Technology Acceptance Model: Constructs, Topic Ideas and Framework Design

Reader outcome

This guide helps a student use the Technology Acceptance Model as a research framework rather than merely placing two familiar constructs in a questionnaire. It explains the original model’s core ideas, shows how to move from context to research question and hypotheses, and identifies limits that must be acknowledged when adapting the model.

The original contribution

Davis’s 1989 study developed and validated measures for two beliefs used to explain user acceptance of information technology:

• Perceived usefulness (PU): the degree to which a person believes that using a system would improve relevant performance.
• Perceived ease of use (PEOU): the degree to which a person believes that using the system would require relatively little effort.

The original research treated these beliefs as fundamental determinants of acceptance and reported strong psychometric results for the developed scales. The paper’s setting, technology and samples belong to a particular historical context. A modern thesis should use the original paper to define the constructs, then verify whether later theory and context-specific evidence justify the chosen model.

Do not confuse original TAM with every later extension

Many studies add trust, perceived risk, social influence, facilitating conditions, self-efficacy, enjoyment or domain-specific beliefs. These may be useful, but an extended model is no longer simply the original TAM. Every added construct needs a theoretical reason, a clear definition and a defensible relationship to the outcome.

Avoid creating a “kitchen-sink” framework in which every convenient questionnaire variable points to behavioural intention. A model should explain a focused problem, not maximise the number of hypotheses.

Develop a topic from a real acceptance problem

Start with a context where adoption, continued use or effective use is genuinely uncertain. Define:

1. User group: for example, first-time users, employees, patients, teachers or small-business owners.
2. Technology: a specific system or service, not “technology” in general.
3. Behavioural outcome: intention to use, actual use, continued use, feature adoption or another clearly measured behaviour.
4. Decision period: initial trial, mandatory implementation, post-adoption continuation or switching.
5. Problem: what practical or theoretical uncertainty makes the study worth conducting?

A focused topic might be: “Perceived usefulness, perceived ease of use and continued intention to use a university learning analytics dashboard among postgraduate instructors.”

Build the conceptual framework

A basic teaching framework can begin with:

• PEOU → PU
• PU → behavioural intention
• PEOU → behavioural intention
• behavioural intention → use behaviour, where actual use is measured appropriately

The exact paths must follow the version of TAM and the sources selected for the thesis. Do not add actual-use paths when the study measures only stated intention. Do not label intention as adoption.

External variables can influence PU and PEOU. Examples might include training quality, system compatibility or task relevance. They should enter the model only when theory explains why they shape a core belief.

Write research questions and hypotheses

A research question should connect constructs to the defined population and technology. Example:

To what extent are perceived usefulness and perceived ease of use associated with postgraduate instructors’ intention to continue using the university learning analytics dashboard?

Illustrative hypotheses:

• H1: Perceived usefulness is positively associated with continued-use intention.
• H2: Perceived ease of use is positively associated with continued-use intention.
• H3: Perceived ease of use is positively associated with perceived usefulness.

These are examples, not universal hypotheses. Direction, causal wording and analysis must match the design. A cross-sectional survey can estimate associations consistent with a model; it usually cannot establish temporal causation by itself.

Measure the constructs responsibly

Use the original source and validated context-specific literature to identify candidate items. Adapt wording only when necessary for the technology and population. Maintain the construct’s meaning, document every change and conduct cognitive or pilot testing.

Do not combine PU and PEOU items into one score merely because they are correlated. Assess dimensionality, reliability and validity using a method appropriate to the study. Translation requires forward adaptation, expert review and testing; literal translation alone does not establish measurement equivalence.

Common response biases include acquiescence, social desirability, common-method variance and restricted response ranges. Procedural remedies—clear anonymity, separated measurement, neutral wording and suitable scale design—should be considered before statistical remedies.

Choose an analysis that matches the model

A small framework may be tested with regression when measurement quality has already been justified and the study’s aim is modest. Latent-variable structural equation modelling can estimate measurement and structural components, but it introduces sample-size, identification and model-evaluation requirements. The software should follow the method, not determine it.

Report effect estimates, uncertainty and model diagnostics. A significant path does not automatically mean the model has strong explanatory or predictive performance. Compare plausible alternatives and acknowledge omitted constructs.

Topic-development matrix

The matrix generates questions; it does not validate the extensions.

Common mistakes

• Calling any technology survey “TAM” without measuring the core constructs.
• Copying items from secondary studies without tracing them to validated sources.
• Adding constructs only to increase novelty.
• Treating behavioural intention as actual use.
• Using causal language for a single cross-sectional questionnaire.
• Reporting only Cronbach’s alpha and ignoring construct validity.
• Ignoring whether use is voluntary or mandatory.
• Presenting the original 1989 relationships as universally complete for every technology and culture.

Source and verification notes

The factual core is based on Davis’s 1989 MIS Quarterly article and its official AIS/MISQ records. This guide distinguishes the original contribution from later extensions and uses original PT Writers examples. A study using a specific TAM variant or application domain should add primary sources that directly support that variant and context.