zc = 1.44
Compare
results: When you increase the amount
of acceptable error (E), the sample size required is smaller.
(b)
If the Normal distribution is used, zc = 1.96, and the 95%
confidence interval for m is (12.811, 13.989). So, the t
distribution gives a wider interval.
Since
1000 hours falls within the confidence interval, it can be concluded that they
are making good lightbulbs.
The 99% confidence interval is wider.
(a)
(b)
(c) If you have an estimate for the proportion (that is not 0.5), the sample size needed will always be smaller. Information about the proportion helps to reduce sample size.
(.00069,.0022) (.0247,.04690)
Not really acceptable, since the criteria calls for s < .025 and the lower limit of the CI ~
.025 (rounding up .0247), meaning within 90% confidence limits the std deviation
is too high.