How to Discuss SPSS Results in a Nursing Dissertation: Step-by-Step Guide

How to discuss SPSS results in a nursing dissertation is not the same task as copying statistics from the SPSS Output Viewer into a discussion chapter. The discussion has to explain what the numbers mean for the research question, how convincing the pattern is, how it fits with nursing evidence, and what can reasonably be said about practice.

This guide concentrates on that interpretive step. If you are still deciding which SPSS tables belong in Chapter 4, start with our nursing SPSS results chapter guide. If you need a full chapter outline, use the separate nursing dissertation discussion chapter structure. Keeping those tasks separate prevents the results and discussion chapters from becoming repetitive.

How to Discuss SPSS Results in a Nursing Dissertation Without Repeating the Results

The first practical rule is simple: the discussion should not reproduce the results chapter. A results chapter tells the reader what the analysis found. The discussion asks what that finding means and how much confidence the reader should place in it.

Before drafting, reduce each important SPSS result to one line. For example:

Nurses who completed the discharge-education workshop had higher patient-education confidence scores than nurses who had not completed the workshop.

That one sentence becomes the starting point for discussion. You do not need to reopen with a full table, every decimal place or every assumption test. Instead, ask five questions:

1. What is the direction of the finding?
2. How large is the difference or association?
3. How precise is the estimate?
4. Does previous nursing evidence show a similar pattern?
5. What does the finding mean for nursing practice, education, policy or future research?

Lang and Altman (2015) argue for statistical reporting that gives enough information to understand both the estimate and its uncertainty. This matters in a dissertation discussion because a p-value alone does not show whether the result is large, precise or clinically useful.

Build an SPSS Discussion Worksheet Before Writing Paragraphs

A useful way to prevent repetition is to create a short worksheet from your final SPSS output. Do this after the results chapter is complete. The worksheet is not another results table. It is a planning tool for deciding what deserves interpretation.

SPSS evidence	What you should extract	Discussion question
Descriptive statistics	Mean, median, spread, proportion or pattern	What does the sample pattern suggest?
t-test or ANOVA	Group difference, confidence interval, effect size	Is the difference meaningful as well as statistically detectable?
Chi-square	Association, cell pattern, Cramer’s V or Phi	Which categories drive the relationship and how strong is it?
Correlation	Direction, strength, confidence interval	What relationship is present, and what cannot be inferred?
Regression	B, beta, confidence interval, model fit	What does the predictor contribute after other variables are considered?
Non-significant test	Estimate, confidence interval, sample size	Does the evidence suggest little effect, or is it too imprecise to decide?

This worksheet gives each statistic a purpose. It also stops a common problem in nursing dissertations: writing one paragraph for every SPSS table simply because the software produced it.

Start With the Estimate, Not With “p < .05”

Nursing students often begin discussion paragraphs with statistical significance. That can flatten the analysis. A p-value helps assess compatibility between the data and a null model, but it does not tell you the clinical size of an effect. Wasserstein, Schirm and Lazar (2019) caution against treating a threshold such as p < .05 as the sole basis for scientific conclusions.

A stronger discussion begins with the substantive pattern. Consider this illustrative result:

Illustrative SPSS result: Mean pressure-injury prevention knowledge was 78.6 in the trained group and 72.1 in the untrained group, mean difference = 6.5, 95% CI [2.1, 10.9], p = .004, Cohen’s d = 0.55.

A weak discussion would say: The finding was statistically significant because p was less than .05.

A better discussion would explain that nurses who received training scored higher, that the observed difference was moderate in size, and that the confidence interval suggests a positive effect is plausible across a range of values. The next question is whether a difference of this size has practical meaning for the knowledge measure used in the study.

Davis, Johnson, Lynch, Gray, Pryor, Azuero, Soistmann, Phillips and Rice (2021) emphasise that effect sizes, confidence intervals and clinical relevance add information that statistical significance alone cannot provide. Barría (2023) makes a similar point in nursing practice, distinguishing statistical evidence from the judgement about whether a result matters clinically.

How to Discuss SPSS Results in a Nursing Dissertation by Test Type

Independent-samples t-test

For a t-test, focus on the direction of the difference, its magnitude, the confidence interval and its possible nursing meaning. You usually do not need to spend a discussion paragraph explaining the t statistic itself.

Illustrative example: Nurses who received structured sepsis simulation reported higher recognition-confidence scores than nurses receiving standard teaching, mean difference = 5.8, 95% CI [1.9, 9.7], p = .004.

The discussion could ask whether previous simulation studies report similar improvements, whether the confidence measure reflects actual recognition performance, and whether the study setting could have amplified the effect. If the research was observational or the groups were not randomly allocated, avoid claiming that simulation caused the improvement.

One-way ANOVA

An ANOVA tells you whether there is evidence that at least one group differs, but the discussion should identify where the meaningful differences sit. Use the appropriate post hoc comparisons and an effect-size measure rather than discussing only the omnibus p-value.

Suppose confidence in end-of-life communication differed across three experience groups, F(2, 147) = 5.96, p = .003, with a moderate overall effect. If post hoc tests showed that newly qualified nurses differed from nurses with more than five years of experience, while the two more experienced groups were similar, that pattern is the finding to discuss. You might explore clinical exposure, mentorship or repeated communication experience as possible explanations, while recognising that a cross-sectional design cannot prove those mechanisms.

Chi-square test

With chi-square, do not stop at “there was a significant association.” Look at the contingency table to identify the pattern behind the statistic. Then consider the strength of the association and the plausibility of competing explanations.

Illustrative example: Completion of a structured handover programme was associated with whether nurses reported omitting critical handover information, χ²(1) = 7.82, p = .005, Cramer’s V = .24.

The discussion should identify which group had the higher or lower proportion, not merely repeat χ² and p. It should then consider whether training exposure, ward culture, staffing pressure or documentation systems could account for the observed pattern. If expected cell counts were problematic, that belongs in the interpretation of confidence rather than being hidden in the methods section.

Correlation

Correlation results need careful wording because “related to” is not the same as “caused by.” If workload and safety-climate scores were negatively correlated, the discussion can say that heavier workload tended to occur alongside poorer safety-climate ratings. It cannot conclude that workload caused the poorer ratings unless the study design supports causal inference.

Illustrative example: Workload score was moderately negatively correlated with safety-climate score, r = -.42, p < .001.

Useful discussion questions include whether the relationship is consistent with previous nursing workforce research, whether both variables were measured by self-report, whether restricted ranges weakened the coefficient, and whether unmeasured staffing or organisational factors could explain part of the association.

Multiple regression

Regression deserves more than a list of significant predictors. Start with the model’s purpose, then discuss the variables that made a meaningful contribution after adjustment.

Illustrative example: After adjusting for age, years of experience and clinical area, workload remained associated with lower safety-climate scores, B = -2.10, 95% CI [-3.25, -0.95], β = -.31, p < .001. The model explained 27% of the variance in safety-climate scores.

That result suggests an adjusted relationship, not proof that workload independently causes changes in safety climate. The discussion should also recognise that 73% of the observed variation remains outside the model. That opens a useful line of reasoning around leadership, staffing mix, organisational culture, shift patterns or measurement limitations.

How to read this output for discussion: Start with the Model Summary to understand how much variation the model explains. Use the ANOVA section to judge whether the regression model as a whole provides evidence of a relationship, then use the Coefficients table for the direction, adjusted association and uncertainty around individual predictors. In the discussion chapter, translate those pieces into what the model suggests about the nursing research question rather than reproducing the output table line by line.

How to Discuss a Non-Significant SPSS Result

A non-significant finding should not be described as “no difference” unless the evidence genuinely supports that conclusion. The result may be small, imprecise or underpowered. The confidence interval helps distinguish these possibilities.

Consider an illustrative comparison where the intervention group scored 2.1 points higher than the comparison group, 95% CI [-1.3, 5.5], p = .22. The interval includes no difference, but it also includes a potentially useful positive effect. The defensible interpretation is that the study did not provide sufficiently precise evidence to establish a difference. It is not proof that the intervention has no effect.

This distinction is particularly important in small nursing dissertations. Davis and colleagues (2021) note that clinically relevant effects can be overlooked when interpretation centres only on significance thresholds.

Move From Statistical Significance to Clinical Meaning

Clinical significance asks a different question from statistical significance. Statistical testing asks how compatible the observed data are with a specified model. Clinical interpretation asks whether the size of the difference would matter to patients, nurses, services or outcomes.

Polit (2017) highlighted the limited attention given to clinical significance in nursing research and argued for greater consideration of meaningful change. Barría (2023) similarly stresses that nurses need to distinguish statistical evidence from practical usefulness.

When discussing clinical meaning, ask:

• Does the outcome measure have a recognised minimally important difference?
• Is the effect large enough to influence care, safety, education or patient experience?
• Would the benefit justify the resources needed for implementation?
• Is the confidence interval narrow enough to support a practical judgement?
• Does the result apply to the patients or nurses who would actually receive the intervention?

A tiny statistically significant difference in a large sample may have little practical value. A clinically important difference in a small sample may fail to reach p < .05 because the estimate is imprecise. Your discussion should show that you understand both situations.

Compare Your SPSS Findings With Nursing Literature

Once you understand your own result, compare it with relevant literature. Avoid searching for papers merely because they use the same statistical test. The important match is conceptual: population, exposure or intervention, outcome, setting and study design.

For each major result, group previous studies into three broad positions:

• Convergent evidence: studies that report a similar direction or pattern.
• Mixed evidence: studies that partly agree but differ in magnitude, subgroup or setting.
• Contrasting evidence: studies that show little association, an opposite direction or a different pattern.

Then examine why. Differences may arise from sample characteristics, measurement tools, clinical setting, follow-up period, intervention intensity, staffing context, analysis choices or study quality. This produces a discussion of evidence rather than a list of authors.

Use Nursing Theory Only When It Explains the Result

A theoretical framework can strengthen interpretation when it genuinely explains the observed pattern. It should not be inserted after the statistics simply because the proposal mentioned a theory.

For example, if a study found an association between self-efficacy and adherence to a nursing intervention, a self-efficacy framework may help explain why confidence relates to behaviour. The discussion should still recognise alternative explanations and the limits of the research design.

Use theory to illuminate a mechanism, boundary or contradiction. Do not force every SPSS result into the same theoretical statement.

Discuss Unexpected SPSS Findings Instead of Hiding Them

An unexpected result can be one of the strongest parts of a discussion if it is handled carefully. Begin by checking whether the result is technically credible. Review coding, missing-data decisions, assumptions, subgroup sizes and influential observations. If the result remains defensible, interpret it rather than trying to make it match the original hypothesis.

Possible explanations might involve contextual differences, measurement limitations, ceiling effects, confounding, sample composition or changes in clinical practice since earlier studies were published. Present these as plausible explanations, not established facts.

Let Limitations Change the Interpretation

A limitation matters because it changes how strongly you can interpret the result. Listing limitations at the end of a chapter without connecting them to individual findings weakens the discussion.

For example, self-reported adherence may overestimate actual behaviour. A single-site sample may limit generalisability. A wide confidence interval may indicate that the study is not precise enough to distinguish a trivial effect from an important one. A cross-sectional design may identify association but not temporal order.

For observational studies, the STROBE statement encourages transparent reporting that allows readers to understand what was done, what was found and what the findings mean (von Elm, Altman, Egger, Pocock, Gøtzsche and Vandenbroucke, 2007). The same discipline is valuable in a dissertation discussion.

Turn SPSS Findings Into Proportionate Nursing Implications

The final interpretive move is to explain what the evidence could reasonably mean for nursing. Keep recommendations proportional to the design and strength of evidence.

If a small cross-sectional dissertation finds that workload is associated with poorer safety-climate scores, it may support further assessment of workload and safety culture in similar settings. It would not justify claiming that a specific staffing policy will reduce adverse events.

If a controlled intervention shows a clear, reasonably precise improvement in a clinically meaningful nursing outcome, the discussion may support piloting or evaluating the intervention more widely. Even then, implementation cost, acceptability, equity and local context need consideration.

SPSS Discussion Language: What to Avoid and What to Write Instead

Avoid	Prefer	Why
“SPSS proved that…”	“The analysis indicated…”	Statistical analysis rarely proves a substantive claim.
“There was no effect because p > .05.”	“The study did not provide clear evidence of a difference.”	Non-significance is not proof of equivalence.
“The result was highly significant.”	“The estimate was statistically compatible with a difference, with p < .001.”	Magnitude should be discussed separately.
“X caused Y.”	“X was associated with Y.”	Use causal language only when the design supports it.
“This agrees with Smith.”	“The direction is consistent with studies in similar clinical settings.”	Comparison should focus on evidence, not author-by-author listing.

A Worked Nursing Discussion Example

Imagine a dissertation investigating whether structured bedside-handover training is associated with nurses’ communication-confidence scores. The final SPSS analysis shows that the trained group had a higher mean score, with a moderate effect and a confidence interval that did not include zero.

A developed discussion might begin by stating that the training group reported greater communication confidence. It would then evaluate the size and precision of the difference, rather than repeating the complete t-test output. Next, it would compare the result with nursing handover or communication research, paying attention to differences in clinical setting and training intensity. The writer could then consider whether repeated practice, standardisation or team feedback might explain the pattern, while making clear whether the study design can establish causation.

The paragraph would finish by considering whether improved confidence is enough to support practice change. Confidence is not identical to communication quality or patient safety. A careful implication might therefore recommend evaluating the training alongside observed handover quality or patient-safety indicators before wider implementation.

That is the difference between reporting a statistic and discussing it.

Final Checklist for Discussing SPSS Results

• Each discussion section starts from an important result, not from an SPSS table number.
• The direction and magnitude of the finding are clear.
• Confidence intervals or other measures of precision are considered where available.
• Effect size is discussed when it adds useful information.
• Statistical significance is not treated as clinical significance.
• Non-significant findings are interpreted without claiming proof of no effect.
• Correlation and observational regression are not described as causal.
• Previous nursing studies are synthesised by agreement, disagreement and context.
• Unexpected findings are explored rather than concealed.
• Limitations alter the strength of the interpretation.
• Practice recommendations stay proportionate to the evidence.

If your SPSS analysis is complete but the written interpretation still feels disconnected from the research question, our nursing dissertation data analysis service can help you review the output, interpretation and chapter alignment. For presentation rules and table selection, return to the SPSS results chapter guide.

Related Nursing Guides

• Missing Data in Nursing Research: SPSS Dissertation Guide
• Nursing Regression Analysis in SPSS for Dissertations

Conclusion

Knowing how to discuss SPSS results in a nursing dissertation means moving beyond output tables and p-values. Start with the substantive finding, examine its size and uncertainty, compare it with relevant nursing evidence, consider clinical meaning, test alternative explanations and finish with a proportionate implication. A strong discussion does not make the statistics sound more impressive. It makes clear what the evidence can and cannot support.

References

• Barría P., R. M. (2023). Use of research in the nursing practice: From statistical significance to clinical significance. Investigación y Educación en Enfermería, 41(3), e12. https://doi.org/10.17533/udea.iee.v41n3e12
• Davis, S. L., Johnson, A. H., Lynch, T., Gray, L., Pryor, E. R., Azuero, A., Soistmann, H. C., Phillips, S. R., & Rice, M. (2021). Inclusion of effect size measures and clinical relevance in research papers. Nursing Research, 70(3), 222–230. https://doi.org/10.1097/NNR.0000000000000494
• Lang, T. A., & Altman, D. G. (2015). Basic statistical reporting for articles published in biomedical journals: The Statistical Analyses and Methods in the Published Literature or the SAMPL Guidelines. International Journal of Nursing Studies, 52(1), 5–9. https://doi.org/10.1016/j.ijnurstu.2014.09.006
• Polit, D. F. (2017). Clinical significance in nursing research: A discussion and descriptive analysis. International Journal of Nursing Studies, 73, 17–23. https://doi.org/10.1016/j.ijnurstu.2017.05.002
• Wasserstein, R. L., Schirm, A. L., & Lazar, N. A. (2019). Moving to a world beyond “p < 0.05”. The American Statistician, 73(sup1), 1–19. https://doi.org/10.1080/00031305.2019.1583913
• von Elm, E., Altman, D. G., Egger, M., Pocock, S. J., Gøtzsche, P. C., & Vandenbroucke, J. P. (2007). The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: Guidelines for reporting observational studies. PLoS Medicine, 4(10), e296. https://doi.org/10.1371/journal.pmed.0040296