A sample mean is 42, its standard error is 3, and μ is 36. What is its z-score?
A 1
B 3
C 6
D 2
Answer: D. 2. The standardized distance is (42−36)/3 = 2.
For σ=18 and n=36, what is the sample mean’s standard error?
A 2
B 3
C 6
D 18
Answer: B. 3. The standard error is 18/√36 = 18/6 = 3.
A statistic’s uncertainty is estimated by repeatedly resampling observed cases with replacement. Which method is used?
A Bootstrap
B Census
C Systematic sampling
D Post-stratification
Answer: A. Bootstrap. The bootstrap approximates a statistic’s sampling distribution by resampling observed data with replacement.
Two strata hold 75% and 25% of a population, with means 8 and 18. What is the combined mean?
A 8.5
B 11
C 10.5
D 13
Answer: C. 10.5. Weighting gives 0.75×8 + 0.25×18 = 10.5.
A sample of 240 uses proportional allocation. What is n for a 25% stratum?
A 60
B 24
C 48
D 80
Answer: A. 60. Proportional allocation assigns 25% of 240, which is 60.
With equal costs, Neyman allocation favors which strata?
A Smaller, less variable strata
B Only the cheapest stratum
C Only the largest stratum
D Larger, more variable strata
Answer: D. Larger, more variable strata. For equal costs, optimal allocation assigns more observations to larger strata with greater variability.
A population list has 2,100 units and n=70. What systematic interval is used?
A 7 units
B 21 units
C 30 units
D 70 units
Answer: C. 30 units. The systematic interval is N/n = 2,100/70 = 30.
Which step gives systematic sampling a random selection basis?
A Fixed interval
B Random start
C Sorted outcomes
D Large sample
Answer: B. Random start. A random start begins selection; the fixed interval then determines subsequent units.
A survey selects 12 colleges and studies all students in them. What are colleges?
A Response categories
B Sampling errors
C Population parameters
D Primary sampling units
Answer: D. Primary sampling units. The first-stage selected clusters are primary sampling units.
In a two-stage design, what is an individual’s inclusion probability?
A Product of stage probabilities
B Sum of stage probabilities
C Largest stage probability
D Inverse sample size
Answer: A. Product of stage probabilities. The chance of selection through successive stages is the product of conditional stage probabilities.
Under PPS selection, which units usually get larger weights?
A Largest-probability units
B Earlier-listed units
C Lower-probability units
D Nonresponding units
Answer: C. Lower-probability units. Inverse-probability weights are larger for units with lower selection probabilities.
A unit has selection probability 0.10. What is its base weight?
A 0.1
B 10
C 1
D 100
Answer: B. 10. The base weight is the inverse probability: 1/0.10 = 10.
How are weights commonly adjusted for nonresponse?
A Set all weights to zero
B Drop the target group
C Reverse the strata
D Inflate respondent weights
Answer: D. Inflate respondent weights. Nonresponse adjustments increase respondent weights to represent similar sampled nonrespondents.
The standard error of x-bar is what when population SD is σ?
A σ divided by n
B n divided by σ
C σ divided by √n
D σ times n
Answer: C. σ divided by √n. For a random sample, the standard error of the mean is σ/√n, before any finite-population correction.
For p=0.25 and n=100, what is the approximate proportion standard error?
A 0.043
B 0.025
C 0.075
D 0.188
Answer: A. 0.043. SE = √[0.25×0.75/100] ≈ 0.0433.
When is a finite-population correction relevant?
A Small sample with replacement
B Large fraction sampled without replacement
C Convenience sampling only
D Every census
Answer: B. Large fraction sampled without replacement. It accounts for reduced variability when a sizable fraction of a finite population is sampled without replacement.
If σ=20, n=100, and z=1.96, what is the mean margin of error?
A 1.96
B 2
C 3.92
D 19.6
Answer: C. 3.92. The margin is zσ/√n = 1.96×20/10 = 3.92.
If target margin is halved, how should n change approximately?
A Double it
B Keep it fixed
C Reduce it by half
D Increase fourfold
Answer: D. Increase fourfold. For fixed confidence and variability, sample size is proportional to 1/(margin)^2.
With a large sample and known σ, which mean interval uses z?
A x-bar ± σn
B x-bar ± zσ/√n
C μ ± z/n
D x-bar ± nσ
Answer: B. x-bar ± zσ/√n. A z-based mean interval uses the sample mean plus or minus z times its standard error.
For repeated random samples, what centers the p-hat distribution?
A Population proportion
B Population mean
C Sample size
D Standard error
Answer: A. Population proportion. The expected sample proportion equals the population proportion under probability sampling.
For a skewed population, what helps the sample-mean distribution become approximately normal?
A Low response rate
B Fixed start
C Large random sample
D Complete census
Answer: C. Large random sample. The central limit theorem says the sample-mean distribution approaches normality as sample size becomes sufficiently large, under its conditions.
Two survey frames contain many of the same households. What issue must be addressed?
A Random start
B Frame overlap
C Item response
D Stratum variance
Answer: B. Frame overlap. Overlapping frames can duplicate units or alter selection probabilities unless overlap is handled.
A complex design’s variance is twice the SRS variance at equal n. What is its design effect?
A 2
B 0.5
C 1
D 4
Answer: A. 2. Design effect is the variance ratio; twice the SRS variance gives a design effect of 2.
A sampled person gives no interview, while another skips income. Which pairing is correct?
A Item, then unit nonresponse
B Coverage, then measurement
C Sampling, then processing
D Unit, then item nonresponse
Answer: D. Unit, then item nonresponse. No interview is unit nonresponse; a missing answer to one question is item nonresponse.
A list excludes a remote village in every survey wave. What issue is most likely?
A Random sampling error
B Item nonresponse
C Stratified allocation
D Persistent undercoverage
Answer: D. Persistent undercoverage. Repeated exclusion of part of the target population is systematic undercoverage.