So America’s community college are working on a new strategy to improve graduation rates. It’s pretty… interesting.

According to this piece at the Chronicle of Higher Education:

A project known as Connect2Complete provides low-income students who are in remedial classes with peer mentors, many of whom have been through the program themselves. During the first few weeks of the semester, the students, mentors, and instructors participate in a service-learning project that is connected to their coursework.

Recently, for instance, math students at Owens Community College, in Ohio, worked at a food bank, charting and graphing the number of boxes that could be filled with the available supplies. At other campuses, English students recorded materials to read to local blind children or helped non-native speakers improve their English-language skills.

Students helped tailor the community-service projects to their interests. The students then wrote about their experiences. One of the biggest benefits, according to an administrator at Owens, is a boost in the students’ self-esteem.

Well good luck with that.

Retention at American community colleges is infamously quite terrible. About a quarter of students who start community college do not even show up for the second semester.

And so colleges have worked on a few things to try and improve completion rates.

There was the coach the poor to success plan. Some others have proposed reducing the number of credit hours students need to earn and reforming college remediation programs. And now we’re trying self esteem training.

This isn’t likely to work very well. According to the article, researchers at the Center for Youth and Communities at Brandeis University, looked into the effectiveness of the program: “After six semesters, the persistence rate among participating students was 32 percent, compared with 26 percent in the comparison groups.”

That’s not nothing, and promoting community service is a good thing, because it really does improve community connections and provide free labor, but a 32 percent persistence rate is still really terrible. The program works, but it doesn’t work very well.

American elementary and secondary schools have also been working on self esteem issues over the last 30 years or so. It makes some degree of sense, because when you’re confident you can succeed, you’re more likely to do well. But self-esteem training hasn’t done much to improve learning for America’s younger students. It’s not likely to do much for college students either.

That’s because self-esteem is a byproduct of achievement. You can’t inject self-esteem into people if they don’t realistically have much reason to be confident about their chances for success.

So what might work better? The article quotes Krista Kiessling, director of service learning, civic engagement, and leadership at the Ohio community college: “‘The students who are most likely to be at risk for not sticking around are usually first-generation, Pell-eligible students who are often the recipient of services’ like food stamps or subsidized housing.” Yes, of course. This is true of struggling college students everywhere in this country. “Putting them in situations in which they are the ones providing services is empowering in all kinds of ways,” Kiessling.

What would be more “empowering”–as in give them the actual power to succeed–is cheaper college. The primary reason students leave school is financial. People don’t drop out of college because it’s too hard or they lack confidence; they drop out of college because it’s too expensive. If community colleges want to improve completion they’re going to have to address the root cause of low completion. Working on cutting community college costs would surely be more expensive than trying to improve self esteem, but it would also be much more effective.

Daniel Luzer

Daniel Luzer is the news editor at Governing Magazine and former web editor of the Washington Monthly. Find him on Twitter: @Daniel_Luzer