What Is the SWOT Model? How to Build and Test It in a Thesis

What is the SWOT model

The SWOT model is an analytical framework that helps you examine an organization, business, product, or project through four groups of factors: Strengths, Weaknesses, Opportunities, and Threats. The first two groups usually concern factors inside the organization, while the latter two usually concern the external environment.

If you are writing a thesis, SWOT is useful when you define the context and identify groups of factors that could enter your research model. A SWOT matrix by itself is not yet a complete quantitative model. You still need to turn the observations into research variables, develop scales, collect data with a questionnaire, and test the model with SPSS or SmartPLS.

The name SWOT comes from the four groups of factors listed above. Classic SWOT materials are usually presented within strategic management, but the sources allowed for this article do not include an original source identifying the author and year in which SWOT began. You should therefore avoid adding an unchecked year or author to your thesis simply to make the model's history look complete.

You can treat SWOT as an initial map. It tells you where to look, but it does not use data to show which factor has the strongest effect, whether the effect is positive or negative, or whether it is statistically significant. To reach those conclusions, you must design and test a separate research model.

Components of the model

The four SWOT components should not be written as a list of general statements. When you apply them to a topic, define the unit of analysis, the scope of the context, and how each factor will be measured.

The boundary between internal and external factors needs careful attention. For example, “the staff have strong skills” usually belongs under Strengths because the business can develop and manage this resource. “Customer demand is increasing” usually belongs under Opportunities because it is a market condition. If you classify these factors incorrectly, the strategic recommendations will lack a clear logic.

In a quantitative thesis, you can use the four groups in three ways. First, you can use SWOT to describe the context and then choose a specific issue for a separate research model. Second, you can convert each group into independent variables and test their effects on a dependent variable. Third, you can use survey results to rank the factors and then propose priority solutions.

The second approach requires the most care. A respondent may evaluate a business strength, but that does not automatically mean the strength creates purchase intention or business performance. A causal relationship must be explained by theory and research literature that fit the context.

If you need to compare SWOT with other strategic tools, you can also read Porter's value chain model, Porter's five competitive forces, or the BCG model. These tools answer different questions, so you should not combine their categories simply because they appear in the same theoretical background chapter.

The original model and common extensions

SWOT usually begins with a four-cell matrix. You list internal factors along one dimension and external factors along the other, then combine them to consider possible courses of action. The four combinations are commonly interpreted as using strengths to take advantage of opportunities, using strengths to reduce threats, overcoming weaknesses to take advantage of opportunities, and reducing weaknesses to limit threats.

In a thesis, describe an extension as the way you operate the analytical framework. Do not automatically call every four-cell table a “new model.” Common extensions include:

• Qualitative SWOT, which focuses on identifying and describing factors.
• Scored SWOT, in which respondents evaluate the importance or presence of each factor.
• TOWS, which places more emphasis on matching factors to form strategic groups.
• SWOT combined with other tools such as PESTEL, industry analysis, or a priority matrix.

If you use SWOT to build a research model, a reasonable extension may be to add a mediator, moderator, or outcome variable. For example, technological capability may affect service quality, while service quality affects satisfaction. In that case, SWOT only helps you identify technological capability as an issue worth studying. The relationships between variables must be supported by the literature review and separate hypotheses.

You can also place SWOT among context-analysis tools alongside the Balanced Scorecard or stakeholder analysis. The Balanced Scorecard focuses on turning objectives into performance measurement perspectives. Stakeholder analysis focuses on identifying the people or organizations that have an interest in and influence over an issue. SWOT is broader when grouping observations into four categories, but it is usually less specific when it comes to measurement.

A common mistake is to take a SWOT table from a textbook, replace the business name, and treat it as a research model. That creates a descriptive section only. If your topic aims to test effects, state clearly which variable is the cause, which is the outcome, who provided the data, and where the scale came from.

Common scales for each concept

SWOT does not have one scale that fits every topic. The “strengths” of a bank will differ from the “strengths” of a tourist destination or an e-commerce platform. Start with a specific concept, then find the original scale in studies that fit your research setting.

Each concept should have multiple items, written as statements that respondents can rate on a Likert scale. Instead of asking, “Does the business have strengths?”, you might measure “The business responds quickly to customer requests,” “The business can maintain consistent service quality,” and “Employees have enough skill to handle customer requests.” These statements are illustrative only. You need to adjust and verify them against the original literature, pilot-test the questionnaire, and consider the actual context.

For a Likert scale, you can use a range from 1 to 5 or 1 to 7, but keep the coding consistent. If an item has the opposite meaning, reverse-code it before running reliability tests. In SPSS, you can create a new variable through Transform > Compute Variable instead of editing the original data column directly.

The usual testing sequence includes Cronbach's Alpha, EFA when required by the research design, and then regression or a structural model. Cronbach's Alpha of 0.7 or above is commonly considered acceptable (Nunnally, 1978). For a new or exploratory scale, Alpha from 0.6 may be considered (Hair et al., 2010). A Corrected Item-Total Correlation of 0.3 or above is a commonly used threshold (Nunnally and Bernstein, 1994).

If you use SmartPLS, you will assess outer loading, composite reliability, AVE, and discriminant validity. An outer loading of 0.7 or above is a reference threshold in PLS-SEM (Chin, 1998). CR of 0.7 and AVE of 0.5 or above are commonly used to assess reliability and convergent validity (Fornell and Larcker, 1981). HTMT should be below 0.85, or below 0.90 when the two concepts are close in content (Henseler et al., 2015).

Applying the model to your topic

Suppose you are studying the factors that affect students' intention to use an e-commerce platform. You could use SWOT to explore the context and then select factors that may explain usage intention. An illustrative model might include technological capability, service quality, perceived convenience opportunities, perceived risk, and usage intention.

Draw this model with clear arrows. Technological capability and service quality may be treated as internal strengths of the platform. Perceived convenience opportunities reflect an external condition or a benefit perceived by users. Perceived risk may represent the threat group. Usage intention is the outcome variable.

Four illustrative hypotheses could be written as follows:

• H1: Technological capability has a positive effect on intention to use the e-commerce platform.
• H2: Service quality has a positive effect on intention to use the e-commerce platform.
• H3: Perceived convenience opportunities have a positive effect on intention to use the e-commerce platform.
• H4: Perceived risk has a negative effect on intention to use the e-commerce platform.

This is an illustrative model, not a default model for every SWOT topic. Replace these variables with concepts that fit your research question, respondents, and the literature you have read. If you use TAM, TPB, or UTAUT, explain why that theoretical framework fits. For example, UTAUT may fit a technology-acceptance topic, while SERVQUAL may be more appropriate when you study service quality.

Sample size must also be considered in relation to the number of items and the analysis method. A commonly cited rule is 5 to 10 observations for each item (Hair et al., 2010). For regression, a reference formula is a sample size of at least 50 plus 8 times the number of independent variables (Tabachnick and Fidell, 2013). This is a planning basis, not a promise that your data will produce significant results.

How to write this in your thesis: “Based on the research objectives and literature review, the study proposes a model comprising [names of independent variables] affecting [name of dependent variable] in the context of [population and location]. The SWOT model is used to organize internal and external factors, while the relationships between variables are tested using quantitative survey data.”

How to test the model with SPSS or SmartPLS

If your model contains observed variables and you need to test direct effects, SPSS can handle Cronbach's Alpha, EFA, correlation, and linear regression. Go to Analyze > Scale > Reliability Analysis to run Cronbach's Alpha, select Statistics, and tick Scale if item deleted. You can then go to Analyze > Dimension Reduction > Factor to run EFA and inspect KMO and Bartlett's Test, Total Variance Explained, and Rotated Component Matrix.

For EFA, KMO of 0.5 or above and a significant Bartlett test at below 0.05 are commonly used criteria (Kaiser, 1974). An Eigenvalue greater than 1 is commonly used as the factor-extraction rule (Kaiser, 1960). Total variance explained of 50% or above and factor loading of 0.5 or above are widely used reference thresholds (Hair et al., 2010). You still need to inspect the item content and the level of cross-loading. Do not remove an item based on one number viewed in isolation.

Illustrative output

If your model has latent variables, several relationships, or a predictive objective, SmartPLS may be a better fit. In SmartPLS 4, import the .csv data file, create constructs, drag the indicators into the constructs, and draw arrows according to the hypotheses. Then run PLS-SEM Algorithm to inspect Outer Loadings, Construct Reliability and Validity, and Discriminant Validity. Run Bootstrapping to inspect Path Coefficients, T Statistics, P Values, and confidence intervals.

The PLS-SEM process should move from the measurement model to the structural model. This is the two-step approach described in the methods literature (Anderson and Gerbing, 1988). When reporting PLS-SEM, check VIF, path coefficients, R², f², and Q². VIF below 5 is a reference threshold (Hair et al., 2019). R² values of 0.75, 0.50, and 0.25 may be interpreted as substantial, moderate, and weak in some PLS-SEM contexts (Hair et al., 2011). Q² above 0 indicates that predictive relevance may be considered (Stone, 1974).

When running Bootstrapping in SmartPLS, you can set 5,000 subsamples according to commonly used guidance (Hair et al., 2022). Do not mix CFI, TLI, and RMSEA criteria into a SmartPLS paper. These indices belong to CB-SEM, with reference thresholds for CFI, TLI, RMSEA, and SRMR (Hu and Bentler, 1999).

Common mistakes when using this model

The first mistake is treating SWOT as causal evidence. A table stating that “the brand is strong” does not prove that the brand creates loyalty. You need measured variables, data, and an appropriate test.

The second mistake is giving a variable an overly broad name. “Business strengths” may include brand, employees, finance, technology, and processes. If you combine all of these into one construct without a theoretical basis, the scale will be difficult to keep consistent.

The third mistake is using sources of unclear origin for a scale. A questionnaire found online is not automatically a validated scale. Trace it back to the original study, check the context, number of items, and wording before translating or adapting it.

The fourth mistake is removing items continuously until the output looks good. Item removal should be based on loading, correlation, content, and the original design. Save each data-file version, record the reason for removing an item, and report the process transparently in your thesis.

The fifth mistake is combining SPSS and SmartPLS results without explaining the sequence. If you use SPSS for EFA and then move to SmartPLS, state the role of each software package, the decision criteria, and the order of analysis. If the model only ranks the priority of SWOT factors, you may not need to build a complex causal model.

Finally, check whether your topic genuinely needs SWOT or whether it needs a more specific theoretical model. You can also review topics in Research models to compare how variables, hypotheses, and scales are developed.

Frequently asked questions

Is the SWOT model a quantitative research model?

SWOT is a strategic analysis framework, not a complete quantitative model. You can use SWOT to identify factors, then convert suitable factors into independent variables, dependent variables, or mediators and test them with survey data.

If your thesis only creates a four-cell matrix and proposes recommendations, it is a SWOT analysis. If your thesis tests effects between variables, you need to present the theoretical basis, scales, hypotheses, and analysis method as well.

What does the SWOT model include?

SWOT includes Strengths, Weaknesses, Opportunities, and Threats. Strengths and Weaknesses usually reflect internal factors, while Opportunities and Threats usually reflect the external environment.

When writing your thesis, describe the classification criteria and unit of analysis. You should place a factor in one group only when you can explain why it belongs to the internal environment or the external environment.

Can you run a SWOT model in SPSS?

There is no button called “SWOT” in SPSS. You use SPSS to analyze data from variables developed from the SWOT context, such as by running Cronbach's Alpha, EFA, correlation, and regression.

If you only need to rank the importance of SWOT factors, you can use descriptive statistics. If you want to test effects, define the dependent variable and select an appropriate analysis model.

Should you use SPSS or SmartPLS for a SWOT model?

SPSS fits a relatively simple model when you want to run EFA, regression, or data tests through a familiar procedure. SmartPLS fits a topic with several latent variables, many indicators, and an objective that includes assessing both the measurement model and the structural model.

The name SWOT does not determine the software. The research question, variable type, sample size, and testing objective are the basis for the choice. You should also avoid using SmartPLS simply to avoid checking scale quality.

Does a SWOT model need hypotheses?

If you only present a SWOT matrix and propose strategies, causal hypotheses are not necessarily required. If you use a survey to test the effects of factors on an outcome, you need to develop a hypothesis for each relationship.

Each hypothesis should state the influencing variable, the affected variable, the expected direction, and the context. The hypotheses in this article are illustrative examples of wording, not conclusions about your topic.

Open your data file and first determine whether you have a descriptive SWOT matrix or a model of variables to be tested. Then choose the corresponding analysis procedure in M4 at DoThesis.