What an enrolment number counts
Reports on online education often lead with a large enrolment figure. It sounds like a measure of how many people are learning, but it usually measures something narrower: how many people pressed a button to register. Registration can be free, instant and made out of curiosity, so it says little about what happens afterwards.
The first thing to check is the unit. A figure might count unique people, or it might count enrolments, in which case one person signing up for five courses appears five times. It might count accounts, which can include duplicates and abandoned profiles. Some figures describe a cumulative total across many years, while others describe a single term or year. Without that detail, two numbers that look comparable may not be.
Enrolled, started, finished
A useful way to think about any course is as a funnel with several stages:
- Registered, meaning an account or enrolment exists.
- Started, meaning the person opened the first lesson or activity.
- Active, meaning they kept returning over some period.
- Completed, meaning they reached the end or met a stated requirement.
- Learned, meaning they can do or understand something they could not before.
Published figures usually describe the first or second stage, because those are the easiest to record. The last stage is the one people care about and is by far the hardest to measure. Moving down the funnel, the numbers typically shrink, and the shrinking is a normal feature of open, low-cost courses rather than a scandal.
Why completion rates are tricky
Completion rates are often quoted as a sign that online courses fail, but they are not always a fair yardstick. People enrol for many reasons. Some want to read one lecture, check whether a topic interests them, or download materials. Counting those people as “dropouts” treats a deliberate choice as a failure.
Completion also depends on how it is defined. A course might count someone as finished if they watched every video, passed all quizzes, or simply reached a final page. Different definitions on different platforms make comparison difficult. Course length, cost, whether a certificate is offered, and whether the learner is paying or has a deadline typically all affect who completes.
There is also the question of who enrols. Learners who already have experience of study, strong motivation or spare time may be over-represented among those who finish. This kind of selection effect means a high completion rate may say as much about who signed up as about how good the course is.
Why sign-ups overstate learning
Several things can push enrolment numbers above the amount of learning that takes place:
- Low cost of entry. Free or cheap courses invite casual sign-ups, and people often register for more than they can complete.
- Bundled access. If a course is included in a subscription, an employer programme or a school licence, learners may be enrolled automatically.
- Promotional counting. Organisations that want to show growth have an incentive to use the largest defensible figure, such as cumulative registrations.
- Time lag. Someone who registers during a surge of interest may not return for months, if at all.
None of this means that online learning is ineffective. It means that registration is a measure of interest and access, not of outcomes.
What measuring learning would require
To say that people learned something, researchers usually need more than platform logs. They might use assessments taken before and after a course, comparisons with similar learners who took a different route, or follow-up some time later. Even then, it is hard to separate the effect of the course from the learner’s prior knowledge and motivation. This is a classic case of confounding variables: motivated learners both choose more courses and do better, so the course may not be the only reason for their results.
Self-reported measures such as satisfaction surveys are easier to collect, but they capture how people felt, and that may differ from what they can do. Employers’ and learners’ own accounts of benefit are valuable but can be shaped by hindsight. The gap between what people say and what they do is a general theme in the study of self-reported behaviour.
Success stories and the missing majority
Stories about someone who changed careers after an online course are real and can be inspiring, but they are a selected sample. The many people who did not finish, or who finished without any change in circumstances, are much less likely to be profiled. This is a form of survivorship bias, and it can make online education look more transformative, or more easily repeatable, than the typical experience.
Comparing online and in-person study
Comparisons between online and classroom learning are difficult for similar reasons. The learners who choose each route often differ in age, work commitments, prior education and location. A difference in outcomes may reflect those differences, not the format. Courses also vary enormously: a live, tutor-supported programme and a self-paced set of recordings are both called online education but may have little in common.
Results also depend on context. Remote study can be a practical way for people who live far from a campus or who work full time to take part, even if overall completion is lower. Whether that trade-off is acceptable depends on what the learner wanted.
Questions to ask of a headline figure
When a report says that online education is growing, a few questions help put it in context:
- Does the number count people, enrolments or accounts?
- Is it cumulative or for a specific period?
- How does the source define completion?
- Who is included: free learners, paying students, employees on required training?
- Is anything said about learning outcomes, or only participation?
- Who produced the figure, and what might they gain from a larger one?
Growth in enrolment can be genuine and still tell us little about how much people learned. Treating sign-ups as the start of a story, not its conclusion, is usually the more careful reading. How the platforms behind these courses are built and judged is a separate matter, and it is covered under e-learning platforms.