Publication:

Advanced statistical methods for the analysis of data in cosmetic science studies

Statistical analysis is fundamental to cosmetic research, enabling rigorous evaluation of product safety, efficacy, and consumer perception. Despite its importance, the selection and application of statistical methodologies in cosmetic studies remain varied and lack a comprehensive synthesis. This systematic review examines the statistical methods employed in clinical cosmetic research, focusing on both commonly used practices and advanced biostatistical techniques. Following PRISMA guidelines, a literature search was conducted using PubMed with the keywords “statistical + cosmetic products,” restricted to studies published between January 2000 and April 23, 2025. After applying inclusion and exclusion criteria, 37 full-text articles involving human subjects and reporting statistical methodologies were included. Data on study design, sample size, outcome measures, and statistical methods were extracted and analysed. Parametric tests such as paired t-tests, independent t-tests, and ANOVA were commonly used for continuous outcomes, while non-parametric alternatives, including the Wilcoxon signed-rank test, Mann-Whitney U test, and Kruskal-Walli’s test, were applied when normality assumptions were not met. Longitudinal data analysis frequently employed repeated measures ANOVA and mixed-effects models, with Bonferroni correction for multiple comparisons. Categorical data were analysed using chi-square tests, McNemar’s test, Fisher’s exact test, and logistic regression. More advanced methods, such as ANCOVA, Bland-Altman plots, two one-sided tests (TOST) for bioequivalence, and meta-analytic techniques using Cochran’s Q and I² statistics, were also identified. Overall, cosmetic research demonstrates broad and largely appropriate statistical application, although improved reporting of assumptions and model specifications is recommended to enhance transparency and reproducibility.