Statistical testing

A1

A2

Sign test

Probability: 0.05 significance level (5%), 1% if drugs

Statistical testing determines whether a hypothesis should be accepted or rejected. Tells us whether variables relationships are statistically significant.

Statistical test gives calculated value to compare to critical value in tables

Which critical value?

How do you choose a stat test?

Main hypothesis = alternative hypothesis H1. Other = null hypothesis H0.

What are the tests?

Level of measurement

Repeated measures design

Nominal data/categorical

Looking for a difference not association

Need the signficance level desired

Number of partiicpants/N value

Directional or non-directional test

Ordinal: can order, rate subjectively, no equal intervals, unsfe, converted into ranks when testing e.g. 1st, 2nd

Interval: equal units, safe and parametric, measured and precise, most sophisticated

Nominal: categorical, discrete

One or two tailed

Degree of freedom (rows-1 x columns-1)

Level of significance

Unrelated = independant groups design

Repeated measures and matched pairs are related

Stat test

Mashed (Mann-Whitney)

With (Wilcoxon)

Come (Chi)

Swede (Spearmans rho)

Should (Sign)

Under (Unrelated-T)

Carrots (Chi)

Roast (Related-T)

Potatoes (Pearsons R)

If difference, which experimental design

Level of measurement

Difference or correlation