Confidence Levels in Statistical Analyses. Analysis of Variances. Case Study.

Abstract

Applying a statistical test to check statistical assumptions offers a positive or negative response regarding the veracity of the issued hypothesis. In case of variance analysis it’s necessary to apply a post hoc test to determine differences within the group. Statistical estimation using confidence levels provides more information than a statistical test, it shows the high degree of uncertainty resulting from small samples and builds conclusions in terms of "marginally significant" or "almost significant (p being close to 0,05) . The case study shows how the statistical estimation completes the application form the analysis of variance test and Tukey test.

Keywords

• analysis of variance
• tukey test
• statistical estimation
• confidence intervals
• graphic representation