Exam One - Hypothesis testing Flashcards
hypothesis testing
inferential statistics provide a mechanism to reject that an observed effect is a chance occurrence
alpha
describes the possibility of concluding two samples are different when they are not
type 1 error
describes the possibility of concluding two samples are different when they are not
alpha is type ___ error
one
means for samples outside the 95% CI are ____
rare
your tolerance for declaring a sample mean “significantly different” if it falls outside the expected range by chance is called the…
alpha level
Beta level
probability of incorrectly concluding that two samples are NOT different
probability of incorrectly concluding that two samples are NOT different
type 2 error
beta level is type _____ error
two
Power =
1 - beta
power is the probability of…
correctly concluding that the groups are different
typical power level acceptable for clinical research is
80%
What are the factors affecting power?
- alpha level
- difference between group means
- within group variability
- sample size
How does alpha level affect power?
- raising a increases power (also increases risk of type one error)
- one tailed tests also increase power
How does the difference between group means affect power?
- larger group differences increases power
how does within group variability affect power?
- decreased variability increases power
- increased variability decreases power
How does sample size affect power?
- more subjects = more power
- narrow sampling distributions
- difference between means of large samples is more likely to be recognized as atypical
prior to data collection, researchers perform a sample size estimate to demonstrate the study has sufficient _______
power (the ability to identify a true difference)
What factors affect calculating sample size?
- alpha level
- difference between group means
- within group variability
- desired power (1-beta)
When do we calculate the effect size?
used for sample size estimation
- estimates the magnitude of a treatment effect or difference between groups
ES =
|diff b/t means|/ avg SD
how to interpret ES
<0.4 = small
0.4-0.8 = medium
>0.8 = large
increasing mean differences _______ effect size
increases
increasing sample variability _____ effect size
decreases
increasing sample size _____ effect size
no effect
increasing effect size _____ sample size estimates
decreases
increasing effect size _______ study power
increases
increase type one error tolerance ______ study power
increases
increasing sample size will ________ study power
increase
