statistical tests Flashcards
parametric tests
-relies upon the assumption that the data you want to test is (or approximately is) normally distributed
-interval level data
non-parametric test
-determine the value of data points by assigning + or - signs based on the data ranking
-less powerful than parametric tests
when use a related T- test
-used in studies with a repeated measures or a matched pairs design, where the data meets the requirements for a parametric test
-interval data
-N in formula means sample size
when use unrelated T- test
-with independent measures
-interval data (parametric)
when use sign test
-test of difference between pared data e.g. repeated measures
-nominal data
-calculated value must be equal to or less than critical
when use chi-squared test
-difference or association
-only nominal, unrelated data
-calculated value must be equal to or greater than critical
when use wilcoxon
-test of difference
-ordinal data
-related data
-calculated value must be equal to or less than critical
when to use spearmans rho
-test of association
-ordinal data
-looking for correlation
-calculated value must be equal to or higher than critical
-order of points, do all increase
when to use pearsons r
-test of association
-correlation
-interval data
-calculated must be equal to or higher than critical
-close to straight line