This Week in Student Success

When our assumptions about students, employers, and AI stop matching reality

Author

Glenda Morgan
July 16, 2026

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’s still hot, the hills around me are still burning and I am suffering from cricket withdrawal. But what is happening with student success?

This week everything I read seemed to be about misaligned assumptions.

The confidence gap

One of my mid-year resolutions is to stop writing about what employers want in graduates.

Not because the question isn't important, but because it has become too dominant. It reduces higher education to workforce preparation, assumes employers always know what they need (they often don't), and too easily shifts responsibility for every perceived skills gap onto colleges and universities.

That said, a new report by Pearson and AWS highlights some interesting misalignments between higher education leaders, learners (hereafter referred to as students), and employers.

The biggest misalignment for me concerns the perceived value of higher education. Students are the least convinced that university education is becoming more important in the age of AI. They were less likely than either employers or higher education leaders to say that university education is becoming more essential, and more likely to believe it is becoming less essential.

The chart below is worth a closer look. The gap isn't huge—62% versus 67%—but it's psychologically fascinating.

Chart showing perceived importance of university education in the ai era - students are less likely to believe university education is more important and more likely to believe it is less important that employers or HE leaders

The people universities are trying to convince are the least convinced. Meanwhile, employers—the people doing the hiring—are the most convinced. That suggests a messaging failure.

The UK results are particularly striking, where students are substantially more likely to believe that higher education has become less important in the age of AI.

A second interesting finding concerns hybrid skills, that is, the combination of technical and human skills. Higher education leaders may be underestimating the importance of technical skills slightly, but once again it is students who are the most interesting group.

Stacked bar chart comparing perceptions of the skills most valued in the AI era among learners, higher education leaders, employers, and respondents overall. Across all groups, the largest share believes that technical and human skills are becoming equally valuable. Employers (60%) and higher education leaders (58%) are most likely to favor a hybrid skill set, while learners are more divided, with 39% choosing hybrid skills, 35% prioritizing human skills, and 25% technical skills.

Across degree programs and markets, students are less convinced that a hybrid skillset of technical and human skills is expected. Instead, they are fractured in their
beliefs about which skills will be most valuable to the future workplace. In some markets, like the US and the UK, they are more likely to expect that critical human skills alone will be of most value.

Universities spend enormous amounts of time trying to understand what employers want. Perhaps the more urgent question is whether students understand what universities actually offer. If students are less convinced of the value of higher education than employers themselves, we don't have an employability problem; we have a confidence problem.

Revisiting my mid-year resolution, I don't want to stop writing about the graduate labor market or the misalignment between higher education and employers. What I do want to stop is treating employers as the sole authority on educational value. That already happens far too often in the literature.

The myth of independent job applications

Graduates finding employment is an increasingly important aspect of student success. But we tend to assume two things about job applications. First, that every employer evaluates applicants independently. Second, that sending more applications improves your chances. A fascinating new article suggests both assumptions may be wrong.

Using data from more than four million job applications, researchers examined hiring decisions made through a widely used AI screening platform. Most research on AI in hiring asks a straightforward question: Is the algorithm biased? This study asks a different one: What happens when hundreds of employers rely on the same algorithm?

Flow diagram showing a four-stage AI hiring process. Applicants apply for multiple jobs, complete a single assessment on the Pymetrics platform, and are evaluated by AI models before employers decide whom to interview or hire. The diagram illustrates how multiple employers can rely on the same AI platform to screen applicants, resulting in similar hiring recommendations across different jobs.

Rather than looking only at aggregate results, the researchers examined hiring decisions job by job. They found that around 10% of jobs showed practically and statistically significant disparities in selection rates for Black applicants. Although employers used different hiring models, many were built on the same underlying platform. As a result, applicants were more likely to receive similar decisions across employers.

The consequence is striking. The researchers estimate that applicants may need to submit roughly 25 applications instead of 10 to achieve the same chance of receiving at least one recommendation. The reason is simple: many hiring decisions are no longer fully independent.

That's a profound shift in how we think about job searching. The challenge isn't just competing for jobs—it's navigating an increasingly interconnected ecosystem of AI-mediated hiring.

There are important limitations. The study examines one vendor that uses game-based assessments rather than the broader universe of AI résumé screening tools, so we shouldn't assume the same pattern exists everywhere. But it does raise an important question: what happens as more employers adopt increasingly similar hiring technology?

Every application feels like a new opportunity. Increasingly, it may simply be another opinion from the same AI. As AI becomes more deeply embedded in hiring, we may need to rethink some of our oldest career advice. Just keep applying assumes every application is a fresh opportunity. Increasingly, that may no longer be true.

Overlaps

I love a good Venn diagram. This one makes me chuckle even though the British now swear loyalty to a King not a Queen.

Humorous Venn diagram comparing British people and ants. The overlap lists "weird loyalty to their queen," "innate instinct to line up single-file and travel in a queue," "takes things that aren't theirs back to their colony," and "love for crumbly pastries." The British-only section says "beans and toast," while the ants-only section says "can lift up to 5,000 times their own weight."

Though I feel obliged to point out that historically the British took things from the colony not to it.

When saying less helps more

I loved this post about the practice of purposeful restraint in coaching. Whether we're coaching, advising, teaching, managing, or mentoring, it's often tempting to jump in with feedback to help a player, student, mentee, or employee avoid making the same mistake.

And sometimes that is exactly what is needed. But sometimes it isn’t.

Sometimes the player already knows.
Sometimes they are too emotional to hear it properly.
Sometimes the answer is right, but the timing is wrong.
And sometimes another sentence, even a good one, would only add noise.

That part took me time to understand, because coaching is not only about whether you know what to say. It is also about whether saying it now will actually help the player and those two are not the same thing.

A coach can be right and still be unhelpful. That is one of the more annoying truths of the job.

The secret to all this is timing, understanding the player or student or employee and in having their trust.

There is a timing to truth:
Say it too late, and maybe you miss the moment. Say it too early, and maybe the player rejects something they might have understood later.

That is why coaching is not just about information. If it were only information, it would be much easier. You would see the problem, say the problem, fix the problem, move on.

But players are not machines. They are human. They have pride, fear, history, sensitivity, tiredness, blind spots, defense mechanisms, and a complicated relationship with their own ambition.

So sometimes the work is not to say more. Sometimes the work is to wait until the player can actually hear.

And this is where trust becomes important. Because if there is no trust, silence can feel like abandonment. If there is trust, silence can feel like space.

Which makes it sound hard—and it is. It's also easy to mistake passivity for purposeful restraint in any of the roles I've mentioned.

To me, there's an obvious student success lesson here. Student success isn't just about providing support; it's about providing support when students are ready to use it. Timing may be just as important as the intervention itself.

Kentucky - the grass is always more blue

Some good news from Kentucky

Bar chart comparing the percentage of bachelor's degrees awarded to low-income students at Kentucky's eight public universities in 2020–21 and 2024–25. Five of the eight universities increased the share of degrees awarded to low-income students over the period, with the largest gains at Kentucky State University (72.1% to 78.8%), the University of Louisville (40.3% to 43.4%), and Eastern Kentucky University (51.9% to 55.5%). Three institutions saw small declines.

Musical Coda

Staying in Kentucky.

Know someone whose assumptions could use a little realignment? Forward them this newsletter.

Thanks for being a subscriber.