Except in few simple cases, classical statistical methods are approximate or asymptotic methods, whereas Exact Methods are based on exact probability statements and exact expected values. A number of widely used Exact Parametric Methods can be performed using freely distributed software. Non standard methods could be carried out by writing code using, say R. In particular, you can (at your own risk; accuracy is not assured)
Use the gp-test option in one-way Tests package at One-Way ANOVA to perform One-Way ANOVA
To perform Two-Way ANOVA under unequal variances, use the R code gpTwoWay . It is also avialble from "gpTwoWay" option of the CRAN R library Package at twowaytests
Use R Code to perform Mixed Model Regressions . Soon, this will also be available as an R Package
Download XPro Absolutely Free:
XPro for Windows
Notes:
Lately Windows lack backward compatibility, and so a number of features of XPro will not work.
In particular, (i) Install Xpro in a folder outside of Program File folders
(ii) Import Data as a tab delimited text file (but can save as .x file) using Data menu, (iii) get help from
Xpro Help, a pdf file.
Detect Efficacy and Safety of experimental drugs early (if truly significant)
XPro procedures are specifically
designed to provide practitioners with powerful
tests while avoiding misleading conclusions.
Statistical techniques for exact parametric inference in numerous linear models are documented in books:
Why Waste Hard Earned Data on Approximate Methods?
Perform exact parametric methods to:
Minimize cost and duration of experiments
Avoid misleading conclusions
Avoid costly investments due to false positive results
Exact Statistical Methods for Data Analysis
from
Springer Series in Statistics
Read Free
Generalized Inference in Repeated Measures:
Exact Methods in MANOVA and Mixed Models
from
Wiley Series in Probabilty & Statistics
About Generalized Inference and Exact Methods
Generalized P-Values
Generalized Estimation
Generalized Intervals
XPro FAQ
Superior Power Demo
Copyright 2011-2020, Weerahandi.org
Contact