Quick answer: An education dissertation on online learning is defensible when you narrow it to one kind of online learner, one design, one instrument you can name and one source of access. “Online learning” is not a variable; it is a setting, so your study has to say what about it you are measuring: students’ sense of presence in a course, their motivation or self-regulation, their persistence, or the effect of a specific online practice. This guide walks through the populations, designs, instruments and access routes that fit an education dissertation on online learning, with illustrative hypotheses and ten researchable titles.

Online learning is a high-volume topic in education research, which means committees have read many dissertations on it and expect precision. If you are choosing among topics first, the list of education dissertation topics with research questions is a good place to compare options. This article is for the next step: turning “online learning” into a bounded study.

Pin down what “online learning” means in your study

Define the setting in your first chapter, and define it in terms a reader can verify. A fully online degree program, a single online course taught asynchronously, a synchronous video class, a blended course and a virtual secondary school are different settings with different students, and results from one do not transfer automatically to another. Say how much of the instruction was online, whether it was planned as online instruction from the start, who taught it, and what platform was used. If your data come from a period when courses moved online at short notice, say so, because students in a course designed for online delivery and students in a course moved online quickly are not the same population.

The reason this matters is that examiners test your claims against your definition. A thesis that studies one online graduate program and concludes that “online learning improves outcomes” has generalized from a setting to a category.

Choose one population

Pick the learner group whose experience your question is about, and keep it narrow enough that you can reach them.

• Undergraduates in online courses at one institution. Easiest to access through a course or department, and the group most often studied.
• Adult and working learners in online degree programs. Their time constraints, motivation and persistence differ from traditional-age students.
• Graduate students in fully online programs. A natural fit for questions about presence and community.
• Secondary school students in virtual or blended settings. Requires parental consent and school-level permission, which changes the timeline.
• Teachers or instructors of online courses. Shifts the question to design, workload and support, and avoids many consent complications that come with minors.

For the consent and gatekeeper steps that apply when your participants are in schools, see the guide to population, sample and access for an education dissertation.

A widely used framework for online learning dissertations: Community of Inquiry

A framework examiners often expect to see is the Community of Inquiry, introduced by Garrison, Anderson and Archer (1999) in “Critical Inquiry in a Text-Based Environment: Computer Conferencing in Higher Education,” The Internet and Higher Education, 2(2-3), 87-105. It describes an online learning experience through three presences: cognitive presence (constructing meaning through communication and reflection), social presence (the ability to identify with the community and develop relationships) and teaching presence (course design, organization and instructor guidance).

It comes with a survey. Arbaugh and colleagues (2008) published “Developing a community of inquiry instrument: Testing a measure of the Community of Inquiry framework using a multi-institutional sample,” The Internet and Higher Education, 11(3-4), 133-136. The Community of Inquiry website states that the survey is an open resource under a Creative Commons license, downloadable in English as PDF and Word, with a Spanish version online and a German version available on request, and that users must give appropriate credit and share modifications under the same license. That makes it unusually easy to use in a student project, but you still need to read the current license text and cite it correctly.

Other frameworks and instruments, with their limits

The Community of Inquiry describes the learning environment. If your question is about the learner, pair it with a learner-level theory and instrument. Two options with documentation you can open:

• Self-regulation and motivation. The Motivated Strategies for Learning Questionnaire (Pintrich, Smith, Garcia and McKeachie, 1991) has 81 items in 15 scales that can be used together or singly, rated on a seven-point scale, and it was designed for a specific course. Its modular scales let you administer only what your hypotheses need. The thesis on academic self-efficacy and the MSLQ shows how to describe it accurately.
• Self-determination theory measures. The Center for Self-Determination Theory lists questionnaires including the Learning Climate Questionnaire and the Basic Psychological Need Satisfaction and Frustration Scale, and states that they may be used for academic, non-commercial research projects. They suit questions about autonomy and need satisfaction in online courses.

For a broader menu of instruments and the reporting expected for each, the guide to validated instruments for an education dissertation compares measures at the student, teacher, classroom and school levels. Whatever you pick, report where the items came from, the version, the number of items, the response scale and the reliability in your own sample.

Designs that fit an online learning dissertation

• Cross-sectional survey of online students. Measure presence, motivation or satisfaction and relate them to an outcome such as persistence intention or course grade. Quick to run, limited in what it can prove.
• Quasi-experimental comparison. Compare an online practice (for example structured peer discussion) against ordinary teaching in comparable sections, controlling for prior attainment. Random assignment is rare in education settings, so describe how groups differ.
• Pre-post design. Measure a construct at the start and end of an online course, ideally with a comparison group.
• Learning analytics study. Use platform logs such as log-in frequency or time on task, which an institution may release in de-identified form. Define each log measure precisely and say what it can and cannot represent.
• Qualitative or mixed-methods study. Interview students or instructors to explain what the numbers hide.
• Synthesis. A review or meta-analytic approach, building on work such as Means, Toyama, Murphy and Baki (2013), “The Effectiveness of Online and Blended Learning: A Meta-Analysis of the Empirical Literature,” Teachers College Record.

Whichever design you pick, match the statistical test to it. The guide to choosing a statistical test for an education dissertation covers the usual choices for classroom and school data, including the problem of students nested within classes.

Illustrative hypotheses you can adapt

These hypotheses are illustrative and written for a generic online course. Replace the population, the course and the outcome with your own.

• H1: Students’ ratings of teaching presence are positively associated with their ratings of cognitive presence.
• H2: Social presence is positively associated with students’ intention to enroll in a further online course.
• H3: Students in sections using structured peer discussion report higher social presence than students in sections without it.
• H4: Self-efficacy for learning and performance predicts final course grade in an online course after controlling for prior grade point average.
• H5: Time on task in the learning platform is positively associated with assessment performance, and the association is weaker for students who report competing work hours.
• H6: Working adult learners report lower perceived autonomy support than traditional-age learners in the same program.

Each names a variable, a direction or contrast and a population. Avoid hypotheses such as “online learning is effective,” which cannot be tested without a defined comparison.

Access: where your data actually come from

Most online learning dissertations succeed or fail on access. There are three routes. The first is through a course or program: an instructor or program director agrees to let you invite students. The second is through institutional data: an office of institutional research or an online learning unit releases de-identified records or platform data under an agreement. The third is through public data. In the United States, the National Center for Education Statistics describes its IPEDS Fall Enrollment component as collecting distance education enrollment information as part of overall enrollment data, and the IPEDS Data Center offers custom data files in CSV format for selected institutions and variables. Public data of this kind describe institutions, not learners, so they fit questions about patterns across institutions rather than about individual students. The guide to US education datasets for a dissertation lists other public sources.

Whichever route you use, plan the ethics application early. Your review board will ask how participants are recruited, how consent is documented, how anonymity is protected and where data are stored. If you plan to link survey responses to platform logs or grades, say so in your application, because linking changes the privacy risk. If you are an instructor studying your own students, say how you will avoid coercion, for example by having someone else collect and hold identifiable data until grades are submitted.

Ten researchable titles

1. Teaching presence and cognitive presence in an online graduate education program: a Community of Inquiry survey
2. Social presence and persistence intention among working adults in fully online degree programs
3. Structured peer discussion and social presence in online undergraduate sections: a quasi-experimental study
4. Self-efficacy for learning and course grade in an online introductory statistics course
5. Autonomy support and need satisfaction in online courses: a self-determination theory study of adult learners
6. Learning platform activity and assessment performance in an online teacher-education course
7. Instructor workload and design decisions in online teaching: a qualitative study of faculty at one university
8. Virtual secondary school students’ experience of teaching presence: a mixed-methods study
9. Do online and in-person sections of the same course differ in student engagement? A comparison with prior attainment as a covariate
10. Peer interaction and course completion in asynchronous online courses: an analysis of discussion participation

Common mistakes that cost marks

• Treating “online learning” as a variable instead of a setting, so the research question has no measurable construct.
• Generalizing from one course or program to all online education.
• Using a survey instrument without naming the version, the number of items or the license terms.
• Reading platform activity as engagement without defining what each log measure represents.
• Concluding that online delivery caused an outcome from a design with no comparison group.
• Collecting data from your own students without explaining how coercion is avoided.

Where Tesify fits

When your setting, instrument and design are fixed, Tesify’s thesis workspace helps you structure and organize your thesis, and 9,000+ students have written 15,000+ chapters with Tesify. Every word stays 100% written by you, and Tesify cannot obtain your institutional access, ethics approval or participants.

Frequently asked questions

Which framework is widely used for an online learning dissertation?

The Community of Inquiry framework by Garrison, Anderson and Archer (1999), which describes online learning through cognitive, social and teaching presence, is a common choice.

Is the Community of Inquiry survey openly licensed?

The Community of Inquiry website states that its survey is an open resource under a Creative Commons license, downloadable in English, with a Spanish version online and a German version on request, so read the license and give the credit it requires.

Which design is best for comparing online and in-person teaching?

A comparison of comparable sections with prior attainment as a covariate is usually the most practical, and you should describe how the groups differ because random assignment is rare.

Can I use learning platform data in my dissertation?

Often yes, through your institution’s research or online learning unit, but you need permission, usually de-identified data and an ethics review, and you must define each log measure and its limits.

Where can I find public data on online enrollment in the United States?

The National Center for Education Statistics describes distance education enrollment as part of its IPEDS Fall Enrollment component, and the IPEDS Data Center offers downloadable data files, though these describe institutions rather than individual learners.

Do I need ethics approval for a survey of online students?

Usually yes, in some form, and your institution’s review board decides the level of review, so submit the survey, consent statement and data plan before recruiting.

How many online students do I need?

It depends on the analysis you plan, so run a power calculation for your chosen test and report it in the methods chapter instead of relying on a rule of thumb.

Can I study my own online class?

You can, but you must show how students are protected from pressure to participate, for example by having someone else collect identifiable data and by keeping responses separate from grading.