Scholars will commonly exude a sigh of relief upon finding a smaller p -value to the statistical tests than 0.05. This statistically significant finding appears to ensure a publication and confirm the hypothesis. But what will occur at a value of p exceeding 0.05? Most researchers perceive non-significant results as failures thus under-developing or demoralizing discussion sections.
As a matter of fact, significant results are useful data. They refute theories, point out complexities and stifle the file drawer problem where only positive findings can escape the light of day. The skill that you can write a strong and balanced discussion even when outcomes are not up to the threshold of significance is a quality of a skillful Research Article Writing. You have to stop thinking about how to prove your suspicions but rather describe the data that you have most thoroughly.
Re-examine and Justify Your Methodology
When this becomes a non-significant finding, it is not necessarily an indication that your hypothesis is incorrect. It usually indicates that your research has not found an effect. The first thing you need to do in your discussion section is to review and analyze your methods stringently to find possible explanations.
- You need to examine power: Was your study of adequate statistical power to be able to observe a meaningful effect in case one was really present? Non-significant results are often due to small sample sizes since the power is low. Measure and report your measured power.
- You evaluate measurement error: Have your measures measured the constructs? A true effect can be obscured in case of high variance or poor reliability in your measures. Talk about possible weaknesses of your tools of measurement.
- You cross-examine implementation: Who were the inconsistencies in the intervention delivery or data collection? The effect may be attenuated by external factors or low fidelity in the treatment resulting in p > 0.05.
Actionable Tip: Before concluding about the hypothesis, you should spend a certain paragraph defending the rigor of your design. This makes your Research Article Writing better.
Put into perspective with Existing Literature
- Your non-significant finding is not a one-time occurrence; this is part of the overall academic dialogue. You have to relate your findings to the body of knowledge.
- You are compared to the past research. Compare your result with studies that did identify a significant effect. Determine and describe the disparities in design, population characteristics, geographic location, or operationalization of a variable. These distinctions help give one a good idea about the boundary conditions of the effect.
- You match conflicting studies and bring out other published studies that also did not find any significant relationship. Your result furnishes important replication that adds to the evidence of non-universality of effect.
- You comment on theoretical implications; How will your non-significant finding change or correct the existing theoretical model? Does your finding imply a necessary condition to the occurrence of the effect which the theory ignores?
When there is no significant result, the field is likely to know that the effect is not present, which is as important as knowing that it is present. Do not present your results as a failure, it is a vital part of the puzzle.
Take into account the Practical or Clinical Significance
There is a difference between statistical significance (p <0.05) and practical significance. Practical significance is defined as the actual significance or size of an effect. When your p-value is not significant, you need to take the stress of your discussion to the magnitude of your effect that you actually did get.
- You give effect sizes: You should report and discuss the effect size (e.g. Cohen d, odds ratio, R 2 ) whether you have a p-value or not. The effect is very small and though it is not significant, it supports the idea that this relationship is trivial.
- You identify confidence intervals: See the confidence interval (CI) of your effect size. When the CI is broad with zero as well, your data are very scattered and the actual impact might be amounting to nothing or no impact at all. When the CI is small and encloses zero, then you can be sure there is a close to zero true effect. The existence of a narrow CI about a trivial effect is a very robust result.
Present Future Research Directions
The non-significant results need to be discussed proactively. You have to shift to explaining what was not discovered and hinting on how the next study can advance the field.
- You suggest changes in design: Suggest how to enhance statistical power (large sample), decrease measurement error (better measures), or isolate the important mechanism (a more specific intervention).
- You suggest alternative hypotheses: According to your current data and literature review, propose totally new hypotheses or moderator/moderator variables which should be explored by future researchers. It may be that the non-finding of yours is due to a significant non measurable variable.
The most important tips in writing Research articles on non-significant results are the ones that are authoritative, analytical, and forward-looking. Your results are not something you are apologetic about; you interpret them. Contact Fast Publications. We help you write the non-significant results.
