STATS: 2 samp Flashcards
when every # in the interval is larger than 0 that means…
there was change (no change would be 0)
Interval –>
PANIC
Test –>
PHANTOMS
POI statement (for BOTH):
The POI is the true DIFFERENCE in (means/proportions) between group 1 and group 2.
Assumptions for proportions
-random sample (for both groups)
-10% condition (for both groups)
-n1(p1) n1(q1) ~ n2(p2) n2(q2) ~ greater than equal to 10 (for approx. normal)
use p-hat for proportion-tests
2-Prop Z Int
define z* and interval
Conclusion for 2 sample INTERVALS
We are __% confident that the true DIFFERENCE in (proportions/means) between group 1 and group 2 is between __ and __.
Assumptions for means
-random sample (for both groups)
-10% condition (for both groups)
-for approx normal:
~both populations are normal
~n greater than or equal to 30 (CLT)
~graph: no strong skew or potential outliers for both data sets
2-Samp T Int
define t* and interval
2-Samp T Test
define t-value and p-value
2-Prop Z Test
define z-value and p-value
ALWAYS STATE YOUR …
SIGNIFICANCE LEVEL (for tests)
p-value < significance level
reject H0, stat. sig –> we DO have convincing evidence
p-value not< significance level
fail to reject H0, not stat. sig –> we DO NOT have convincing evidence
Conclusion for 2-sample tests
We do/do not have convincing evidence of a different (average/proportion) of ____ between group 1 and group 2
Hypotheses for proportions
Ho: p1=p2
Ha: p1<,>,not=to p2
Hypotheses for means
Ho: M1=M2
Ha: M1<,>,not=to M2
Hypotheses are used for …
tests !! (PHANTOMS)
write down the memorization phrases:
605-ZAp-hat
42-CLT8x
when looking for a matched pair …
-before and after (of ONE group)
-when subjects are paired then assigned different treatments
-if each subject got both treatments and your comparing effects of treatments
when looking for 2 sample …
-completely randomized design
-when comparing 2 SEPARATE groups
(look to find 2 DIFFERENT groups in a question)
when looking for the amount of time…
MEANS
when someone does something PER DAY …
MEANS
key words for intervals
-estimate(difference)
-find
-describe
key words for tests
-when needing to answer a question
-to find enough evidence
-comparing
-suggest
don’t forget to …
assign which group is group 1 and group 2!!
A double blind experiment …
reduces bias (hippocratic oath)
for matched pairs …
-POI is the true AVERAGE DIFFERENCE between…
-Ho: M=0
Ha: M<,>,not=to 0
-matched pair t-test for M