What to Do When Your Results Aren’t Significant (and How to Write About It Properly)

What to Do When Your Results Aren’t Significant (and How to Write About It Properly)
Dissertation Results

You ran your analysis, checked the output twice, and there it is: p > .05. Your result is not statistically significant.

Before you panic: a non-significant result is still a valid research finding. A dissertation does not need every hypothesis to be supported. What matters is whether you analyse, interpret and report the result correctly.

By the Analytivio team5 min readUpdated November 2026

The short answer

Report the result exactly as you found it, say the hypothesis was not supported, look at the effect size, confidence interval and sample size, and discuss possible reasons. Never change the analysis just to reach significance.

What a non-significant result means

A non-significant result means your analysis did not provide enough statistical evidence to reject the null hypothesis at your chosen significance level, commonly p < .05.

It does not prove there is no relationship or difference. If SPSS gives p = .218, the correct interpretation is that the finding was not statistically significant, not that your hypothesis was “wrong” or that the variables are unrelated.

Do not try to make the result significant

One of the biggest mistakes in data analysis is changing the analysis just to get a significant result. These practices undermine the credibility of your research:

×

Removing data without justification

Dropping inconvenient observations to change the outcome.

×

Testing until something works

Repeatedly trying different methods until one gives p < .05.

×

Selective reporting

Reporting only the favourable findings and hiding the rest.

✓

Report it as obtained

Present the result honestly and explain what it means for your research question and hypothesis.

Look beyond the p-value

Good interpretation does not stop at statistical significance. Consider the effect size, confidence interval, sample size and the direction of the relationship. A non-significant result with a meaningful effect size may still deserve discussion, especially when statistical power is limited. Our guide to effect sizes explains how to report them.

How to write it up in your dissertation

State the result clearly and neutrally, then explain it.

Example wording“The analysis did not identify a statistically significant relationship between X and Y, p > .05. Therefore, the hypothesis was not supported.”

Possible reasons to discuss

  • Sample size or sample characteristics
  • Measurement limitations
  • Insufficient statistical power
  • How the finding fits with earlier research

The lesson: your research does not become weaker because one hypothesis is not supported. Your interpretation becomes stronger when you report the result accurately.

Transparent reporting shows that your analysis follows the data, not the result you hoped for. Keeping a clear SPSS syntax file also makes it easy to show exactly what you tested.

Frequently asked questions

Is a non-significant result a failed dissertation?

No. A non-significant result is a valid finding. Examiners look at whether you analysed, interpreted and reported it correctly.

Does p > .05 mean there is no effect?

No. It means there was not enough evidence to reject the null hypothesis. It does not prove the effect is zero.

Should I remove outliers to get significance?

Only remove data for a justified, pre-planned reason, and report it. Removing cases just to change the result is not acceptable.

How do I explain a non-significant result?

Discuss the effect size and confidence interval, sample size and power, measurement issues and how the finding compares with previous studies.

Should I still report effect sizes?

Yes. Effect sizes and confidence intervals help readers judge whether the result may still be practically meaningful.

Unsure how to interpret your results?

Analytivio’s statisticians can review your SPSS output, check your analysis and help you interpret and report every result accurately.

Keep reading

This article is general guidance. The right metrics and methods depend on your industry, data and goals.

Tags:

2 Responses

Leave a Reply

Your email address will not be published. Required fields are marked *