12: MORE STUFF

Question 1

power

Answer 1

• the ability to find a statistically significant effect (difference or relationship) when one exists
• measured on a scale of 0 to +1, where 0 = no power
• Power = 1- β
• Power = .1 means you have a 10% chance of finding an effect (if one exists)

low power: 0 - 0.3
high power: 0.8 - 1

Question 2

factors influencing power

Answer 2

• effect size (small effect size - hard to find effect)
• criterion significance level (alpha)
• sample size (small bad)
• statistical test (better with parametric)
• subjects design (RM>independent)
• 1 or 2 tailed hyp (1 better)

Question 3

calculating power

Answer 3

need: no. of Ps, effect size obtained, criterion significance level (alpha)
- .8 recommended
- uesful to determine how many Ps we will needs to obtain enough power to minimise type 2 error