Statistics: Population and Sampling

Population and Sampling

• Population: the entire group under study as defined by research objectives
• Census: requires information from everyone within a population instead of a sample
• Sample: a subset of the population that should represent the entire group
• Sample Unit: the basic level of investigation
• Sample Frame: a master source of sample units in the population
• Sampling Frame Error: the degree to which the sample frame fails to account for all of the population
• Sampling Error: any error in a survey that occurs because a sample was used

Reasons for Sampling

• Cost considerations
• Population size is too big to analyze the entire census
• Inability to analyze a whole census

Probability Samples

• Members of the population have a known chance (probability) of being in the sample
• Types:
  • Simple Random Sampling
  • Systematic Sampling
  • Cluster Sampling
  • Stratified Sampling

Probability Sampling Methods

Simple Random Sampling

• The probability of being selected into the sample is equal for all members
• Probability of selection = sample size / population size
• Methods:
  • Random Device Method
  • Blind draw
  • Random number
  • Random digit dialing (RDD)

Systematic Sampling

• An efficient method of selecting a random sample from a directory or list
• Uses a skip interval for selection (selects every other qualified member for the sample)

Cluster Sampling

• The population is divided into subgroups (clusters) that represent the entire population
• Area Sampling: divides geographic areas into clusters for sampling
• One-step area sample: various geographic areas are similar, so one of these could be representative for all
• Two-step area sample: researcher selects a random sample of areas and then decides on an appropriate probability method to sample individuals within the chosen areas

Stratified Sampling

• Separates the population into different subgroups (strata) and then samples them all
• Used with a skewed population and requires calculating a weighted mean
• Types:
  • Proportionate Stratified Sample: has sample sizes scaled to population size
  • Disproportionate Stratified Sample: sample sizes not scaled to population size, may be more statistically efficient

Nonprobability Sampling

• Not based on fairness, equity, or equal chance
• Methods:
  • Convenience Sampling
  • Chain Referral Sampling
  • Purposive Sampling
  • Quota Sampling

Nonprobability Sampling Methods

Convenience Sampling

• Samples drawn at the convenience of the interviewer (e.g., Mall interviews)
• Sample frame error occurs in the form of members of the population who are infrequent or nonusers

Chain Referral Sampling

• Requires respondents to provide the names of prospective respondents

Purposive Sampling

• Requires a judgment or an educated guess as to who should represent the population
• Subjectivity is introduced, so certain population members have less chance to be selected

Quota Sampling

• Specified percentages of the total sample for various types of individuals to be interviewed
• Uses demographics or product use factors; sizes of quota determined by researcher
• Ensures convenience samples include desired proportion of different respondent classes

Online Sampling Techniques

• Online Panels: large number of individuals who have agreed to participate in online surveys
• River Samples: invites website visitors to take a survey via banners, pop-ups, or other online devices
• E-mail List Samples: purchased/procured from a person/firm that has a list of emails of population opt-in members

Sample Plan Steps

• Define the population
• Obtain sample frame
• Decide sample method
• Decide sample size
• Draw the sample
• Validate the sample (inspect characteristics of respondents to see if it matches with population characteristics)