What Is a Theoretical Framework? How to Build One in a Quantitative Thesis
What Is a Theoretical Framework
A theoretical framework is the part of your thesis that explains the concepts, theories, and relationships between the variables you study. It gives you the basis for answering three questions: What problem does the study examine, how are its concepts understood, and why do you predict that the variables are related.
In a quantitative thesis, the theoretical framework usually appears in Chapter 2, after the introduction to the research problem and before the methodology chapter. It should not be a sequence of copied definitions. You need to select theories that directly relate to your model, use academic sources to define the variables, and then build the argument for each hypothesis.
You can view the theoretical framework as the following sequence:
foundational theory → research concepts → variables and scales → model → hypotheses → testing with data
For example, if your topic examines intention to use an online learning application, the theoretical framework might present TAM, explain perceived usefulness and perceived ease of use, identify intention to use as the dependent variable, and propose hypotheses about the effects of those two variables. You would then test the model with a questionnaire, SPSS, or SmartPLS.
You can also read the section on scientific research to distinguish a theoretical framework from the entire research process. If you are starting with a broad topic, the article What Is Scientific Research can also help you define your research question and scope before writing Chapter 2.
Why the Theoretical Framework Matters in Quantitative Research
Your theoretical framework determines how you build the model and collect data. When this section is clear, you know which variables are independent, dependent, mediating, or moderating. You also have a basis for selecting indicators instead of putting items in the questionnaire based only on intuition.
First, the theoretical framework gives the concepts a consistent meaning. For example, service quality can be measured in different ways depending on the context. If you do not select a definition and a specific measurement scale source, the indicators in your questionnaire may not measure the same concept.
Second, the theoretical framework explains why a variable belongs in your model. A good hypothesis does more than state that X affects Y. You need to explain the mechanism or reasoning: when X changes, what behavior or perception among the research participants could cause Y to change as well.
Third, the theoretical framework connects your study to earlier research. You may build on TAM from Davis (1989), TPB from Ajzen (1991), UTAUT from Venkatesh et al. (2003), or another suitable model from your verified reference list. Building on an existing model does not mean copying it unchanged. You need to state the context, participants, and variables that you retained, added, or removed.
Fourth, the theoretical framework gives you a basis for explaining the results. When a hypothesis is supported or not supported, you return to the theoretical reasoning and previous studies to interpret the finding. This is especially important when your result differs from the initial prediction.
A theoretical framework is also different from the research object. The research object tells the reader who or what phenomenon you survey. The theoretical framework explains the concepts and relationships used to study that object.
How Much Is Enough for a Theoretical Framework
There is no fixed page count that applies to every thesis. A theoretical framework is sufficient when the reader can understand the concepts, see where the model comes from, follow the reasoning that leads to the hypotheses, and know how the variables will be measured.
The table below is a checklist of reference thresholds. Use figures related to analysis and measurement scales in the correct context. They should not become rigid conditions for every topic.
| Item to check | Reference level or requirement | Source |
|---|---|---|
| Cronbach's Alpha of a scale | Usually accepted at 0.7 or above | (Nunnally, 1978) |
| Cronbach's Alpha of an exploratory scale | May be considered from 0.6 | (Hair et al., 2010) |
| Corrected Item-Total Correlation | At 0.3 or above | (Nunnally and Bernstein, 1994) |
| KMO when running EFA | At 0.5 or above | (Kaiser, 1974) |
| Bartlett's Test | Statistically significant with p-value below 0.05 | (Kaiser, 1974) |
| Factor loading in EFA | At 0.5 or above | (Hair et al., 2010) |
| Total Variance Explained | At 50% or above | (Hair et al., 2010) |
| Observations per indicator | About 5 to 10 observations per indicator | (Hair et al., 2010) |
| CR in PLS-SEM | At 0.7 or above | (Fornell and Larcker, 1981) |
| AVE in PLS-SEM | At 0.5 or above | (Fornell and Larcker, 1981) |
| HTMT | Below 0.85, or below 0.90 for closely related concepts | (Henseler et al., 2015) |
You need to distinguish having enough content from meeting statistical thresholds. Cronbach's Alpha, KMO, AVE, and HTMT are checked after you have data. They cannot show that your theoretical framework is sound if the model lacks logic or the measurement scale source is unsuitable.
For length, prioritize relevance. A short section focused on the four or five variables in your model is usually more valuable than many pages summarizing theories that you do not use. Each theory section should explain what it contributes to your topic.
How to Read the Theoretical Framework Through the Model and Output
A theoretical framework does not appear as a single results table in SPSS. You assess it through the consistency between the conceptual model, coding table, test results, and hypotheses. If Chapter 2 says that X affects Y but the questionnaire contains no indicator measuring X, the model has a broken link.
The following is illustrative output, not the result of an actual study. Suppose you are studying the effects of perceived usefulness and perceived ease of use on intention to use. After running EFA and regression in SPSS, you might read some tables as follows:
| SPSS output table | Statistic or row to read | Illustrative output | How it connects to the theoretical framework |
|---|---|---|---|
| KMO and Bartlett's Test | KMO | 0.812 | The data are suitable for considering EFA according to (Kaiser, 1974) |
| KMO and Bartlett's Test | Sig. of Bartlett's Test | 0.000 | The variables are sufficiently correlated for factor analysis |
| Total Variance Explained | Cumulative % | 63.480 | The extracted factors explain variance at the reference level according to (Hair et al., 2010) |
| Rotated Component Matrix | PU3 on the PU factor | 0.774 | The indicator fits the perceived usefulness concept in the model |
| Reliability Statistics | Cronbach's Alpha of PU | 0.856 | The scale has reference-level reliability according to (Nunnally, 1978) |
| Coefficients | PU → INT, Sig. | 0.003 | There is statistical evidence to consider the hypothesis about the PU and INT relationship |
| Coefficients | PU → INT, Standardized Beta | 0.218 | This shows the direction and relative effect in the illustrative model |
When reading the Coefficients table, do not draw a conclusion from Standardized Beta alone. Check the sign of the coefficient, Sig. or p-value, the confidence interval if available, and how the variables were coded. If the hypothesis predicts a positive effect but Beta is negative, check reverse-coded items, data entry, and the theoretical reasoning.
If you use SmartPLS, the corresponding tables may be Outer Loadings, Reliability and Validity, Discriminant Validity, and Path Coefficients. With PLS-SEM, do not put CFI, TLI, or RMSEA from CB-SEM into the model assessment. The PLS-SEM procedure usually assesses the measurement model first and then the structural model, following (Hair et al., 2022).
What to Do When the Theoretical Framework Is Weak
If your supervisor says that the theoretical framework is not yet acceptable, revise it from the model toward the wording. Do not begin by making Chapter 2 longer or adding large numbers of sources that you never use.
Identify Variables Without a Basis
Create a table with the variable name, definition, theoretical source, measurement scale source, and related hypothesis. If a variable has no clear definition or does not lead to any hypothesis, ask why it remains in the model. You may need to remove it, or you may be missing part of the argument.
Check the Source of Each Concept
Prioritize the original source of the theory and a measurement scale that fits the context. For example, if you study service quality, clarify whether you are adopting SERVQUAL from Parasuraman et al. (1988) or using another approach. Do not take a definition from a blog and present it as an academic concept.
Rewrite the Reasoning for Each Hypothesis
Each hypothesis should have three parts: how concept X is understood, why X may affect Y, and which direction of effect your study predicts. Only then should you write the hypothesis statement. If you cannot predict the direction, review the literature or the way you framed the research question.
Adjust the Model Before Removing an Indicator
When Cronbach's Alpha or EFA does not meet the threshold, do not delete an indicator simply to make the output look better. Check the content of the indicator, reverse-coded items, missing values, and data entry. If you must remove an indicator, record the reason and the run number so you can explain the decision.
A decision with a clear reason and transparent reporting is usually easier for the committee to accept. Adding sources after seeing the results only to justify a relationship without a prior basis will make the discussion weaker.
Distinguishing the Theoretical Framework from the Literature Review
The theoretical framework and literature review are related, but they are not identical. The literature review answers what previous studies have done, which variables they used, in what contexts, and what gaps remain. The theoretical framework answers which concepts and theories your study uses to build its model.
For example, in the literature review, you might compare five studies on intention to use e-wallets. In the theoretical framework, you define intention to use, explain the relevant theory, select the influencing variables, and build the reasoning for H1, H2, and H3. The two sections may use the same sources, but their purposes are different.
A theoretical framework is also easy to confuse with a practical basis. The practical basis describes the actual situation of an industry, business, or participant group. It can explain why the topic deserves investigation, but it does not replace the conceptual definitions and theoretical model.
If your topic uses qualitative data only as a reference, state the methodological scope clearly. What Is Qualitative Research is a separate article about that approach. This article focuses on the theoretical framework for quantitative research using questionnaires and numerical data.
A quick check is to ask what each paragraph is doing: explaining a theory, synthesizing previous studies, describing the current situation, or leading to a hypothesis. If you cannot identify its function, the paragraph may be repetitive or in the wrong section.
Common Errors When Writing a Theoretical Framework
The first error is writing definitions one after another without producing a model. You may have ten concepts, but if you do not explain which variable affects which other variable, the reader still cannot see the research logic.
The second error is using unverified sources or citing the wrong year. Check the original source, publication details, and the claim that the source actually supports. A source defining a concept does not automatically support a Cronbach's Alpha threshold or an EFA procedure.
The third error is mixing terminology. If you use “perceived usefulness” and perceived usefulness for the same variable, choose one main name and give the English term the first time it appears. Variable names in Chapter 2, the questionnaire, and the SPSS file need to remain consistent.
The fourth error is putting results in Chapter 2. The theoretical framework can state expected criteria and the sources for those criteria, but KMO, Alpha, Beta, or p-value values from your sample belong in the results chapter.
The fifth error is using every familiar theory in one topic. TAM, TPB, UTAUT, and SERVQUAL should not appear together simply because they are well known. Each theory needs a specific role in the model. Otherwise, Chapter 2 becomes long but loses focus.
The final error is failing to update the variable coding table after changing the model. When you remove an indicator or rename a variable, check the questionnaire, .sav or .csv file, SmartPLS model, and methodology description. The committee often notices this inconsistency as soon as it moves from a hypothesis to the corresponding results table.
Frequently asked questions
What is a theoretical framework in scientific research
A theoretical framework explains the theories, concepts, and relationships used to build a research model. In quantitative research, it leads to the variables, measurement scales, and hypotheses that must be tested with data.
Are a theoretical framework and a theoretical foundation the same
In many theses, the two terms are used almost interchangeably. However, “theoretical framework” often has a broader scope, including foundational theories, concepts, previous research, and the reasoning used to form the model. “Theoretical foundation” usually places more emphasis on the theories being adopted.
What sections does a theoretical framework include
A common structure includes the main concepts, foundational theory, previous research, research model, hypotheses, and a chapter summary. Depending on the topic, you may add the research context or definitions of control variables, but anything added should serve the research question.
How many pages should a theoretical framework be
There is no page count that applies to every university and topic. Use your department's requirements or your supervisor's guidance as the reference, then check whether every variable in the model has a definition, source, and supporting argument. Relevant content matters more than extending the page count.
What should I do if a hypothesis lacks supporting sources
First, identify which theory supports the hypothesis and find the original source for the relationship rather than searching only for a ready-made quotation. If no suitable basis exists, consider narrowing the model or stating clearly that the relationship is exploratory, instead of presenting it as an established conclusion.
Do I need to run SPSS when writing the theoretical framework
You do not need to run SPSS to write the foundation, but you need to design the theoretical framework so that its concepts can be measured with data. After data collection, SPSS or SmartPLS is used to test reliability, scale validity, and the relationships in the model.
Open your outline file now and create a table with the variable, definition, source, indicator, and hypothesis. Check each row against Chapter 2. If you need support building the model, hypotheses, measurement scales, and questionnaire, use module M3 at DoThesis.