Parametric versus non-parametric statistics in the analysis of randomized trials with non-normally distributed data
- PMID: 16269081
- PMCID: PMC1310536
- DOI: 10.1186/1471-2288-5-35
Parametric versus non-parametric statistics in the analysis of randomized trials with non-normally distributed data
Abstract
Background: It has generally been argued that parametric statistics should not be applied to data with non-normal distributions. Empirical research has demonstrated that Mann-Whitney generally has greater power than the t-test unless data are sampled from the normal. In the case of randomized trials, we are typically interested in how an endpoint, such as blood pressure or pain, changes following treatment. Such trials should be analyzed using ANCOVA, rather than t-test. The objectives of this study were: a) to compare the relative power of Mann-Whitney and ANCOVA; b) to determine whether ANCOVA provides an unbiased estimate for the difference between groups; c) to investigate the distribution of change scores between repeat assessments of a non-normally distributed variable.
Methods: Polynomials were developed to simulate five archetypal non-normal distributions for baseline and post-treatment scores in a randomized trial. Simulation studies compared the power of Mann-Whitney and ANCOVA for analyzing each distribution, varying sample size, correlation and type of treatment effect (ratio or shift).
Results: Change between skewed baseline and post-treatment data tended towards a normal distribution. ANCOVA was generally superior to Mann-Whitney in most situations, especially where log-transformed data were entered into the model. The estimate of the treatment effect from ANCOVA was not importantly biased.
Conclusion: ANCOVA is the preferred method of analyzing randomized trials with baseline and post-treatment measures. In certain extreme cases, ANCOVA is less powerful than Mann-Whitney. Notably, in these cases, the estimate of treatment effect provided by ANCOVA is of questionable interpretability.
Figures
![Figure 1](https://www.ncbi.nlm.nih.gov/pmc/articles/instance/1310536/bin/1471-2288-5-35-1.gif)
![Figure 2](https://www.ncbi.nlm.nih.gov/pmc/articles/instance/1310536/bin/1471-2288-5-35-2.gif)
![Figure 3](https://www.ncbi.nlm.nih.gov/pmc/articles/instance/1310536/bin/1471-2288-5-35-3.gif)
![Figure 4](https://www.ncbi.nlm.nih.gov/pmc/articles/instance/1310536/bin/1471-2288-5-35-4.gif)
![Figure 5](https://www.ncbi.nlm.nih.gov/pmc/articles/instance/1310536/bin/1471-2288-5-35-5.gif)
Similar articles
-
Imputation strategies for missing continuous outcomes in cluster randomized trials.Biom J. 2008 Jun;50(3):329-45. doi: 10.1002/bimj.200710423. Biom J. 2008. PMID: 18537126 Review.
-
A simple sample size formula for analysis of covariance in randomized clinical trials.J Clin Epidemiol. 2007 Dec;60(12):1234-8. doi: 10.1016/j.jclinepi.2007.02.006. Epub 2007 Jun 6. J Clin Epidemiol. 2007. PMID: 17998077
-
Randomized clinical trials with a pre- and a post-treatment measurement: repeated measures versus ANCOVA models.Contemp Clin Trials. 2007 Nov;28(6):713-9. doi: 10.1016/j.cct.2007.04.002. Epub 2007 Apr 21. Contemp Clin Trials. 2007. PMID: 17524958
-
Statistics in review Part I: graphics, data summary and linear models.Crit Care Resusc. 2007 Mar;9(1):81-90. Crit Care Resusc. 2007. PMID: 17352672 Review.
-
Robustness and power of analysis of covariance applied to ordinal scaled data as arising in randomized controlled trials.Stat Med. 2003 Apr 30;22(8):1317-34. doi: 10.1002/sim.1433. Stat Med. 2003. PMID: 12687657
Cited by
-
Turning pain into strength: prosocial behaviours in coping with trauma.Eur J Psychotraumatol. 2024;15(1):2330302. doi: 10.1080/20008066.2024.2330302. Epub 2024 Apr 4. Eur J Psychotraumatol. 2024. PMID: 38573118 Free PMC article. Clinical Trial.
-
Single- and Dual-Task Gait Performance in Patients With Open-Angle Glaucoma: A Cross-sectional Study.Transl Vis Sci Technol. 2023 Nov 1;12(11):31. doi: 10.1167/tvst.12.11.31. Transl Vis Sci Technol. 2023. PMID: 38015169 Free PMC article.
-
Empagliflozin improves high-sensitive cardiac troponin-I and high-density lipoprotein cholesterol in patients with type 2 diabetes mellitus and coronary artery disease: a post-hoc analysis of EMPA-CARD Trial.J Diabetes Metab Disord. 2023 Sep 22;22(2):1723-1730. doi: 10.1007/s40200-023-01305-2. eCollection 2023 Dec. J Diabetes Metab Disord. 2023. PMID: 37975102
-
The impact of age-relevant and generic infographics on knowledge, attitudes and intention to attend cervical screening: A randomized controlled trial.Br J Health Psychol. 2024 Feb;29(1):204-220. doi: 10.1111/bjhp.12695. Epub 2023 Sep 28. Br J Health Psychol. 2024. PMID: 37770383 Free PMC article. Clinical Trial.
-
A systematic review and meta-analysis on the efficacy of glibenclamide in animal models of intracerebral hemorrhage.PLoS One. 2023 Sep 27;18(9):e0292033. doi: 10.1371/journal.pone.0292033. eCollection 2023. PLoS One. 2023. PMID: 37756302 Free PMC article.
References
-
- Altman DG. Practical Statistics for Medical Research. London: Chapman and Hall (monograph)
-
- Jekel JF, Katz DL, Elmore JG. Epidemiology, Biostatistics and Preventive Medicine. Philadelphia, W.B. Saunders Company; 2001.
-
- Heeren T, D'Agostino R. Robustness of the two independent samples t-test when applied to ordinal scaled data. Stat Med. 1987;6:79–90. - PubMed
-
- Sawilowsky SS. Comments on using alternative to normal theory statistics in social and behavioural science. Canadian Psychology. 1993;34:432–439.
-
- Zimmerman DW, Zumbo BD. The effect of outliers on the relative power of parametric and nonparametric statistical tests. Perceptual and Motor Skills. 1990;71:339–349.
MeSH terms
LinkOut - more resources
Full Text Sources