and we’re back! with BRAND SPANKIN' NEW tangents!
welcome redditors!to snoo-finity ... and beyond!

nbme21/Block 4/Question#11

A study is conducted on patients with asthma to ...

If the treatment really changes the mean asthma score by 0.4, there is an 80% chance that a study of this size will find a p-value < 0.05

+5  upvote downvote
submitted by nosancuck(30),

Black Powah!

Powah = 1 - BETA = da rate of dem TRUE POSITIVES; wuz da rate at which you get da true positives across a buncha studies? 80% here bruh

lostweightthxnorovirus  wherever you learned biostats I wanna learn ;) +6

+2  upvote downvote
submitted by imgdoc(18),

Basically this question just asks "what is power" and asks you to explain that 80% power. Power is 1 - beta (type 2 error), basically when a difference exists and the null hypothesis is rejected is power. So if the Rx detects a mean difference of 0.4 in asthma in the patients in the treatment group, then that data falls in the 80% power range, and its significance is 95% (p<0.05), P value is just the probability of something happening by chance, so you what this to be less that 5% so whatever you observed isn't bullshit.

I hope this helped, and correct me if I'm wrong.

yssya1992  I have a question : whats the relation then between power and P value ? +
privatejoker  This one took me a minute and was during the last block so my brain was already fried. But my reasoning was that, as stated above, since it gave you power, it is basically just a long-winded way of asking what Power is, and how this relates to p-value. P-value is the odds that the finding was due to chance alone. Obviously a p-value set to <0.05 implies a greater than 95% chance that the finding is legit. Since the power is said to be 80%, this means that there is an 80% chance that the study finding is legit, at least insomuch that it met the pre-set criteria of being 95% non-chance related. +
sahusema  80% chance (power) the study correctly identifies the existence of an association in reality. If an association is determined to exist, >95% chance the study and reality agree with each other (p<.05) +

+0  upvote downvote
submitted by sahusema(12),

80% chance (power) the study correctly identifies the existence of an association in reality. If an association is determined to exist, >95% chance the study and reality agree with each other (p<.05)