statistics

greenspun.com : LUSENET : History & Theory of Psychology : One Thread

If you only use one participant who took the experimental conditions several times,could you still use a within t-test or anova?

-- maggie (lo_maggie@hotmail.com), April 12, 2004

Answers

I am afraid you can't use parametrics tests like T or ANOVA when employing designs with n equal to 1. As the word implies these tests assume that your data have certain properties, such a normal distribution, among others. Even tough there exist some test aimed to analyze single participant studies, it is a widespread tradition to visually inspect the plot of data and extract conclusions upon it.

-- Ricardo Marcos Pautassi (rpautassi@immf.uncor.edu), April 12, 2004.

They would not be "independent observations," which is an asumption of most such tests.

-- Christopher Green (cgreen@utoronto.ca), April 12, 2004.

Moderation questions? read the FAQ