How do you deal with non significant results?
5 tips for dealing with non-significant results
- #1: Perform an equivalence test.
- #2 Collaborate to collect more data.
- #3 Use directional tests to increase statistical power.
- #4 Perform sequential analyses to improve data collection efficiency.
- #5 Submit a Registered Report.
- Read next:
What are non significant results?
Null or “statistically non-significant” results tend to convey uncertainty, despite having the potential to be equally informative. When the probability does not meet that condition, the program result is null, i.e. there is no statistically significant difference between the treatment and control groups.
Is there a publication bias in a meta-analysis?
Publication bias is a well-recognised issue in meta-analysis. As statistically significant studies are more likely to be published, it is likely that there are several non-significant studies left languishing in researcher’s file-drawers for any given research area, which cannot contribute to meta-analyses.
Are there journals that accept a meta analysis?
Importantly, paper acceptance is not contingent on the statistical significance of results. As of August 2018, there are 42 journals that offer Registered Report meta-analysis, with the majority of these journals publishing research from the biobehavioral sciences. Publication bias is a well-recognised issue in meta-analysis.
What are some common mistakes in meta-analysis?
“As hypothesised, a meta-analysis synthesising eight effect sizes from seven studies suggests that social media usage is positively associated with self-reported levels of anxiety [r = .073, 95% CI ( .004, 0.142); p = .038; Fig. 1A]. Egger’s regression test suggested no evidence of publication bias (p = .19; Fig. 1B).”
How to calculate statistical significance in a meta-analysis?
A relatively simple funnel plot modification, called a contour-enhanced funnel plot, can help better understand the risk of publication bias compared to small study bias. For a given range of effect sizes and variances, it is possible to calculate levels of statistical significance at every point.