GraphPad Statistics Guide

PRINCIPLES OF STATISTICS

PRINCIPLES OF STATISTICS

Previous topic Next topic No expanding text in this topic  

PRINCIPLES OF STATISTICS

Previous topic Next topic JavaScript is required for expanding text JavaScript is required for the print function Mail us feedback on this topic!  

The first half of this Guide reviews general principles of statistics, and is not at all specific to GraphPad Prism. It includes discussions of some important issues that many statistical text books barely mention, including:

The problem of multiple comparisons  and the many ways you can get trapped by multiple comparisons.

Testing for equivalence

The danger of using outlier tests with lognormal distributions and the problem of masking which can make it harder to find two outliers than to find one.

Why it doesn't make sense to automate the decision to use a nonparametric test or not.

The distinction between SD and SEM, and when to display each.

The advantages of reporting confidence intervals.

The most common misunderstanding about P values, and other misunderstandings.

Why you can't peek at the results and add more subjects if the results are not quite significant yet.

A simple analogy to understand statistical power.

A set of analysis checklists. Each checklist lists questions you should ask yourself before accepting the results of a statistical analysis.

The second half of the guide explains how to analyze data with Prism. Even so, much of the content explains the alternative analyses and helps you interpret the results. These sections will prove useful no matter which statistical program you use.