On Student Success
Posts
Student Success Charts

Student Success Charts

Fewer friends, fewer summer jobs, more deadly sins

Glenda Morgan
September 02, 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.

What chart and listicle goodness do I have to share this month?

For education to lead to good jobs there need to be good jobs (obligatory John Burn-Murdoch post)

We spend a lot of time asking whether higher education is preparing students well enough for the labor market. The Financial Times’ John Burn-Murdoch turns the question around. Perhaps we should also ask whether the labor market is producing enough good jobs for the people education prepares.

Six line charts compare inflation- and cost-of-living-adjusted annual earnings in the US, Germany, and UK across degree holders, professional workers, technical and vocational workers, people without degrees, retail and sales workers, and factory workers. US earnings are generally highest across all six groups, while UK earnings are generally lowest.

What makes the chart striking for me is that the US advantage isn't confined to highly educated workers. Americans earn more than their British counterparts among degree holders and professional workers, but also among people without degrees, retail workers and factory workers. Education can create human potential, but the economy determines how much of that potential gets translated into good jobs and wages.

Hollowing out the middle

A fascinating analysis in the Journal of Economic Geography looks at how Americans have become geographically sorted by education. The researchers define high-skill workers as those with a bachelor's degree or higher, and middle-skill workers as those with a high school diploma, some college or an associate degree.

The charts take a little explaining. In 1970 there was a wide spread of local labor markets, but places with lots of high-skill workers also tended to have lots of middle-skill workers—the upward slope. By 2019 that had completely reversed. High- and middle-skill workers are now increasingly concentrated in different labor markets—the very strong downward slope.

Six scatterplots show the changing geographic relationship between high-skill and middle-skill workers in US labor markets from 1970 to 2019. In 1970, places with more high-skill workers also tended to have more middle-skill workers, with a correlation of +0.39. The relationship steadily reverses over time, reaching a strong negative correlation of −0.88 by 2019, as high- and middle-skill workers increasingly become concentrated in different labor markets.

What does it mean for higher education when people with different levels of education increasingly don't even inhabit the same local labor markets? Among other things, it changes both the populations local institutions serve and the labor markets they are preparing them for.

And geography isn't the only way Americans are becoming more socially separated.

Increasingly it includes how often we hang out with others and whom. In 1990 there was almost no educational divide in friendship. By 2024 there was a large one. Americans without a college degree were much more likely to report having no close friends and much less likely to have six or more.

Two charts show a growing education divide in Americans’ close friendships between 1990 and 2024. The share reporting no close friends rose from 2% to 10% among people with a bachelor’s degree or higher, and from 3% to 24% among those with a high school education or less. The share reporting at least six close friends fell from 45% to 33% among college graduates and from 49% to 17% among those with a high school education or less.

And this may be becoming more important. A separate analysis shows that young Americans are spending much more time physically alone. Before COVID, only about 3.5% of 23–29-year-olds spent an entire waking day without being physically with another person. By 2025 that was 9.6%—almost three times the pre-pandemic level

Line chart showing the share of US 23- to 29-year-olds who spent an entire waking day with no in-person contact. The share remained around 2–4% from the early 2000s through 2019, then rose sharply during the pandemic and remained elevated afterward, reaching about 10% in 2025.

Some of that appears related to remote work, but not all of it. Society reopened; young adults haven't completely returned to their old social routines. Put these two charts together and all our conversations about belonging in higher education start to look rather more consequential.

Children of the 1960s

Apparently the 1960s really were far out, man.

Community college enrollment grew nearly five-fold during the decade, a rate of expansion unlike anything before or since. It is a useful reminder of just how much of the higher education system we now take for granted was built during one extraordinary period of expansion.

Line chart comparing public community college enrollment growth during each decade from the 1950s through the 2020s. The 1960s stand out dramatically, with enrollment growing to nearly five times its level at the start of the decade. No other decade approaches that rate of growth; the 1950s roughly doubled, while most later decades saw much smaller increases or declines.

Too much success?

CUNY has succeeded spectacularly in attracting students to computer science. Now it has a different problem: faculty capacity hasn't remotely kept pace.

Line chart showing the number of computer science students per faculty member at five CUNY senior colleges from 2014 to 2024. Student-to-faculty ratios rose substantially at all five institutions as enrollment outpaced faculty growth. By 2024, ratios ranged from 20 students per faculty member at John Jay College to 40 at Queens College, all well above CUNY’s target of roughly 14 students per STEM faculty member.

And this isn't simply a matter of institutions failing to respond. Recruiting and retaining computer science faculty is difficult when colleges are competing with much higher salaries outside academia.

Bar chart comparing changes in computer science enrollment and faculty at CUNY from 2014 to 2024. At senior colleges, enrollment grew 90% while faculty grew only 21%. At community colleges, enrollment grew 34% while the number of computer science faculty fell 22%, showing that faculty capacity has not kept pace with rising student demand.

Growing demand for a program doesn't help students much if institutions can't provide the capacity needed to serve them.

The seven deadly sins of student success

Jobs and the young

Teen employment has fallen enormously over the past several decades. Even after a recent rebound, only 35.5% of 16–19-year-olds worked this summer, compared with well over half for much of the postwar period.

Line chart showing the percentage of US 16- to 19-year-olds employed during the summer and the rest of the year from 1948 to 2026. Summer employment fell from 56.5% in 1948 to 35.5% in 2026, with a particularly steep decline after 2000. Despite a recent rebound, teen summer employment remains well below historic levels.

I'm not sure this is entirely good news for student success. Jobs aren't simply a source of income. They are also one of the places young people learn how to work—showing up, dealing with other people, taking responsibility, and acquiring experience.

And work isn't the only part of young people's lives policymakers are worried about.

Bans don’t work, Part 75

It's early days, but Australia's social media restrictions don't appear to have changed young people's actual use very much yet.

Before the restrictions, 85.9% of children surveyed had used an age-restricted platform in the previous four weeks. Three months later: 81.5%. Statistically significant, perhaps. Transformative, not so much.

Table showing social media use among Australian children ages 10–15 before and three months after age restrictions took effect. Use of any age-restricted platform fell only modestly, from 85.9% in the four weeks before the restrictions to 81.5% three months later. YouTube remained the most widely used restricted platform at 73.6%, followed by TikTok at 34.6%, Snapchat at 27.9%, and Instagram at 25.5%.

China to US re demographic cliff: hold my Tsingtao

China's preschool enrollment has fallen by a third since 2020, from 48 million to 32 million, and nearly 30,000 kindergartens and primary schools closed in 2025. The US demographic cliff suddenly looks rather more like a curb.

Screenshot of a post showing China’s sharp preschool enrollment decline. Preschool enrollment fell 33%, from 48 million children in 2020 to 32 million in 2025. Two accompanying charts show kindergarten numbers and preschool enrollment rising for much of the 2010s before declining steeply after 2020. A second post notes that China closed nearly 30,000 kindergartens and primary schools in 2025.

Things I wish I’d known

I wish someone had given me this list when I first became an analyst. It's nominally aimed at junior strategists, but much of it applies to anyone trying to think critically about a market or practice—which includes most of us working in EdTech and student success.

Screenshot titled “68 tips for junior strategists to add value on the job,” offering practical advice for analysts and strategists. Visible tips include building a story around research, making presentation titles tell the story, reading unusual sources, sharing work early, focusing on what the work should achieve, consuming information widely, and remembering that context matters more than theory.

Envy may be a deadly sin. Sharing isn't. If you enjoyed this post, send it to someone who might enjoy it too.

Thanks for being a subscriber.