TAM vs UTAUT: Which Technology Acceptance Model for Your Dissertation? (2026)

If your dissertation asks why people adopt — or refuse — a technology, two named models dominate the field: the Technology Acceptance Model (TAM) and the Unified Theory of Acceptance and Use of Technology (UTAUT). Supervisors will assume you know both; examiners will expect you to justify choosing one. This comparison sets out what each model actually claims, where each fits, and how to make the choice defensible in a methodology chapter — the model-specific decision that generic framework guides, including our own on theoretical vs conceptual frameworks, deliberately leave open.

The two models in one table

What each model actually claims

TAM is radical in its economy: whether people accept an information technology comes down to two beliefs — that it would enhance their performance (perceived usefulness) and that using it would be free of effort (perceived ease of use). Two constructs, a handful of questionnaire items each, and four decades of studies to compare against. That parsimony is why TAM survives: it fits a dissertation-sized sample, and its measures are so well-worn that instrument design becomes adaptation rather than invention — with the re-validation duties that adaptation carries, covered in our Likert questionnaire guide.

UTAUT was built as a consolidation. Venkatesh and colleagues reviewed the competing acceptance models of the 1990s and condensed their 32 variables into four main effects — performance expectancy (the technology helps me achieve gains), effort expectancy (it is easy to use), social influence (people who matter to me think I should use it), and facilitating conditions (the infrastructure and support to use it exist) — with four moderating factors: gender, age, experience and voluntariness of use. The first two constructs are TAM’s pair renamed and broadened; the second two are what TAM leaves out. If your research question involves whose adoption differs and under what conditions — students vs staff, mandated vs voluntary systems, first-years vs finalists — UTAUT’s machinery exists precisely for that.

How to choose, by research design

• Small sample, single user group, voluntary use (e.g. “do students accept this revision app?”): TAM. Fewer constructs means fewer items, a smaller required sample for the same statistical power, and a cleaner story.
• Organisational or mandated context (e.g. “why do clinicians resist the new records system?”): UTAUT. Facilitating conditions and social influence are usually where the findings live in workplace studies, and voluntariness moderation is built for mandated-use settings.
• Comparative user groups (age bands, experience levels): UTAUT, because the moderators are the hypothesis.
• Qualitative designs: either model can serve as a sensitising framework for interview coding — but say so explicitly; neither model is “tested” by twelve interviews, and claiming a test invites the examiner’s easiest kill.

Whichever you pick, the justification pattern examiners reward is the one our guide to building a theoretical framework step by step recommends for any theory: name the alternatives you considered and say why you rejected them. For this pair the honest sentence usually writes itself — parsimony versus contextual coverage.

The traps in using either model

• Citing the wrong origin. TAM is Davis (1989) in MIS Quarterly; UTAUT is Venkatesh et al. (2003) in MIS Quarterly. Resolve both by DOI (10.2307/249008 and 10.2307/30036540) rather than trusting a secondary source’s reference list — miscited foundational papers are among the most common reference-list errors examiners notice.
• Kitchen-sinking. Bolting UTAUT’s constructs onto TAM “for completeness” produces a model that is neither, with no published psychometric history. Extensions exist in the literature; if you extend, cite the specific extension you are following, not a homemade hybrid.
• Measuring constructs with invented items. Both models come with validated item sets in their origin papers and decades of adaptations. Write your questionnaire from those, adapt minimally, and report reliability for your version.
• Forgetting the dependent variable. Both models predict intention and use. If you can measure actual use (logs, records) rather than self-reported intention, your dissertation immediately outclasses the intention-only average — and if you cannot, say plainly that intention is your outcome.
• Treating model fit as the finding. “UTAUT fitted the data” is a manipulation check, not a contribution. The contribution is what the coefficients say about your population and technology, and what that implies for the people who must act on it.

A worked sketch: UTAUT on a concrete question

Suppose the question is “what predicts postgraduate students’ use of AI writing assistants for coursework?” A UTAUT operationalisation would run: performance expectancy — items on whether the tool improves the quality and speed of academic work; effort expectancy — items on how easy the tool is to learn and use; social influence — items on whether supervisors, peers and departmental policy encourage or discourage use (in this setting, often the most interesting construct, because the social signal is genuinely mixed); facilitating conditions — access, training, and whether the institution’s rules make legitimate use practicable. The moderators earn their place naturally: experience (first exposure vs habitual users) and voluntariness (free choice vs course-mandated tools) are live distinctions in this population, and age and gender complete the specification. The dependent variable should be stated honestly — self-reported use frequency, unless you can obtain actual usage records.

Now run the same question through TAM: perceived usefulness and perceived ease of use, two scales, one regression. If your sample is one cohort using one tool voluntarily, TAM answers the question with half the questionnaire length. The worked contrast is the justification paragraph, nearly verbatim: UTAUT if the mixed social signal and mandated-vs-voluntary distinction are the point; TAM if they are noise.

Writing the framework chapter around your choice

Structure the section in four moves. First, situate the problem: technology adoption in your specific setting. Second, introduce the candidate models with their origin citations and core claims — the table above is one paragraph each. Third, choose, with the design-based reasoning from the list above; the choice belongs to your question, not to the models’ fame. Fourth, operationalise: name each construct, the items measuring it, their source, and any adaptation. That final move is where the framework stops being decoration and becomes the analysis plan — the transition that separates a framework chapter from a name-drop, as our framework comparison guide puts it, between the elevation blueprint and the floor plan.

Frequently asked questions

What is the difference between TAM and UTAUT?

TAM (Davis, 1989) explains technology acceptance through two beliefs: perceived usefulness and perceived ease of use. UTAUT (Venkatesh et al., 2003) synthesises eight earlier models into four predictors — performance expectancy, effort expectancy, social influence and facilitating conditions — moderated by gender, age, experience and voluntariness.

Which is better for a student dissertation?

Neither is “better”; they trade parsimony against coverage. TAM suits small, single-group, voluntary-use studies; UTAUT suits organisational, mandated or group-comparative designs where context and user characteristics carry the question.

Can I use TAM or UTAUT in qualitative research?

As a sensitising framework for interview design and coding, yes — state that explicitly. What a qualitative design cannot do is statistically test the model’s paths, so frame the contribution as exploring how the constructs manifest, not as a model test.

How many participants do I need to test these models?

It depends on the analysis (regression or structural equation modelling), the number of constructs, and the effect sizes you expect — run a power analysis rather than borrowing a rule of thumb. UTAUT’s larger construct count and moderation tests demand meaningfully more data than TAM’s two predictors.

Do I have to use the original questionnaires?

Use the validated items from the origin papers or published adaptations as your base, adapt wording minimally to your technology and population, and report reliability for your adapted version. Inventing items from scratch discards the models’ main practical advantage.

Is UTAUT just TAM with extra steps?

No — performance and effort expectancy do descend from TAM’s constructs, but social influence, facilitating conditions and the four moderators come from the other synthesised models, and they are usually where organisational studies find their results.

From model choice to methodology

The model is a commitment: constructs, items, sample size and analysis all follow from it. Tesify helps you carry that commitment through the framework and methodology chapters without losing the thread — 100% written by you.

Build your dissertation with Tesify