Question

5. A manufacturer of PCs uses a chip supplier for which the proportion of defective chips is an unknown parameter p. The PC manufacturer wants to estimate p by sampling n chips and determining the fraction Mn of sampled chips that are defective. The manufacturer wants to be 90% confident that the estimate is within 0.05 of the true value p. (a) If the manufacturer bases its analysis on Chebyshev's Inequality, how many chips would it need to sample? (b) If the manufacturer bases its analysis on the Central Limit Theorem, how many chips would it need to sample? (Table of standard normal cdf is attached)

Answer 1

Hese M, derote the Asocfion of chips that are defectlve out of n sampled chips og p{} Mn-P) so.cs)? }-.eaaつ o.05)2 we houst have Co.os)n C0.05)% PCH- Co. 05)2n vn 「우 go. = 1000 shoulde bte ctleast lood Sampes to be ao % Gnadent

どBy aamsal limit theorem, coe have PCi-P PCI-P) evn y/70の 6.。 2.05), By the cLT, the manu-factweea shauld toke aleast -Sawply to be 90% Ca, fident that the egガmate, g crit ?r) 0.05 of the ue value of p. CS Ca