(4) Lastly, we test for model stability and residual analysis to check for normality, autocorrelation and heteroscedasticity of the models residuals in the fourth step.
4) The fourth step involves checking for model stability and residual analysis to check for normality, autocorrelation and heteroskedasticity of the residual process.
An analysis of variance for normality was calculated using the Shapiro-Wilk test.
The distributions of all calculated values from conventional and topographic analysis were tested for normality using q q-plots and a Kolmogorov–Smirnov-test, and log-transformed when necessary.
For quantitative analysis of DNA methylation, percentages of DNA 5-methylcytosine and 5-hydroxymethylcytosine derived from the formulas listed above were transferred into the statistical analysis package GraphPad, verified for normality, and an analysis of variance (ANOVA) run to assay differences between experimental treatments.
When these normally distributed Kh values were combined into one data set for normality analysis, they became log-normal distribution.
All continuous variables were found to be normally distributed by analysis by Shapiro Wilks test for normality, apart from age, which was almost normally distributed and accepted for parametric testing with Student's t-test due to the central limit theorem, and alcohol consumption per week, which was not normally distributed.
For statistical analysis, data were evaluated for normality, equal variance and sphericity, and analyzed by one-way repeated-measures analysis of variance with multiple dependent measures (MANOVA).
Prior to analysis data was tested for normality using Cochran's test, and for homogeneity of variance using Shapiro Wilk's test.
Before statistical analysis, data were tested for normality and homoscedasticity using Lilliefors and Bartlett's tests, respectively.
