Supporting code for ''Considerations for missing data, outliers and transformations in permutation testing for ANOVA, ASCA(+) and related factorizations''

Oliver Polushkina Merchanskaya, Michael D. Sorochan Armstrong, Carolina Gómez Llorente, Patricia Ferrer, Sergi Fernandez-Gonzalez, Miriam Perez-Cruz, María Dolores Gómez-Roig, José Camacho

missing_data

Code related to the simulations for missing data, in Section 4 of the paper. As a function of an increasing percentage of missing data, a comparison between unconditional mean replacement (UMR) conditional mean replacement (CMR) and permutational conditional mean replacement are considered using random data with and without a simulated interacting term.

power_curves_analysis

Code related to the power curve analysis for different types of randomized data: normally distributed, uniformly distributed, exponentional cubed data, and uniformly distributed with one outlier. Visualization of the effects of different distributions on the semi-parametric estimation of apparent significance as a function of effect size.

simulated_example

Code related to the simulatino of two factors, with one known a-priori to be significant and another insignificant. Comparison of the drops in Sum of Squares (SS) as a function of the different missing value imputation methods for 5% missing data.