

• “Support is not for students like me.”

• “I doubt it will be helpful.”

Taken together, these point to something much deeper than a communication problem. They suggest a lack of trust in the support itself.

In other words, this is not primarily a visibility problem. It is a legitimacy problem. Support that lacks legitimacy becomes invisible, no matter how visible it is.

Many students do not experience institutional support as something designed for people like them, in situations like theirs, and at moments when they actually need it. Instead, support is often perceived as remedial, stigmatizing, bureaucratic, or disconnected from real academic and personal challenges. For some students, using support feels like admitting failure. For others, it feels like entering a system that does not understand them.

This helps explain why awareness alone does not translate into use.

In many ways, this mirrors a familiar challenge in faculty development. Institutions regularly offer training in areas such as online course design, assessment, or inclusive teaching. Faculty are aware that these programs exist, and many agree they are important. Yet participation remains uneven. Highly skilled professionals often resist guidance that feels generic, misaligned with their context, or disconnected from their day-to-day work.

With students, the dynamic is similar—but the stakes are higher. Institutions are working with learners who may be unfamiliar with how universities operate, uncertain about their own standing, and sensitive to signals about belonging and competence. When support systems fail to account for this, they quietly lose credibility.

Some barriers are still practical. Long wait times, limited hours, lack of walk-in availability, and inflexible modalities continue to matter, and there is little excuse for allowing these problems to persist.

But the larger challenge is cultural and relational.

If students do not believe support will help them, will respect them, or will understand their circumstances, they will not use it, no matter how well funded or widely advertised it is. Addressing that gap requires more than better marketing. It requires rethinking how support is designed, delivered, and embedded in students’ everyday academic lives.

It also suggests a need for much more qualitative research into how students actually experience institutional support, and how those experiences shape trust over time. Until institutions understand that, they will continue to invest in services that exist on paper but fail to function in practice.

A sad horse for an answering year

This odd story captures something about how many people in higher education seem to be feeling right now. A factory in China accidentally stitched a scowl onto the face of a stuffed horse -it is the Chinese Year of the Horse—turning it into an unintended bestseller.

In the Year of The Horse at least some aren’t liking the answers so far.

Moving on up

Student social mobility, that is, the ease with which students can achieve high levels of income, is a critical aspect of student success. A new report from the Sutton Trust provides some fascinating insights into this issue across multiple countries.

The report examines the relationship between higher education and social mobility, defining mobility as reaching the top 20 percent of the national earnings distribution. It is important to note, however, that this definition focuses on reaching the top quintile without referencing where students might have started.

The authors find that mobility varies enormously across countries, but that it is deeply shaped by whether students’ parents have a degree themselves.

If parental education shapes access to high-paying occupations, then advice focused on choice of major or institution misses the point. This is an angle that is implied but not fully explored in the otherwise excellent work of scholars such as Raj Chetty.

If it works, why aren’t you using it?

A report from RNL (now part of Encoura) left me puzzled for two different reasons.

On the one hand, the data suggest that institutions are making very limited use of AI-driven tools for student success. On the other, the report describes extraordinarily high effectiveness among the small number of institutions that say they are using them.

Those two findings do not sit comfortably together.

According to the report, only small fractions of institutions say they are using AI in areas such as early alert systems, academic advising, or proactive outreach. For example, just 6 percent of four-year private institutions, 11 percent of four-year publics, and 3 percent of two-year institutions report using AI in early alert communications.

These usage rates are lower than I would have anticipated.

One possibility is that they understate reality. Many early alert systems and CRM platforms now embed some form of machine learning or automated pattern detection. It is plausible that some respondents are using AI-enabled tools without labeling them as such.

But even if that explains part of the gap, it cannot explain all of it. The overall picture still points to very limited and uneven adoption.

The second puzzle is the report’s claim that early alert systems using AI are 100 percent effective.

At best, this reflects the limitations of self-reported survey data. At worst, it reflects wishful thinking.

Over the past decade, I have repeatedly asked institutional leaders whether early alert systems have clearly and consistently improved student outcomes at their institutions. In all that time, I have received only one unequivocal yes. More often, the answer is some version of “it helps in certain cases” or “it works when everything else is in place.”

That experience makes a universal effectiveness rating deeply implausible.

What the RNL data likely captures is not objective impact, but perceived usefulness among a small, self-selected group of early adopters. Institutions that invest heavily in these tools, integrate them into workflows, and assign staff to act on the data are more likely to view them positively. Institutions that struggle to operationalize them either do not adopt them at all or abandon them quietly.

Seen this way, the real story in this report is not about AI’s technical potential. It is, once again, about institutional capacity.

AI tools for student success do not work in isolation. They require clean data, integrated systems, clear governance, trained staff, aligned incentives, and sustained attention. Without those conditions, they generate dashboards rather than decisions, and alerts rather than action.

This helps explain the paradox of low adoption and high reported effectiveness. Where institutions have built the surrounding infrastructure, AI can amplify human work and improve targeting. Where they have not, it adds another layer of complexity to already strained systems.

It is the student success Groundhog moment: new tools, new reports, new dashboards, same structural constraints. Across earnings, support, mobility, and AI, the same ecosystem problem keeps resurfacing. This is not primarily a technology problem. It is an organizational and systemic one.

I suspect this will not be limited to Groundhog Day, and that we will keep returning to it in this year of answers.

Musical coda

I came across this, oddly enough, through a recommendation from Adam Savage, and I’ve been enjoying it. Bowie is a tough artist to cover, and “Life on Mars” is an especially difficult song.

On Student Success is free, and you are welcome to share it (with attribution).

Last week, I asked you not to forward it, in order to preserve the illusion of a small, slightly secret society of people who enjoy thinking too much about higher education.

This week, I am reluctantly conceding that some of your friends probably belong in that society too.

Thanks for being a subscriber.

Was this forwarded to you by a friend? Sign up for the On Student Success newsletter and get your own copy of the news that matters sent to your inbox every week.

There are so many stressful things going on in the world, and especially in higher education, that even the folks on my pre-recorded meditation app are starting to sound tired. But what is happening in student success?

Much of what I read this week points to the same uncomfortable conclusion: our biggest student success problems are not about data, tools, or even resources, but about organizational capacity and institutional will.

Online students as afterthoughts

A white paper based on research by Don Hossler, J. T. Allen, and Luke Schultheis sheds some sobering light on the kinds of support provided to students studying online. It draws on a larger research study, which I will cover when it is released. The report is sponsored by MyFootpath, a vendor that provides enrollment and re-enrollment support, so its framing should be read with that in mind. Even so, it is solid research and worth reading, as it offers much-needed insight into the state of online student success, albeit based on a small, self-reported sample.

The study is exploratory in nature, relying on interviews with university administrators who met the following criteria.

These thresholds meant that fully online institutions were excluded. At the same time, the institutions that were included represent a wide range of experience with online learning, from Stony Brook and Illinois State University, with just 1% and 2% of students studying fully online, to online heavyweight Oregon State University, where 35% study fully at a distance (yes, that is more than 33%). The findings should therefore be interpreted in light of this variation and the fact that these institutions are at very different stages of online development.

Even with this caveat, the picture that emerges is not encouraging. In their interviews, the authors found that support for online student success is, for the most part, haphazard.

• 83% of institutions reported having no well-defined organizational structure for online student success.

• Institutions employed a range of organizational models, from more centralized approaches to highly distributed arrangements in which responsibility was split across central units and academic departments. Tracking of student success was far more systematic in the centralized models.

• Only 42% of institutions believed their university had a strong institutional focus on online retention and graduation.

• 69% did not regularly track online student retention, and staff often lacked independent access to these data.

Taken together, these findings point to a deeper problem: many institutions have expanded online learning faster than they have built the organizational capacity to support it. Online students generate enrollment and revenue, but not always institutional priority.

Early alerts, late responses

The organizational failures described in the online student study are not unique. A similar pattern appears in how institutions use—or fail to use—early alert systems. I have written before about the way higher education tends to be far more interested in collecting data than in acting on it. A recent post persuaded me that this is not just a US phenomenon.

In the UK, the higher education regulator, the Office for Students (OfS), has for several years required universities to meet benchmarks of at least 80% continuation (or persistence, as we would call it here—potato, potahto) and 75% completion for first-time degree students. In principle, such mandates should encourage widespread and systematic use of early alert systems that flag students at risk. Writing on Wonkhe, Carmen Miles argues that this has often not happened. Even where these systems are in place, universities frequently fail to act when students are flagged as at risk. She writes.

She argues that this is a readiness and a governance issue.

I think this gets at part of the problem, but not all of it. There are also important weaknesses in early alert systems themselves. Some are calibrated to be overly sensitive, triggering flags so frequently that staff begin to ignore them. At the same time, students, especially first-generation students, may interpret “at risk” warnings as confirmation that they do not belong at university, turning what is meant to be support into a self-fulfilling prophecy.

Governance matters. But so does system design. Over-sensitive alerts, weak calibration, and poor communication can turn well-intentioned tools into sources of anxiety rather than sources of support.

Three guys from Boston

After all that grim news, a little levity. For reasons I do not entirely understand, Massachusetts accents seem to provoke disproportionate hilarity. Although I am not from Boston, my own accent appears to have much the same effect. I once had a delightful colleague who, every time she passed my office, would stick her head through my doorway, say the word “banana” in my accent, roar with laughter, and disappear.

And now back to scheduled depressing programming

But dashboards don’t fix structures

I have long threatened to write a book about the graduate school experience titled The Pedagogy of the Depressed. For many people, especially in PhD programs, it is a long, grinding process in which classmates gradually fall by the wayside.

Despite this, graduate persistence and completion remain surprisingly understudied. Much of higher education’s attention is focused on undergraduate student success—which is rightly seen as critical—but this emphasis leaves a significant gap in our understanding of the challenges facing graduate students.

A new report by Jeffrey Denning and Lesley Turner from the PEER Center begins to fill that gap. Drawing on administrative data from Texas for cohorts entering between 2003–04 and 2012–13, the authors document several important patterns.

Most strikingly, they find that only 62% of graduate students complete a degree within six years, with wide variation across fields of study.

All of this makes the report a useful start. But serious analysis of graduate completion needs to move well beyond simple breakdowns by discipline.

Master’s degrees and PhDs are fundamentally different enterprises and should not be analyzed together. The same is true of professional programs such as JDs and MBAs, which are typically more structured, time-bound, and better supported than disciplinary master’s programs. Treating these programs as interchangeable obscures important differences in risk, resources, and student experience.

At the same time, much graduate study now takes place online, making it essential to disaggregate outcomes by modality. Many online students, and some in-person students, also study part time, which complicates standard completion timelines and calls for more flexible benchmarks.

Without this level of disaggregation, we risk reproducing in graduate education the same blind spots we have long tolerated in undergraduate and online programs. Any serious effort to improve outcomes is likely to falter if it rests on such incomplete analysis.

I also have reservations about the authors’ policy recommendations, which I will explore in more detail in a separate post. They lean heavily on the familiar call for “more data,” while largely sidestepping the harder question of how graduate education is structured and supported.

Even so, it is encouraging to see sustained attention to this neglected area. This is a solid study and an important starting point.

Taken together, this week’s readings paint a sobering picture. Whether we are talking about online learners, undergraduates flagged as “at risk,” or graduate students struggling to finish, the same pattern appears again and again: institutions have invested heavily in data and systems, but far less in the organizational capacity needed to act on what those systems reveal. Student success remains too often a technical project layered on top of structures that were never designed to support it.

Musical coda

In keeping with the overall grim tone of this post, I am sharing my favorite cover of what I believe is a truly depressing song, delivered in the way I think it should be delivered (sorry, The Boss).

On Student Success is free, and you are welcome to share it (with attribution).

Last week I encouraged enthusiastic forwarding.

This week, I would like to retract that advice. Please do not forward this. I prefer to imagine On Student Success as a small, slightly secret society for people who enjoy thinking too much about higher education.

Was this forwarded to you by a friend? Sign up for the On Student Success newsletter and get your own copy of the news that matters sent to your inbox every week.

I have a modest political pin collection, and one of my favorites dates back to the 1936 US presidential election, when Alf Landon of Kansas ran against the incumbent Franklin Roosevelt. It’s a small artifact from one of the most lopsided elections in modern American history, and a useful reminder of how badly conventional wisdom can misread reality.

This image is from one for sale on Etsy - my photography skills being what they are

Prior to the election, The Literary Digest ran a large-scale survey. Having correctly predicted the winners of the previous five presidential elections, the magazine was widely trusted as a political forecaster. Based on 2.38 million responses (equivalent to 18.7 million people in today’s currency!), it confidently predicted that Landon would win in a landslide.

He didn’t.

Instead, Landon carried only Vermont and Maine, winning a mere eight electoral votes, one of the lowest totals in U.S. history.

Many explanations have been offered for this failure, but most come down to two related problems.

• First, the magazine’s readership skewed far more Republican than the general population.

• Second, the survey relied on voluntary participation, meaning it captured the views of people who were especially motivated to respond rather than a representative cross-section of voters.

The result was a massive dataset that was systematically misleading.

Which is to say: surveys can be enormously valuable, but their findings, and the conclusions drawn from them, must be interpreted with care.

I’ve been thinking about this a lot lately, and a recent survey report from AAC&U offered a timely reminder of why survey-based evidence deserves close, critical scrutiny.

A good survey, read carefully

AAC&U recently published a fascinating, and methodologically sound, report on employers’ attitudes toward the skills of recent graduates. It is a strong piece of work, with many interesting and carefully presented findings. This is exactly the kind of survey we need more of in higher education.

One of the biggest surprises for me was the extent to which employers value higher education’s role in educating the whole person. In particular, they see universities as responsible not only for preparing students for work, but also for helping them become informed citizens. Fully 94 percent of employers say it is very or somewhat important for institutions to help students develop as informed citizens—the same proportion who believe it is important for universities to prepare educated workers for the economy.

This runs counter to the common impression that universities should focus narrowly on workforce skills and immediate job preparation.

Less encouragingly, the survey also revealed more ambivalent views about different kinds of out-of-classroom learning experiences. Internships, as expected, remain highly valued. But employers’ relatively muted enthusiasm for more widely accessible experiences—such as on-campus employment, senior projects, and study abroad, is more concerning, particularly given how much more accessible many of these are to students, compared to internships.

Who you ask changes the answer

he AAC&U report is both valuable and fascinating, and under normal circumstances I would spend much more time unpacking its details, especially because many of its findings run counter to what we often hear in the media as received wisdom.

To be clear, this is a well-executed survey, with a strong sample and careful, thoughtful analysis. It is well worth reading in full and discussing more widely.

One aspect of the report, however, stood out to me in particular. The way AAC&U disaggregated its data served as a useful reminder of how much survey findings depend on whom you ask. This became especially clear in their treatment of one of the most contested questions in higher education today: how well institutions are preparing students for the workplace.

Contrary to much of the current prevailing narrative, a majority of employers in the AAC&U survey believe that higher education is doing a reasonably good job of preparing graduates for work.

But what is most striking are the breakdowns by age and political affiliation. Republicans and Democrats are remarkably similar in their views on this issue, while Independents emerge as a clear outlier.

There are also substantial differences by age. Employers aged 50 and older are far more critical of how well higher education prepares students for the workforce. This age-based divide becomes even clearer when AAC&U asks more focused questions about the skills and knowledge needed to succeed in entry-level positions and to earn promotion.

There is a 12-point gap between employers under 40 and those over 50 in their assessments of how well students are prepared for entry-level jobs, and a 16-point gap in their views of whether recent graduates have the skills needed for promotion. Both differences are substantial, and too large to dismiss as statistical noise.

Why do we keep forgetting this in higher ed?

This made me wonder how much of the recent coverage criticizing the skills of new graduates, and questioning whether they are prepared to thrive in the workplace, reflects who is being surveyed rather than what is actually happening. If samples lean disproportionately toward older employers and political independents, it is not surprising that the dominant narrative skews toward disappointment and decline.

I’ve seen this dynamic play out repeatedly in institutional decision-making. In workforce alignment projects, for example, colleges often rely on aggregated employer survey results to guide curriculum redesign. Those summaries tend to emphasize broad themes, “communication skills,” “adaptability,” “AI literacy.” But when the data are disaggregated by industry, firm size, or respondent role, the picture is far more fragmented. Those nuances rarely survive into strategy documents. Institutions end up designing programs around a simplified narrative that reflects who responded, not necessarily what students will encounter in the labor market.

A similar pattern appears in student support design. Institutions frequently build advising, retention, or engagement strategies based on surveys or focus groups that under represent key segments: online students, commuters, working adults, stop-outs, or caregivers. In other cases, programs are shaped primarily by staff perceptions of what students need. All of these perspectives matter. But each represents only one slice of a much larger and more complicated picture. When that slice becomes the whole, blind spots become baked into institutional design.

This is why findings from surveys and other data-gathering exercises need to be questioned more rigorously. So much depends on whom you ask, how you ask, and whose voices end up dominating the results. The patterns are often far less obvious than headline numbers suggest.

Higher education repeatedly builds policy and strategy on fragile evidence. The problem is not that surveys are useless. It is that we too often treat them as settled truth in environments where the stakes are enormous. More data will not solve this. What institutions need instead is greater interpretive discipline: the capacity to parse evidence carefully, understand its limits, and use it to inform, rather than dictate , policy and practice.

If you enjoyed this post or found it valuable, please forward it to everyone you know, and even some people you don’t!

Thanks for being a subscriber.