Introduction
Many nursing and healthcare students have questionnaire data in SPSS but are unsure how the items should be grouped. The dataset may contain 15, 20, or 30 Likert-scale items, yet the student may not know whether those items represent one construct, several subscales, or a weak structure that needs revision.
Learning how to do exploratory factor analysis in SPSS helps solve that problem. Exploratory factor analysis, or EFA, examines the hidden structure behind related questionnaire items before you create subscales, remove weak items, or interpret scale dimensions.
The difficult part is not only running the SPSS menu. You must decide whether EFA is appropriate, choose an extraction method, select rotation, review KMO and Bartlett’s test, interpret the pattern matrix, handle cross-loading items, and report the results in APA 7th edition format. This article is a deeper supporting guide to our broader SPSS Factor Analysis article.
Need help running exploratory factor analysis in SPSS? Our SPSS Data Analysis Help service can help you prepare your dataset, run EFA, interpret factor loadings, and write APA-style results for your nursing research.
What Is Exploratory Factor Analysis in SPSS?
Exploratory factor analysis in SPSS is used when a researcher wants to explore possible latent factors behind several related observed items. In simple terms, EFA helps you see whether questionnaire items naturally cluster into smaller dimensions.
For example, a patient education confidence questionnaire may include items about teaching patients, communicating with families, documenting education, and checking understanding. EFA can help show whether these items form one broad confidence scale or several subscales. It is useful when the questionnaire structure is uncertain, the student wants to explore possible subscales, the items may measure several related dimensions, or a nursing instrument is being adapted to a new sample.
EFA does not prove validity by itself. It supports early evidence about item grouping, but interpretation still depends on theory, item wording, sample quality, and previous research.
When Should You Use EFA Instead of PCA?
Students often confuse EFA and PCA because SPSS places both under the Factor menu. However, they answer different questions.
Use EFA when your goal is to explore underlying constructs. This is common when you want to understand whether Likert-scale items represent latent dimensions such as patient education confidence, clinical competence, communication confidence, or medication adherence beliefs.
PCA is mainly a data-reduction method. It summarizes observed variables into components, but it does not model latent constructs in the same way. Fabrigar, Wegener, MacCallum, and Strahan argue that researchers should choose factor-analysis methods based on the purpose of the analysis, not software convenience (Fabrigar et al., 1999).
For dissertation work, choose EFA when your research question concerns scale structure or hidden constructs. Choose PCA only when your purpose is component-based reduction and your methodology supports that decision.
Before Doing Exploratory Factor Analysis in SPSS
Before running EFA in SPSS, check whether the items belong together conceptually. Do not analyze unrelated survey questions only because they appear in the same questionnaire.
Confirm that items use similar response scales. A group of 5-point Likert items is easier to justify than a mixture of Likert items, yes/no items, demographics, and open-ended responses.
Check reverse-coded items before EFA. If negatively worded items are not recoded, the correlation matrix may be distorted. This can affect KMO, communalities, factor loadings, and the final factor solution.
Review missing values, out-of-range values, variable labels, value labels, and coding errors. These steps connect closely with Data Cleaning Steps in Nursing Research because SPSS output is only as reliable as the dataset behind it.
Also inspect frequencies, means, standard deviations, and item distributions. This is part of Descriptive Data Analysis in Nursing Research and helps identify unusual response patterns before EFA.
Suitability Checks Before Running EFA
Sample Size
Larger samples usually produce more stable EFA solutions, but there is no single rule that works for every study. Sample adequacy depends on the number of items, communalities, factor strength, expected number of factors, and research context. Fabrigar et al. explain that sample size needs depend partly on communalities and whether factors are represented by enough items (Fabrigar et al., 1999).
Correlation Matrix
EFA requires meaningful relationships among items. If items are weakly related, there may be no useful structure to extract.
KMO Measure
KMO helps assess sampling adequacy. As a practical guide, values below .50 are usually weak, .60s are mediocre, .70s are acceptable, .80s are good, and .90 or above is excellent. Treat this as guidance, not a universal rule.
Bartlett’s Test of Sphericity
Bartlett’s test checks whether the correlation matrix differs from an identity matrix. A significant result usually supports proceeding because the items show enough relationships for factor analysis. IBM includes KMO and Bartlett’s test among the SPSS factor-analysis descriptives used to evaluate factorability (IBM Corp., n.d.).
Step-by-Step: How to Do Exploratory Factor Analysis in SPSS
To run EFA in SPSS:
- Open the dataset in SPSS.
- Click Analyze.
- Select Dimension Reduction.
- Click Factor.
- Move the related questionnaire items into the Variables box.
- Click Descriptives.
- Select Initial solution, Coefficients, and KMO and Bartlett’s test of sphericity.
- Click Continue.
- Click Extraction.
- Choose an EFA extraction method, such as Principal Axis Factoring, if appropriate.
- Select Scree plot.
- Review or adjust the number-of-factors option based on your study plan.
- Click Continue.
- Click Rotation.
- Choose Direct Oblimin or Promax if factors may correlate, or Varimax if factors are expected to be independent.
- Click Continue.
- Click Options.
- Choose Sorted by size.
- Suppress small coefficients if this helps interpretation.
- Click Continue.
- Click OK.
IBM lists the SPSS factor procedure under Analyze > Dimension Reduction > Factor, with options for extraction, rotation, scores, and output displays (IBM Corp., n.d.).
Choosing the Extraction Method in SPSS EFA
Extraction determines how SPSS identifies the factors from the item correlation matrix. Your choice should match the research purpose, methodology chapter, questionnaire type, and supervisor guidance.
Principal Axis Factoring is commonly used when the purpose is latent factor exploration. It focuses on shared variance among items, which fits many questionnaire-based nursing studies.
Maximum Likelihood is another EFA option. It may be useful when assumptions are reasonable and the researcher wants additional model-based information. However, it should not be chosen automatically.
PCA appears in the same SPSS Factor procedure, but it is not the same as EFA. PCA may be appropriate for data reduction, but it is not the best choice when the goal is latent construct exploration.
Costello and Osborne emphasize that extraction choices affect EFA results and should be made carefully rather than left to default software settings (Costello & Osborne, 2005).
Choosing the Rotation Method in SPSS EFA
Rotation makes the factor solution easier to interpret. Without rotation, items may not show a clear pattern.
Orthogonal rotation assumes factors are unrelated. Varimax is the most common orthogonal option. It may be suitable when the study assumes that factors are independent.
Oblique rotation allows factors to correlate. Direct Oblimin and Promax are common oblique options in SPSS. In nursing and healthcare research, constructs such as confidence, communication, self-efficacy, satisfaction, and competence are often related. In those cases, oblique rotation may be more defensible.
Watkins notes that EFA requires thoughtful decisions at each stage, including extraction and rotation, because poor choices can weaken interpretation (Watkins, 2018).
Do not choose varimax only because it is familiar. Choose rotation based on theory, expected relationships among factors, previous literature, and supervisor guidance.
SPSS EFA Output: What to Read First
KMO and Bartlett’s Test
Read this table first. KMO helps evaluate sampling adequacy, while Bartlett’s test shows whether the correlation matrix has enough relationships to support EFA.
Communalities
Communalities show how much variance in each item is explained by the extracted factors. Low extraction communalities may suggest weakly represented items.
Total Variance Explained
This table shows eigenvalues, extracted factors, and variance explained. It helps you understand how much item variance is represented by the retained solution.
Scree Plot
The scree plot helps identify the point where eigenvalues begin to flatten. It should support factor-retention decisions, not replace them.
Pattern Matrix
When using oblique rotation, the pattern matrix is usually central. It shows the unique relationship between each item and each factor.
Structure Matrix
The structure matrix shows correlations between items and factors. Beginners should not confuse it with the pattern matrix.
Factor Correlation Matrix
This table shows whether retained factors are correlated. It helps justify oblique rotation when factors are meaningfully related.
How to Decide the Number of Factors to Retain
Factor retention is one of the most important EFA decisions. Do not rely only on the eigenvalue-greater-than-one rule. That rule is common, but it can retain too many or too few factors.
Use several types of evidence: eigenvalues, scree plot, theoretical meaning, previous validation studies, factor interpretability, number of strong item loadings per factor, cross-loading problems, and supervisor or committee expectations.
Parallel analysis can also help when available, although students may need software outside standard SPSS menus to conduct it.
A factor should make statistical and conceptual sense. For example, if SPSS suggests four factors but one factor has only one strong item, that factor may be difficult to defend. If three items load together and share a clear nursing meaning, the factor may be more interpretable.
Watkins describes EFA as a decision-based process, not a single automatic test (Watkins, 2018).
How to Interpret Factor Loadings in SPSS EFA
Factor loadings show the relationship between an item and a factor. Larger absolute values indicate stronger item-factor relationships.
Items should ideally load clearly on one factor. If an item loads strongly on more than one factor, it is called a cross-loading item. Cross-loading items can make the factor structure harder to interpret.
Weak-loading items may need review, but do not remove them automatically. Check item wording, coding direction, reverse coding, theoretical importance, and previous literature.
Negative loadings require careful interpretation. Sometimes they make sense because of item wording. Other times, they may suggest a coding problem.
Many studies use practical thresholds such as .30, .40, or .50, but no single cutoff fits every dissertation. Loading decisions should reflect sample size, discipline, research purpose, theory, and supervisor guidance.
What to Do With Weak or Cross-Loading Items
When an item loads weakly or cross-loads, start with the item itself. Ask whether the wording is unclear, double-barreled, too broad, or inconsistent with the construct.
Next, check reverse coding. A wrongly coded item can create weak or unexpected loadings.
Then compare the item with previous literature. If the questionnaire has been used before, check whether the same item behaved poorly in earlier studies. If it is a new questionnaire, decide whether the item truly belongs to the factor.
Do not delete items only to make the matrix cleaner. Removing items can improve appearance while weakening theoretical coverage.
If you remove an item, report the decision transparently. State the reason, the number of items removed, and how the final factor solution was retained.
Naming Factors After EFA
After reviewing the pattern matrix, name each retained factor based on the items that load on it. Do not name a factor based on one item unless there is a strong methodological reason.
Look for the shared meaning among the items. If several items describe explaining care instructions, confirming understanding, and answering patient questions, the factor may be named patient teaching confidence.
Use terms that match nursing practice, the questionnaire theory, and previous literature. Avoid vague names such as “Factor 1” or “General Issues” in the final discussion unless the factor cannot be interpreted clearly.
Good factor names should be clear, defensible, and connected to the research questions.
Common Mistakes When Doing Exploratory Factor Analysis in SPSS
Common mistakes include:
- Running EFA on unrelated items
- Treating PCA as EFA
- Ignoring KMO and Bartlett’s test
- Forgetting reverse-coded items
- Relying only on eigenvalues greater than 1
- Ignoring the scree plot
- Using varimax automatically when factors may be related
- Confusing pattern matrix and structure matrix
- Keeping cross-loading items without explanation
- Removing items without theoretical justification
- Claiming EFA proves validity
- Reporting output without explaining decisions
EFA can support evidence about questionnaire structure, but it does not prove validity by itself.
If your exploratory factor analysis output has unclear factors, weak loadings, cross-loading items, or confusing pattern and structure matrices, our Dissertation Data Analysis Help service can help you interpret the results and write your findings clearly.
How to Report Exploratory Factor Analysis in APA 7th Edition
Exploratory factor analysis APA reporting should explain the purpose, sample size, number of items, extraction method, rotation method, KMO, Bartlett’s test, retention criteria, number of factors, variance explained, key loadings, item removals, final factor names, and any follow-up reliability analysis.
APA Style guidance supports clear statistical reporting with readable numbers and enough context for interpretation (American Psychological Association, 2024).
Example:
“An exploratory factor analysis was conducted to examine the structure of the 15-item patient education confidence scale. Principal axis factoring with oblique rotation was used because the underlying factors were expected to correlate. The Kaiser-Meyer-Olkin measure indicated adequate sampling suitability, KMO = .82, and Bartlett’s test of sphericity was statistically significant, χ²(105) = 684.21, p < .001. A three-factor solution was retained based on the scree plot, factor loadings, and interpretability. The retained factors were patient teaching confidence, communication confidence, and documentation confidence.”
Exploratory Factor Analysis Example in Nursing Research
A nursing student has a 20-item patient education confidence questionnaire measured on a 5-point Likert scale. The student wants to know whether the items form meaningful subscales.
First, the student checks missing values, out-of-range responses, item labels, and reverse-coded items. Descriptive screening shows that all items are coded correctly.
Next, the student runs EFA in SPSS. KMO is acceptable, and Bartlett’s test is significant, so the item relationships support proceeding.
The student selects Principal Axis Factoring because the goal is latent factor exploration. Oblique rotation is chosen because patient teaching confidence, communication confidence, and documentation confidence are likely related.
The pattern matrix suggests three factors. Most items load clearly, but two items cross-load. The student reviews their wording, checks previous literature, and discusses whether they should be retained or removed.
The final factors are named patient teaching confidence, communication confidence, and documentation confidence. After EFA, the student runs Cronbach’s alpha for each factor before creating subscale scores.
How EFA Connects to Reliability Analysis and Later Testing
EFA helps identify possible factors or subscales. After that, students often run Cronbach’s alpha for each retained factor to assess whether items within each factor work together consistently.
If reliability is acceptable, the student may create subscale scores. These scores can then be used in descriptive statistics, correlations, regression, t-tests, ANOVA, or other quantitative analyses when appropriate.
This sequence matters. EFA helps identify structure, reliability analysis checks item consistency within that structure, and later tests examine relationships or group differences.
When to Get Help With EFA in SPSS
Get help when you are unsure whether EFA is appropriate for your dataset, when your extraction method is unclear, or when your rotation choice is difficult to defend.
Support may also be useful when KMO is low, Bartlett’s test is not significant, the scree plot is unclear, or the pattern matrix has weak and cross-loading items. These issues require careful judgment, not only SPSS screenshots.
Students also need help when supervisor feedback asks them to justify factor retention, explain item deletion, connect EFA with Cronbach’s alpha, or rewrite the results in APA format.
For dataset setup, EFA decisions, SPSS output interpretation, and APA-style reporting, SPSS Data Analysis Help can support your questionnaire analysis from preparation to final write-up.
Conclusion
Knowing how to do exploratory factor analysis in SPSS helps nursing and healthcare students examine the hidden structure of questionnaire items before creating subscales or reporting scale results. Strong EFA requires more than clicking Analyze, Dimension Reduction, and Factor. It requires careful decisions about data preparation, suitability checks, extraction, rotation, factor retention, item review, factor naming, and APA 7th edition reporting.
The best EFA write-ups explain what was tested, why the method was chosen, how many factors were retained, what the factor loadings showed, and how item decisions were made.
Need expert help with exploratory factor analysis in SPSS, pattern matrices, factor loadings, factor retention, or APA 7th edition reporting? Upload your dataset, questionnaire, research questions, and rubric through our SPSS Data Analysis Help page for focused nursing research support.
FAQs
1. What is exploratory factor analysis in SPSS?
Exploratory factor analysis in SPSS is used to explore how related questionnaire items group into underlying factors or subscales.
2. How do I run exploratory factor analysis in SPSS?
Click Analyze, select Dimension Reduction, choose Factor, move related items into the Variables box, request KMO and Bartlett’s test, choose extraction and rotation methods, then review the output.
3. What extraction method should I use for EFA in SPSS?
Principal Axis Factoring is commonly used for latent factor exploration. Maximum Likelihood may be appropriate when assumptions fit the study. The method should match your research purpose and methodology.
4. Should I use varimax or oblimin rotation in SPSS EFA?
Use varimax when factors are expected to be unrelated. Use oblimin or promax when factors may be correlated. Many nursing constructs are related, so oblique rotation is often defensible.
5. How do I report exploratory factor analysis in APA 7th edition format?
Report the purpose, sample size, items, extraction method, rotation method, KMO, Bartlett’s test, retained factors, variance explained, factor loadings, item decisions, and final factor names.
References
American Psychological Association. (2024). Numbers and statistics guide.
Costello, A. B., & Osborne, J. W. (2005). Best practices in exploratory factor analysis: Four recommendations for getting the most from your analysis. Practical Assessment, Research, and Evaluation, 10(7), Article 7.
Fabrigar, L. R., Wegener, D. T., MacCallum, R. C., & Strahan, E. J. (1999). Evaluating the use of exploratory factor analysis in psychological research. Psychological Methods, 4(3), 272–299.
IBM Corp. (n.d.). Factor analysis. IBM Documentation.
IBM Corp. (n.d.). Factor analysis descriptives. IBM Documentation.
Watkins, M. W. (2018). Exploratory factor analysis: A guide to best practice. Journal of Black Psychology, 44(3), 219–246.
