Statistical Power/underpowered Stats

Question 1

Power curve

Answer 1

Horizontal axis is the true probability of an event, the vertical axis is the probability of concluding the correct outcome???

Question 2

Power for any hypothesis test

Answer 2

The probability that it will yield a statistically significant outcome.

Affected by 3 factors

1. Size of bias looking for.
2. The sample size
3. Measurement errors

Question 3

Underpowered

Answer 3

If a trial isn’t powerful enough to detect the effect it is looking for.

Usually 0.8
Meaning an 80% chance of detecting a real effect of the expected size.

Question 4

Assurance

Answer 4

Determines how often the confidence interval must beat our target width.

Choose a sample size so that the confidence interval will be suitably narrow (99%)(insert assurance number here)

Question 5

AIPE - Accuracy In Parameter Estimation

Answer 5

Sample size selection methods based on assurance.

Question 6

Statistical power

Answer 6

Is the probability that it will distinguish an effect of a certain size from pure luck.