Week Nine Flashcards
What is delta
A value used in calculating power that combines effect sizes (gamma or d) and some function of the sample size (N) depending on the test used.
Enables us to calculate power for different sample sizes: therefore estimate appropriate sample sizes.
What is power
The ability to detect a difference or find a relationship where on exists
What is SE formula
SD divided by the N squared
What is a type one error
Rejecting the null h when the null h is correct so finding a difference when there isn’t one
a used to find probability of a type one error
What is a type two error
Accept the null hypotheses when there is a difference / relationship
Find no difference when there is a difference
B (slope) helps show probability of a type two error
What influences power
Effect size: true difference between the null hypotheses and research hypotheses
Significance level (alpha): probability that results are due to sampling error (usually varies from .01 to .05)
Sample size: the greater the sample size the more power you have greater ability to detect a significant effect
Type of statistical test: parametric test has more power than non-parametric test
Research design: repeated measures design have more power because within participants variability is reduced.
One or two tailed test: more power in a one tailed test (need a larger sample size to reach critical value for two-tailed test)
How can you increase power
Large effects (y)
Large samples
Research design choice
Varying sample size is easiest way to increasing power
