statistical testing

unrelated t-test

• testing for a difference

• independent measures

• interval

related t-test

• testing for a difference

• repeated measures

• interval

chi squared

• testing for difference or association

• independent measures

• nominal

sign test

• testing for a difference

• repeated measures

• nominal data

spearman’s rank order correlation coefficient

• testing for a correlation

• related pairs of scores

• ordinal

pearson’s production moment correlation coefficient

• testing for a correlation

• related pairs of scores

• interval

interval data

data measured using units of equal intervals eg., the time taken to complete a questionnaire

ratio data

where there is a true zero point e.g., age

ordinal data

data ordered in some way e.g., listing football teams in an order of liking

nominal data

data in separate categories e.g., grouping people to their favourite football team

mann whitney u tets

• testing for a difference

• independent measures

• ordinal

wilcoxon

• testing for a difference

• repeated measures

• ordinal

what is the aim of inferential statistical testing?

to discover if results are statistically significant

what is significant?

unlikely to have occured through chance

what is the usual level of significance?

5%

how is significance expressed?

in terms of null hypotheses

what is a null hypothesis?

where there is no stated difference between variables

what is a type I error?

• false positive

• too lenient when setting probability level

• too much margin for error

• e.g., p as >0.1

• incorrectly rejecting null hypothesis

what is a type II error?

• false negative

• too stingy when setting probability level

• e.g., p as <0.01

• incorrectly accepting null hypothesis