Lecture 10 Flashcards
what is pvalue
it is the conditional probability of the sample effect size being observed or one larger, given that the null hypothesis is true
pvalue = pr ( t(obs) | H0 true )
what is a type 1 error
falsely rejecting the null hypothesis
means that null hypothesis is true (pvalue > alpha) but we (falsely) reject it.
if null hypothesis is false, can we commit to type 1 error?
no.
type 1 error is when H0 is TRUE but we falsely reject it.
we are supposed to reject if H0 is false… which is a CORRECT thing to do.
what is alpha value
the (conditional) probability of rejecting the null hypothesis H0 given the null hypo is in fact true
alpha = pr (reject H0 | H0 true)
what is priori (alpha value) for
indicative to the maximum chance of FALSELY REJECTING a true H0
how can pvalue be thought as a measure of the consistency of our sample with the null hypothesis population parameter?
the smaller the pvale, the less consistent is the data with the null hypothesis
“p value tells us null hypothesis is true or false” true or false?
false.
truth and falsity of H0 is not determined by pvalue.
it is pvalue = pr ( t(obs) | H0 true )
not pvalue = pr ( H0 true | t(obs) )
“p value tells us observed effect is due to chance alone. “ true or false?
p value reflect both assumed null hypothesis and all statistical assumptions, it CANNOT tell us the probability that the observed effect is due to chance alone.
p value is calculated on the assumption that chance alone is operating but does not indicate the probability of chance being the ONLY explanation
“p > alpha means NO effect is present” true or false?
no effect present means H0 is CORRECT and we fail to reject H0 (because pvalue was larger than alpha value)
this is FALSE.
you just FAIL to reject the statement that there is effect. It’s important to note that the null hypothesis is never accepted; we can only reject or fail to reject it.
how does confidence interval relate to pvalue?
lower bound in ci tells us that any null hypothesis value lower than this corresponding pvalue would result in rejection of h0
upper bound in ci tells us that any null hypothesis value greater than this corresponding pvalue would result in rejection of h0
“95% confidence interval tells us that 95% of the time our data would capture the population parameter value” true or false
true
“95% confidence interval tells us that there is 95% probability that our data would capture the population parameter” true or false
false
