8.2 Estimating a Population Proportion Answering these types of questions: What proportion of U.S. adults are unemployed right now? What proportion of high school students have cheated on a test? What proportion of Ash trees have the Emerald Ash Beetle? What proportion of Cicadas actually survive the cycle of 7 years? Activity: The Beads
1
8.2 Estimating a Population Proportion Conditions for Estimating p Shape: If the sample size is large enough that both np and n(1p) are at least 10 (Normal Condition) the sampling distribution of phat is approx. Normal. Center: The mean is p. That is the sample proportion phat is an unbiased estimator of the population proportion p. Spread: The standard deviation of the sampling distribution of phat is σphat = √p(1p)/n provided that the population is at least 10 times as large as the sample (10% condition)
2
Example: The Beads p486 nphat = 251(107/251) = 107 and n(1phat) = 251(1(107/251)) = 251(144/251) = 144 so both are at least 10
3
Constructing a Confidence Interval for p Recall: We can construct a confidence interval for an unknown population proportion p: statistic ± (critical value) x ( standard deviation of statistic) where phat is the statistic we use to estimate p When Independent condition is met, the standard deviation of the sampling distribution of phat is σp = √[(p)(1p)/n] but since we do not know the value of p we put phat in its place σp = √[(p)(1p)/n]
This quantity is called the standard error (SE) of the sample proportion phat. Definition: Standard Error When the standard deviation of a statistic is estimated from data, the result is called the standard error of the statistic. Calculating the Critical Value z* Define alpha as the % left out of the confidence interval i.e. if 95% CI then alpha = .05 z* = zalpha/2 = z.025 then refer to table
4
Example: p488 80% Confidence
z
.07
.08
.09
1.3 1.2 1.1
.0853 .1020 .1210
.0838 .1003 .1190
.0823 .0985 .1170
You can also find the critical value using the command invNorm(.9, 0, 1) which tells the calulator to find the zvalue from the standard Normal curve that has area .9 to the left of it.
5
Example: In a study of 400 adults, the credit reports of 79% of these adults contained errors. Find the 95% CI for the proportion of all adults in the US whose credit report contains errors. n = 400
phat = .79
alpha = .05 so alpha/s = .025
phat ± z*√p(1p)/n = .79 ± z.025√(.79)(.21)/400 = .79 ±1.96(.020365) = .79 ± .0399 (.7501, .8299) so with 95% confidence the proportion of all adults whose credit reports contain errors is between . 7501 and .8299.
6
OneSample z Interval for a Population Proportion Choose an SRS of size n from a large population that contains an unknown proportion p of successes. An approximate level C confidence interval for p is p ± z*√(p)(1p)/n where z* is the critical value for the standard Normal curve with area C between z* and z*. Use this interval only when the numbers of successes and failures in the sample are both at least 10 and the population is at least 10 times as large as the sample.
7
Example: The Beads p489
z 1.7 1.6 1.5
.03 .0418 .0516 .0630
.04 .0409 .0505 .0618
.05 .0401 .0495 .0606
statistic ± (critical value)(st. deviation of statistic) = phat ±z*√phat(1phat)/n
8
PUTTING IT ALL TOGETHER: THE FOURSTEP PROCESS 1. STATE: what parameter do you want to estimate, and at what confidence level? 2. PLAN: Identify the appropriate inference method. Check conditions. 3. DO: If the conditions are met, perform the calculations. 4. CONCLUDE: Interpret your interval in the context of the problem.
9
Example: Teens Say Sex Can Wait p491 nphat = 246≥10 and n(1phat) = 193≥10
phat = z*√phat(1phat)/n = .56 ± 1.96√(.56)(.44)/439 .56 ± .046 = (.514, .606)
10
AP Exam Tip: If a freeresponse question asks you to construct and interpret a confidence interval, you are expected to do the entire fourstep process. That includes clearly defining the parameter and checking conditions. Technology Corner p492
11
Example: A survey of 900 US Adults, 48% of Americans say there is never enough time in day to get things done. Find with a 90% CI for the proportion fo Americans who say this. n = 900
phat = .48
alpha = .1 alpha/2 = .05
phat ±z.05√(p)(1p)/n = .48 ± 1.645√(.48)(.52)/900 = .48 ± .0274 so (.4526, .5074)
12
CHOOSING THE SAMPLE SIZE Sample Size for Desired Margin of Error To determine the sample size n that will yield a level C confidence interval for a population proportion p with a maximum margin of error ME, solve the following inequality for n: z*√[p(1p)/n] ≤ ME where phat is a guessed value for the sample proportion. The margin of error will always be less than or equal to ME if you take the guess phat to be .5 It may be easier to go ahead and solve for n in the above equation b/c that is what you are looking for: n = [(z*)2p(1p)]/(ME)2 ME is the desired margin of error (what you want to add and subtract to phat in the confidence interval) If phat is not known, typically we set phat = .5 which yields the most conservative estimate of n (i.e. the largest possible sample size)
13
Example: Customer Satisfaction p493
14
Example: A large city is considering legalizing casino gambling. The mayor would like to estimate p, the proportion of all adults who support legalizing casino gambling. What would be the most conservative estimate of n, the sample size that would limit the margin of error to be .05 for a 95% CI of p? phat = .5
ME = .05
alpha = .05
alpha/2 = .025
n = (z.025)2(phat)(1phat)/(ME)2 = (1.96)2(.5)(.5)/(.05)2 = 384.16 people Always Round Up = 385 people
15
Example: An electronics company has just installed a new machine that makes a part used in clocks. The company wants to estimate the proportion of these parts defective, p. The manager wants to be within .02 of p for a 99% CI. What is the most conservative estimate of the sample size, n , needed to achieve this CI? phat = .5
E = .02
alpha = .01
alpha/2 = .005
n = (z.005)2(phat)(1phat)/(E)2 = (2.576)2(.5)(.5)/(.02)2 = 4147.36 = 4148 parts
16
Ex: In a poll of 2000 adults, 1280 have money in regular savings accounts. Find with a 95% CI for proportion of adults with savings. n = 2000 phat = 1280/2000 = .64 alpha = .05
alpha/2 = .025
phat ± z.025√(.64)(.36)/2000 = .64 ± (1.96)(.010733) = .64 ± .0210 ( .6190, .6610)
17
Ex: You want to estimate with maximum error of .03 the true proportion (p) of all TV households tuned to a particular show and you want 90% CI . What is the most conservative number of households to survey? phat = .5
E = .03
alpha = .1
alpha/2 = .05
n = (z.05)2(phat)(1phat)/E2 = (1.645)2(.5)(.5)/.032 = 751.673 = 752 households
18
Homework: p 497 35, 37, 41, 43, 47
19