UNIT 6 Flashcards
What are the conditions that have to be met in order to use a t model?
random, 30 OR.. sample normalish
for means, what is sample size condition
30 OR sample looks normalish
How can you check for normalish?
a diagonal straightish normal prob plot, a unimodal symmetric historgram
What are mean and sd of sampling distribution for mean?
mean at mu, st dev at sigma over root n
What is a 2 sample T interval
When you have two samples and want to know if the populations they come from have different means. Stand at your diff and reach out and try to catch true diff
Where does the st error for 2 sample t stuff come from?
the added variances
do alpha and beta work with means?
YEP all the same…
do you pool with means? (t tests)
NOPE
how are 2 samp t and paired t tests different?
2 samp T you are looking at difference between 2 averages from two samples.. Paired t tests look at LIST 3, Just a single list of differences so you do a one sample test.
3 questions to decide right test?
One samp or Two samp.. means or %,… test or INT
how do you find degrees of freedom?
One sample and paired: n-1. Two sample: calculator regression: n-2 proportions: NO DF!!
How do you calculate sample size with means?
first use Z crit, calculate n, use that for t crit, then calculate again. N= (t*s/ME)2
why would someone pool with means?
when variances are equal it could increase sample size for more precise conclusions
what happens to t models as n gets larger
they look more like normal model (t looks like z)
what is a t crit?
same as z crit.. how many sd you reach out for confidence. For 95% CI, reach out about 2, for 99.7% CI reach out about 3
what is normal enough condition?
when you think sample could have come from normalish population
1 prop t test?
not possible. T for MEANS, Z for PROPS
when is test a PAIRED test?
2 measurements on same subject. Pre-Post test. OR..
who invented the “student’s T table?”
wild BILL GOSSET
what are the x bar and mu in the tests?
x bar is sample mean, mu is population or hypothesized mean
when you are doing paired, matched or blocked.. what are you finding
the average difference (not the difference of two averages)
WHat is difference between 2 sample t test and paired t test?
in a 2 sample test, you are looking for the difference between two means.. In a paired t test you are looking at the mean of a bunch of differences…