Transcript Things you must know for exam 3

Things you must know for exam
3
• Sampling distribution of the mean and
proportion
– When to use it, what is the model (mean, SD etc)
– How to use invNorm and Normalcdf to answer
questions about the normal model
– How sample size affects spread (standard
deviation) of sampling distributions
Ch 19
• Formula for computing the sample size n
needed to guarantee a certain confidence
level and a certain margin of error.
– n = (z*)2 (p^)(q^) / (ME)2
– z* found using invNorm for confidence interval
desired
• Relationship between confidence interval and
margin of error
Hypothesis testing
• Must must must know how to write hypothesis…
• Must must must know the names of the tests and which
one to do when.
–
–
–
–
–
–
–
–
–
–
–
1-prop Z-interval
2-prop z-interval
1-prop Z test
2 prop zTest
T-test (one sample)
Tinterval (also one sample
2-sample T test
2-sample T interval
T-Test and T-interval on pairwise differences (for paired data)
X2 test for Goodness of Fit and X2 test of independence
(and don’t forget we also looked at sampling distributions when
we KNOW true population parameters – mean, sd, proportion)
Hyp testing continued
• What conditions must be satisfied and what
those conditions mean (understand what you
are actually checking
– E.g. 10% condition, success-failure condition,
randomization condition, counted data condition,
paired data condition, nearly normal condition,
etc
• How to draw (sketch) distributions (indicating
test statistic) and shade the p-value
– E.g. shade p-value region of T-test
– Shade p-value region of Z-test
– Chi sq test etc.
Type I and Type II errors
• When they occur, how to write about them in
terms of the situation
How to accurately interpret results of a
hypothesis test
•
•
•
•
•
Can only reject or fail to reject the null
Don’t talk about accepting anything
Only talk about evidence suggests
Ok to talk about statistically significance…
Reject null
– Sufficient evidence to suggest that alternate hyp is
true
• Fail to reject null
– Was not sufficient evidence to conclude that alternate
hypothesis is true.