sig test Flashcards
Ho
- made for reject/not reject
- NEVER USE CE FOR HO
- “always” =
- null hypothesis
Ha
- made for CE or not Ce
- NEVER REJECT HA
- > , <, and unequal
- alternative hypothesis
Conclusion p value ( p>0.05)
Because the p value is 0.07 which is greater than or equal to alpha (unless stated, is 0.05), we do NOT reject Ho and do NOT conclude Ha . The data do not provide convincing evidence that the true prop of _____ is different from 0.72 (percentage provided in question)
** no on average when making conclusions***
Conclusion p value ( p<0.05)
Because the value 0.015 is less than or equal to alpha (0.05 unless provided), we reject Ho and conclude Ha. The data do provide convincing evidence that the true mean weight of _____ is more than/less than (depending on question) 31.4 oz
Interpret p value (0.08)
If the true prop/mean of ___ is actually 0.72, then there is a 0.08 probability of getting a sample mean/proportion of (sample mean or prop ex. 90/150) BY CHANCE !!!!!!!!!!!!!!!!!!!!!!!!!in an srs of 150!!!!!!!!!!!!!!!
If you reject Ho, you can conclude Ha
If you do not reject Ho, you do not conclude Ha
PLAN/SETUP:
Ho: =
Ha: >, <, not =
Parameters
P1-P2
Ud
U
P
Interpret p as 0.02
If I truly have 10 dogs, There is a 0.02 probability that I will get a sample mean/true proportion of 94 (use number) by chance in a srs of 25
Conclude:
Because p= 0.02 is less than or equal to alpha = 0.05, we reject Ho conclude Ha. The data do provide convincing evidence that the true prop/the sample mean of ____ is greater than/less than 10 (if ho= 10)
(no number, only greater than/less than)
interpret- use second number (Not HO ) because it’s that probability that you can get that number
conclude- do/don’t reject ho, find/don’t find ce that the sample mean or true prop is greater/less than/not equal to the first number
Type 1 error
FALSE POSITIVE- you incorrectly reject Ho and the data do provide convincing evidence that you have covid but in reality you don’t
Type 2 error
FALSE NEGATIVE- you do not reject Ho and do not find convincing evidence that you have covid but in reality you do have covid
type 1 consequence
you do reject ho and find convincing evidence (no mention of ha) that ce that that this place is affordable but this place is actually not affordable.
—A consequence can be that I move to this place incorrectly thinking this place is affordable and I waste money and possibly go broke :(
type 2 consequence
i incorrectly do not reject ho and (no mention of ha) don’t find convincing evidence that that this place is affordable but this place is actually affordable.
—A consequence can be that I do not move to this place incorrectly thinking this place is not affordable and I MISS OUT on possibly saving money :(
Alpha and beta have
an INVERSE RELATIONSHIP!!
- when alpha increases, beta decreases
- when beta increases, alpha decreases
Alpha and beta do NOT equal 1
Power = 1-Beta
Power is the proportion of
correctly rejecting Ho when Ho is actually FALSE (Ha is true)
ex. testing positive for covid and you have it
Power is the probability of
NOT making a type 2 error (false negative)
ex. testing negative for covid but you have it
Alpha
P(Type 1 error)
Beta
P(Type 2 error)
How to increase power
- increase sample size!!!!
- increase alpha
- decrease beta
- if the true parameter/sample data is further away from the Ha’s direction, the more power you have
ex. if Ha (!!!) does not equal 10, 3 is a better number than 9,
more power= smaller p value
the further away the p value is from from Ha, the less Ho which is more CE
we want the sample mean/true parameter furthest away from Ha in Ha’s direction
