BUAL 2650 Auburn Exam 1 |51
Complete Q’s and A’s
sample proportion - -sample of the population, p-hat, that we use because
we do not know the parameter of the whole population, p. p=p-hat most of
the time but not always
- standard deviation - -typical difference between p and p-hat. the
proportion from sample, p-hat, is not equal to p, typically the estimate p-hat
will be off by the sq.rt of pq/n,
- confidence interval - -assume symmetry, p-hat +/- 2*SD(p-hat) for 95%
confidence interval, so 95/100 will contain p.
- conditions to check - -randomization condition, 10% condition (no larger
than 10% of the population), success/failure (nq >10, np>10)
- confidence intervals for proportions - -68%- (p-sq.rt.pq/n,p+sq.rt.pq/n)
95%- (p-2sq.rt.pq/n,p+2sq.rt.pq/n)
99.7%- (p-3sq.rt.pq/n,p+3sq.rt.pq/n)
- z-score - -p-hat - p / SD(p-hat)
mu(0,1) standard normal distribution
- positive z-score - -outlier > 3 is unusual
- negative z-score - -outlier < -3 is unusual
- null hypothesis - -we assume someone is innocent until proven guilty,
retain the hypothesis until the facts make it unlikely beyond a reasonable
doubt, consider if the data is consistent with the hypothesis
- stat hypothesis testing - -the population perimeter is the initial hypothesis,
p=x, collect data to challenge the hypothesis and form p-hat, then decide if
the data proves likely or unlikely
- Ho - -null hypothesis, population parameter, hypothesized value
- Ha - -alternative hypothesis, the parameter we deem plausible when we
reject the null hypothesis
- Two-sided test - -population parameter does not equal hypothesized value
- One-sided test - -population paramater > or < hypothesized value
Complete Q’s and A’s
sample proportion - -sample of the population, p-hat, that we use because
we do not know the parameter of the whole population, p. p=p-hat most of
the time but not always
- standard deviation - -typical difference between p and p-hat. the
proportion from sample, p-hat, is not equal to p, typically the estimate p-hat
will be off by the sq.rt of pq/n,
- confidence interval - -assume symmetry, p-hat +/- 2*SD(p-hat) for 95%
confidence interval, so 95/100 will contain p.
- conditions to check - -randomization condition, 10% condition (no larger
than 10% of the population), success/failure (nq >10, np>10)
- confidence intervals for proportions - -68%- (p-sq.rt.pq/n,p+sq.rt.pq/n)
95%- (p-2sq.rt.pq/n,p+2sq.rt.pq/n)
99.7%- (p-3sq.rt.pq/n,p+3sq.rt.pq/n)
- z-score - -p-hat - p / SD(p-hat)
mu(0,1) standard normal distribution
- positive z-score - -outlier > 3 is unusual
- negative z-score - -outlier < -3 is unusual
- null hypothesis - -we assume someone is innocent until proven guilty,
retain the hypothesis until the facts make it unlikely beyond a reasonable
doubt, consider if the data is consistent with the hypothesis
- stat hypothesis testing - -the population perimeter is the initial hypothesis,
p=x, collect data to challenge the hypothesis and form p-hat, then decide if
the data proves likely or unlikely
- Ho - -null hypothesis, population parameter, hypothesized value
- Ha - -alternative hypothesis, the parameter we deem plausible when we
reject the null hypothesis
- Two-sided test - -population parameter does not equal hypothesized value
- One-sided test - -population paramater > or < hypothesized value
