Power

Question 1

What is power?

Answer 1

The correct decision made, (rejecting H0 when in reality H0 is false) 1-beta

Question 2

How does the alpha level affect power?

Answer 2

Relaxed alpha level, e.g., 0.10 or 0.20, increases power but comes with the trade-off with increased Type I error (may not be worth it)

Question 3

How does the effect size affect power?

Answer 3

The greater the effect size, the higher the power. However, sometimes we cannot change the effect size.
e.g., people’s heights cannot be changed.

Question 4

How does variability influence power?

Answer 4

Variability: how spread out the curves are
Wider curves (higher variability) increases Type II error (decreases power). Type I stays the same because alpha level does not change.

Narrower curvers (lower variability), decreases Type II error (increases power).
e.g., it would be good to study homogenous population (more similar, lower variability) groups of individuals or types of effects, but may limit generalizability!

Question 5

What is the influence of sample size on power?

Answer 5

Increasing the sample size, the smaller the variability of the sampling distribution, therefore reduces the standard error of the mean.

Increased sample size = increased power

Question 6

What can a researcher do to increase power?

Answer 6

1. make a larger (but this increases probability of Type I error – not always a good idea)
2. increase the effect size (not always possible)
3. decrease variability of populations studies (not always possible or preferable)
4. increase sample size