Until recently, data limitations have stymied research on online college programs, especially in the nonprofit four-year sector of higher education. Substantial growth in online education over the last several years has now made it possible to identify, with greater precision, who attends college exclusively online and how attending online may shape crucial educational outcomes.

In this commentary, we present key takeaways from our analyses in “Predatory Inclusion in Non-Profit and For-Profit Online Education,” a paper currently under review for academic publication. Those analyses triangulate findings from multiple datasets. The data come from the 2012–2019 Integrated Postsecondary Education Data System (or IPEDS), which includes aggregate, institutional-level information on very recent college entry cohorts and covers the bulk of U.S. four-year higher education institutions; and a sample of four-year students in the 2012–2017 Beginning Postsecondary Students Longitudinal Study (or BPS), which offers insight into student-level experiences and outcomes.

Evidence from both data sources suggests that online four-year college programs have a negative impact on student completion and student loan repayment outcomes, and that groups traditionally underrepresented in higher education are concentrated in this potentially problematic online modality. These troubling patterns arise in both for-profit and nonprofit online four-year education.

Specifically, we see a greater concentration of Black students and Pell grant recipients in exclusively online four-year education (henceforth called online education, for brevity). The concentration of Black students and Pell grant recipients sits alongside evidence of worse completion for students enrolled in online education, relative to similar peers attending online. Repayment outcomes are also poor, even though being enrolled in online programs is associated with accruing less student debt overall. Students enrolled online may have less student debt, but they fare worse in repayment than similar students who are attending in-person.

Who Attends College Online?

It is useful to first understand the distribution of exclusively online enrollments across nonprofit and for-profit four-year colleges and universities. We use IPEDS enrollment data, starting from 2012, the first year that IPEDS asked colleges to report on online enrollments. Immediately visible in the figure below is the rise in enrollments at nonprofits, especially public institutions. Online students make up a smaller share of the nonprofit sector: in 2019, 13 percent of nonprofit students were online versus 40 percent in the for-profit sector. However, the sheer size of the nonprofit sector means that, in total, there are now far more online students at nonprofits than at for-profits.

Figure 1


Next, we use BPS data to show that Black students in our sample are more likely to attend online programs than other students. Being Black increases the odds of attending an exclusively online program by an estimated 124 percent among students at nonprofits, even when accounting for basic institutional features. Although Black students make up a larger share of online for-profit student bodies versus in-person, for-profit student bodies, the difference between online versus in-person enrollment is most pronounced at nonprofit colleges and universities. This is true even in analyses that hold constant the social class background of students.

Our institution-level data similarly show that institutions with more online students tend to have more Black students. A ten-percentage-point increase in online prevalence is conditionally associated with a 0.8-percentage-point increase in the percent of the student body that is Black. Here, the association is about the same regardless of sector, in contrast with the student-level results that show a stronger association at nonprofits.

BPS data indicate that Pell grant recipients are more likely than non-recipients to attend online, regardless of sector. These data indicate that Pell grant receipt increases the odds of attending an exclusively online program by an estimated 104 percent at nonprofits and 45 percent at for-profits. Comparably, institution-level analyses show that a ten-percentage-point increase in online enrollments is conditionally associated with a 0.4 percentage point increase in the percent of the student body receiving Pell grants—but only in the nonprofit sector.

BPS and IPEDS data indicate greater concentration of Black and/or Pell students in online education. Any differences between BPS findings and IPEDS findings come from measuring aggregate patterns versus individual differences, the greater recency of the IPEDS versus BPS data, or both factors. Consistent findings across datasets highlight a clear pattern, especially in nonprofit online enrollments.

Online Enrollment and Completion

The greater representation of underrepresented students in exclusively online programs could be positively viewed as evidence of increasing access. However, as we will illustrate, attending online is also associated with substantial negative student outcomes.

Our analyses of BPS data indicate that ever attending an exclusively online program decreases the odds of bachelor’s degree completion within six years by an estimated 67 percent in the nonprofit sector, and 60 percent in the for-profit sector. These analyses account for an array of individual, familial, and institutional characteristics that might independently make online enrollment more likely.

The figure below offers a visual representation of completion gaps by learning modality. It shows that while predicted probabilities of four-year degree completion are higher among nonprofit students in our sample than for-profit students, it is the case that within each sector, online students are less likely than other students to complete a degree.

Figure 2

Our IPEDS data provide corroborating evidence: Nonprofits with higher online enrollments have lower graduation rates. In fact, a ten-percentage-point increase in online prevalence is associated with a 1.48-percentage-point decrease in aggregate graduation rates, holding constant institutional factors that independently impact graduation, such as acceptance rate and student body composition.

In our IPEDS analyses, we do not observe a significant difference in completion by modality in the for-profit sector. However, completion is a lagging indicator of academic struggles: first-to-second year retention rates can capture these struggles more immediately. In IPEDS, six-year graduation rate estimates refer to cohorts of students who began between 2012 and 2014. When we look at first-to-second-year retention rates, which include more recent cohorts, we find that online prevalence is related to worse retention rates at both nonprofits and for-profits. Holding constant other institutional factors, a ten-percentage-point increase in online prevalence is associated with a 1.13-percentage-point decrease in first-to-second-year retention at nonprofits, and a 0.68 decrease at nonprofits.

These data suggest that online enrollment, regardless of sector, hinders progress to the degree, such that students would have fared better in in-person programs.

Taken together, these data suggest that online enrollment, regardless of sector, hinders progress to the degree, such that students would have fared better in in-person programs.

Online Enrollment and Student Loans

Our analyses reveal that students who attend online have lower debt burdens. For example, BPS data indicate that six years after the start of college, students in our sample who attended online programs in the nonprofit sector have, on average, around $1,900 less student debt than their in-person peers. In the for-profit sector, the corresponding difference is about $2,500. These differences make sense, as students who attend online may be able to save on cost-of-living expenses. However, cost savings are partly a function of lower rates of completion among those who attend online, as these students simply spend less time in school.

Perhaps more crucially, students attending online have worse loan repayment outcomes.

In BPS data, respondents are asked to indicate if they have ever been delinquent on student loan payments. Attendance in exclusively online programs increases the odds of ever being delinquent by an estimated 50 percent in the nonprofit sector and 40 percent in the for-profit sector. In our sample, of four-year BPS attendees, students who have attended online in the nonprofit sector face an estimated 42-percent increase in the odds of ever going into forbearance. Individual-level repayment analyses account for student and institutional features.

Default occurs when students do not make scheduled student loan payments on federal loans for at least 270 days. Student-level data indicate no significant differences in default rates by modality in either sector; however, BPS offers a short time-frame in which to observe ever experiencing loan default, given that default requires a much longer period of nonpayment than delinquency.

Compared to BPS data, IPEDS data (merged with College Scorecard data, which offer information on student loan outcomes) examine a much wider array of student loan outcomes over a longer period; however, these data aggregate information on whether individuals have a given loan progress status at a moment exactly two years after they enter repayment, so the rates of some outcomes are very low.

Institution-level data indicate that, at nonprofit universities, higher rates of online attendance are associated with higher rates of student default. Indeed, for every ten-percentage-point increase in online prevalence, student loan default rates increase by an estimated 0.2 percentage points. Rates of forbearance, a student loan status in which distressed borrowers are granted a temporary stop or smaller payments, also increase by an estimated 0.87 percentage points with every ten-percentage-point increase in online prevalence. In the nonprofit sector, greater online prevalence is also associated with lower rates of making progress on student loan debt (a 0.7-percentage-point decline for a ten-percentage-point increase in online prevalence) and lower rates of having paid off student loans in full (a 0.4-percentage-point decline). All of these analyses hold constant other institutional factors.

In the for-profit sector, there is evidence of student loan outcome data manipulation. Our data indicate that as online enrollments increase at for-profit universities, rates of default and delinquency decrease while rates of forbearance increase (a 0.37-percentage-point increase for a ten-percentage-point increase in online prevalence). This puzzling set of findings appears to be a function of for-profit schools’ response to the requirement by the U.S. Department of Education that these schools report default rates for three-year cohorts, with a threat of losing Title IV funding eligibility if default rates become too high.

Specifically, for-profits circumvent high default rates by hiring firms that aggressively push borrowers, who might otherwise be delinquent or headed into default, into forbearance. These students often end up defaulting outside of the three-year window considered by the Department of Education, thereby preserving low aggregate-default rates within the reporting window but inflating forbearance rates. Our interpretation of the for-profit repayment data is also supported by the fact that higher online enrollments are associated with significantly lower rates of making progress. In fact, for every ten-percentage-point increase in online enrollment, there is a 0.41-percentage-point decline in rates of making progress.

It should raise a red flag that students in online programs still struggle more than their in-person peers to pay off their debts. Nonprofits may have the most consistently troubling associations between online enrollments and repayment outcomes, but our data point to efforts to cover up problematic student loan repayment outcomes at for-profits.

Preventing Online Predation

Our findings provide novel evidence that online programs: 1) enroll underrepresented students at higher rates and 2) have worse completion and repayment outcomes. Scholars Louise Seamster and Raphaël Charron-Chénier refer to increased access without corresponding benefits as “predatory inclusion.” They define predatory inclusion as “a process whereby members of a marginalized group are provided with access to a good, service, or opportunity from which they have historically been excluded but under conditions that jeopardize the benefits of access.” In this case, online education increases inclusion for students historically excluded from higher education—but at a cost. Indeed, marginalized groups are more likely to face the negative consequences of online education.

Online education increases inclusion for students historically excluded from higher education—but at a cost. Indeed, marginalized groups are more likely to face the negative consequences of online education.

In the nonprofit sector, online predation may be a function of contracts with for-profit online program managers (or OPMs), who provision many exclusively online programs at public colleges and universities. When for-profit companies are involved in nonprofit student recruitment and receive a share of tuition revenue, there are strong incentives to engage in predatory recruitment behaviors and cut corners in student support. A “bundling loophole” in current guidance by the Department of Education allows universities receiving federal aid to hire for-profit OPMs and financially reward them for recruitment. As The Century Foundation has repeatedly argued, the Department of Education could enforce the higher education ban on incentive compensation by rescinding this loophole.

The manipulation of institution-level default rates may obscure potential hazards of for-profit online education. We concur with policy advocates who argue that the Consumer Financial Protection Bureau and Federal Trade Commission should scrutinize and take action against the companies that for-profit universities hire to manipulate their default rates.

We also recommend that the Department of Education apply a holistic set of student outcome metrics, linked to Title IV funding, to online programs in both nonprofit and for-profit sectors. These metrics might include graduation rates and negative repayment outcomes, beyond default rates. The goal would be to flag predatory online programs and make it more difficult for both for-profits and nonprofits to evade sanction for poor student outcomes.

Higher education in the United States is often touted as a fail-safe avenue for social mobility. However, predatory inclusion in online education may mar this promise, and leave students worse off than before they enrolled.