One mark is awarded to a correct answer for a multiple choice question; zero mark, otherwise. Denote the probability that a student knows the correct answer to the question by p. A test consists of n such multiple choice questions. Assume that the questions are independent and equally difficult to the student. How large does n need to be so that the length of a 95% confidence interval for p is smaller than 10%, if we believe p(1 − p) < 0.2? [Hint: consider only the situations with large n.]
do you recall the formula for an error?
the length of 95% is smaller than 10% ? thats hard to parse
in general. the setup for a confidence interval is: \[\bar x\pm Z_\%\sqrt{\frac{p(1-p)}{n}}\] or written another way \[\bar x\pm E\] so:\[E=Z_\%\sqrt{\frac{p(1-p)}{n}}\] i spose they want E to be .10 by that other stuff, solve for n
or would it be .10/2 ?
95% confidence interval mean something else 10% meaning that for example i have X and in this X +-10/2 is in interval where 95% of data is placed.
I guess 10 = +-5% on each side of the mean
yeah, that does seem like a better read :)
and how to understand that p(1 − p) < 0.2? as far as i know it is variance but why it is <0.2? and how it works than?
do we put 0.2 in and that it?
just leave that part as an unknown for now and work out the n, then we can use that to find a range with
ok
I got that\[n=p(1-p)/0.00065077\]
since they ask for 95% CI than z=1.96 is it right?
(1.96/.05)^2 = 1536.64, times p(1-p) \[n = 1536.64p(1-p)\] so n = 0 to 1536.64(.02)
yeah, 0 to 31 is what i get
i see you did (E/z)^2 for a denominator which is fine too
ok than answer in from 0 to 307.4?
i used .02 instead of .2 didnt i ... yeah, that fine. id rnd up to 308
ok, thank you i think a get the point.
Join our real-time social learning platform and learn together with your friends!