14 - Sampling & Surveys

Question 1

how to estimate unknown population mean µ

Answer 1

use sample mean xbar

Question 2

N vs. n

Answer 2

N = population size

n = sample size

Question 3

Good samples

Answer 3

probability samples for which each unit in the pop has a known probability of being in the sample

• SRS = Simple Random Sample = equal probability sample = each unit has the same chance of being in the sample

Question 4

Sampling error

Answer 4

the error in the estimate due to the fact that we did not see the entire population (did not take a census)

• can be driven to 0 by taking a bigger and bigger sample

Question 5

Sampling variability

Answer 5

different samples would give slightly different estimates

Question 6

Non-sampling error

Answer 6

• do not get smaller as the sample size gets larger
• attributable to a systematic lack of representativeness in the sampling mechanism
• best dealt with by avoiding them

Question 7

Sample size

Answer 7

So long as the sample size is small w/respect to the population size, then the sample size required to reach a certain level of precision is independent of the population size

Ex. n/N <= 0.1

Ex. do you need to take a bigger sample in CA than in WY bc CA has more people?

-no - take the same size sample in each state

Question 8

Stratified sample

Answer 8

group the population into homogeneous strata, then take an SRS w/in each stratum

• helpful when a few observations may dominate the population total

Question 9

Cluster samples

Answer 9

good when the pop naturally falls into groups (Ex. houses on a street, children in classrooms)

• can cut time & costs

Question 10

Sequential/Skip sampling

Answer 10

sample every i^th observation

• helpful when data is flowing in a stream (Ex. visitors on a website, containers at a packaging plant)