FIN Flashcards
(34 cards)
If something is correlated is it associated? If it is associated, are they correlated?
Yes. Correlation is association. No. Association does not mean correlation.
SRS
Random # Generator
Stratified
By year, 10 F,S,J,S
Biased Estimator
Asking everyone outside gym to find % of people who workout.
Unbiased Estimator
Pile of stats centered at parameter
Assign treatments correctly
Cards, shuffle, choose 20
Experiment V Study: Random #s
In a study choose subjects to survey, in an experiment assign subjects to treatments.
Experiment V Study: Inference
In a study you infer about the population, in an experiment infer about treatment.
Experiment V Study: Manipulating
No manipulation in studies, only in an experiment.
Experiment V Study: Prove Causation
Can’t prove causation in studies, only in experiments.
Volunteers are ok where?
In experiments.
Describe completely randomized
All subjects into one hat, pick treatment
Describe randomized block
More hats, female, male, etc and pick.
Describe matched pairs
Fuel efficiency, both fuel A and B for a month and compare.
Law of large #s?
In the long run, relative frequency settles to true probability.
What is probability?
The long run relative frequency.
Can independent events be disjoint?
No
When combining random variables
Add variances
What happens to a pile of stats when we take larger samples?
X bars or P hats will get closer to parameter.
Difference between SE and SD
SE is distance a statistic is from the mean.
SD is the distance a datum is from the mean.
Difference between dist of a sample and a sampling dist?
Dist of a sample is a histogram of the data
A sampling dist is made from sample statistics.
CLT says
No matter shape of pop, large enough samples, close to normal
What happens to models as N gets larger
Models look closer to normal value
What is a critical value?
How many SE you reach out