Comparing the mean score on the GREs of individuals before and after they took a prep class
Paired t-test
Goal of this test?
Relative frequency of successes in the population matches null expectation
Difference between binomial and Chi squared?
Binomial = only two outcomes
Chi squared = two or more outcomes
How many and what variable is observed?
Two categorical variables
Difference between paired t-test and independent samples t-test
Paired = both treatments applied to every sample unit
Independent = each group receives only one treatment
Is the proportion of individuals who chose chocolate, vanilla, or strawberry ice cream equal?
X2 test
Assumptions of the test?
Data are dichotomous (two parts) and individuals are independent
Assumptions of this test?
Random sample and expected frequencies are not too small
Assumptions of this test?
Random sample and expected frequencies are not too small (same as X2)
What test do we use the F value? What does it mean to be closer or farther from 1?
ANOVA
Closer = means are not different
Farther = means are different
Is there an association between handedness (left or right) and what state they are from (Pennsylvania, New Jersey, or New York)
X2 contingency test
How many and what variable is observed?
A single categorical variable
What does the X2 value measure?
Difference between observed and expected frequency
Goal of this test?
To see if the two variables are associated or independent
What is pseudoreplication? Give an example.
Non independent values analyzed as independent
Elephant tusks
Is the proportion of individuals who kick a soccer ball with their right foot equal to those with their left foot?
Binomial
If there is an equal probability, what is our test proportion?
0.5 (50%)
What does a large X2 value mean? Close to zero?
Large =observed and expected are different
Small/close to zero = observed and expected are not different
If a p-value is less than 0.05 and X2 is large, what does this mean?
The two variables are associated/not independent
What does the t-value mean?
How different the two means are from one another
Large t = bigger difference
Small t = less difference
Comparing the mean tree height of three different National Parks
ANOVA
Is there a test statistic? What two things should we include in our results statement?
No, sample size and p value
If you have two categories and a small sample size what should you do?
Use the binomial test, as there are fewer assumptions about low frequencies.
What is the preferred graph type for this test?
Stacked bar graph
Ways to reduce bias? Ways to reduce sampling error?
Bias = Control group, randomization, and blinding
Sampling error = Replication, balancing, and blocking