All Rights Reserved (wwww.statconsultingzoo.com) Statistical Consulting,Z.0.0 Copyright@20112012 J.Wrotniak Email: yvucic@hotmail.com

Statistical Methods

Statistical Procedures applying to continuous and categorical data
 Descriptive statistics (mean, median, mode, SD, SE, 95%CI for mean)
 ttest(paired sample or independent sample) with the appropriate assumptions
 Chisquare test applied for 2 X 2 tables and used only for small sample sizes
 MantelHaenszel ChiSquare test stating a linear association between the row variable and the column variable with both variables lying on the ordinal scale
 CochranArmitage test testing for trend in binomial proportions, appropriate for twoway table where one variable has two levels and the other variable is ordinal
 Fisher's Exact Test defining association between row and column tables that is appropriate for small sample sizes and sparse tables. That test does not depend on any largesample distribution assumptions.
 McNemar test applied to the binomial distribution for 2 X 2 tables, appropriate when analyzing data from matched pairs of subjects with dichotomous (yesno) response
 Measures of Association appropriate of the ordinal variables: gamma, Kendall's taub, Stuart's tauc and Somers'
 Anova performing analysis of variance Oneway, Twoway and multivariate analysis of varianceManova, which performs multivariate analysis of variance when multiple dependent variables are listed in the model
 MIXED Models adhering to the randomized complete blocks design having one fixed effect(treatments) and one random effects(blocks)
 GLIMMIX procedure fitting statistical models to data with correlations or nonconstant variability and response variable not necessarily normally distributed
 CATMOD procedure fitting linear models and performing analysis of variance and repeated measures analysis of variance for categorical responses
 LOGISTIC procedure fitting logistic regression models for binomial and ordinal outcomes
 CANCORR procedure fitting continuous data to correlate simultaneously several dependent variables and several independent variables
 Discriminant Analysis procedure classifying observations where dependent variable is categorical and independent variables are continuous
 Time Series Analysis (forecasting, ARIMA, ARCH,GARCH models)

Nonparametric Methods requiring the minimal or none assumptions about the form of the distribution of the population
 Wilcoxon ranksum test carries by doing permutation test on the ranks and using the sum of ranks of one of the treatments as the statistics, it is performed to do the pairwise comparisons
 KolmogorovSmirnov test defines the maximum absolute value of the difference between the two treatments
 KruskalWallis test detects differences among treatments
 Friedman's test used for larger number of treatments when performing means comparisons

