Stats Flashcards
+LR
change in likelihood that a patient has the condition after a positive test
-LR
change in likelihood that a patient does not have condition after a negative test
t-tests
statistical hypothesis test used to determine if 2 sets of data are significantly different from one another
Can be Unpaired/Paired or 1 tailed/2 tailed
Null Hypothesis
you are always testing null with p-values
there is no significant difference between groups
u1 = u2
Alternate Hypothesis
there is a significant difference between groups OR group 1 is significantly greater/less than group 2
u1 does not equal u2
u1>u2 OR u1<u2
Research Hypothesis
could be either the null of the alternative hypothesis
you initial guess when starting an experiment
Unpaired/Independent samples t-test
Groups are independent, not the same people
Paired/dependent samples t-test
groups are the same people, generally pre vs post test
1 Tailed T-Test
directional
looks for an increase or decrease, testing either greater OR less
2 Tailed T Tests
looks for a change
tests both greater AND less than
stronger test
Alpha Level
value you set as the researcher, usually .05
the probability that you will commit a type 1 error
a lower alpha level means you are taking a smaller risk that you will report an effect when there is actually no effect
Type 1 error
reporting an effect when there is no effect
False positive
Type 2 error
reporting no effect when there is an effect
False negative
p-value
value that is obtained with t-test
compared to alpha level to determine significance
If p value is < alpha level
reject the null, significant difference between groups
if p value is > alpha level
fail to reject the null, no difference between the groups
Hypothesis rules
your options are: fail to reject the null, reject the null
you NEVER accept the null
you NEVER accept or reject the alternative hypothesis
p-hacking
the more t-tests you run, the greater chance you will get a significant result
can be abused by researchers
One-way ANOVA
used to determine whether there are any statistically significant differences between the means of three or more independent/unrelated groups
used with unpaired t test
Repeated Measures ANOVA
used to determine if there are significant differences within the same group over time
used with paired t test
time is almost always an independent variable
Post Hoc Testing
tells you which groups or time points are different after doing a pvalue and t-test
Power
the probability that a test of significance will pick up on an effect that is present, probability that a test of significance will detect a deviation from the null hypothesis, should such a deviation exist
1-Type 2 Error = power
