t-tests

Question 1

Is it possible to totally avoid making a Type I error?

Answer 1

 Only if you test the entire population!
 There is always a chance of a Type I error when inferring a
significant effect from sample to population

Question 2

What does a positive skew show?

Answer 2

Low scores over represented

 ‘Tail’ in the +ive direction

Question 3

What does a negative skew show?

Answer 3

-High scores over represented

 ‘Tail’ in the –ive direction

Question 4

Homogeneity of variance?

Answer 4

Want variance to be the same (or similar) in all groups

• Homogeneity = same
• Heterogeneity = different

Question 5

Why no homogeneity of variance in repeated measures?

Answer 5

• Parametric assumption: Independence of observations
• One data point should not influence another
• BUT – repeated measures, so can’t avoid this!
• If my memory is bad it will be bad in both conditions
• In repeated designs random variance is reduced
• No individual differences between condition
• In repeated designs there is a different assumption - Sphericity: