Effect Sizes

Question 1

How to obtain a partial eta squared on jamovi

Answer 1

Tick the partial n2 option under effect size, as pictured

Question 2

What is a partial eta squared and how is this different from p values

Answer 2

A measure of effect size

P value indicates probability that the data would have been observed if the effect didn’t really exist (ie if null was true) so we want it as low as possible

Effect sizes indicate the magnitude of effect - which can be used to say something about its importance in terms of explaining a DV

Question 3

Why are effect sizes important

Answer 3

Unlike p values, they’re unaffected by sample size (where small and maybe trivial effects can be significant in large samples). this…
- allows comparison of an effect across several studies of varying sample size and significance values
- provides a good basis for meta-analysis
- is a basis for power calculations to estimate sample sizes needed to find an effect

Question 4

Effect sizes vary in terms of range and meaning of values - True or False?

Answer 4

True

various effect size measures aren’t directly equivalent to each other and need to be converted. 0.2 has a very different meaning depending on the type of effect size

Question 5

What are the estimates for what a small/medium/large effect size is (for cohen’s d, r, eta squared, generalised eta squared)

Answer 5

Answer pictured

Question 6

What is Eta Squared - brief summary of issues and solution

Answer 6

The proportion of total variance accounted for by a given effect

Basically it has its drawbacks and partial eta squared is the solution to this

Question 7

Two caveats for partial eta squared

Answer 7

It’s a biased measure (alternatives are omega squared or epsilon squared)

Effect sizes vary depending on methodological features