Sampling

1
Sampling
Partially Adapted from The Research Methods
Knowledge Base, William Trochim (2006). Methods
for Social Researchers in Developing Counries,
The Ahfad University for Women Online
Statistics Education A Multimedia Course of
Study, David M. Lane, Rice University.
2
Agenda

• Introduction
• Sampling Terminology
• Types of Sampling
• Sampling Biases

2
3
Introduction

• The purpose of research is to learn something
  about a larger group the population
• It is often impractical, however, to enumerate an
  entire population
• The alternative is to obtain a representative
  sample

3
4
Sampling Terminology

• Population abstract vs target
• Sampling element A single member or unit of
  target population
• Sample frame list of all sampling elements in
  target population
• Statistic finding based on a sample used to
  estimate a parameter
• Parameter finding based on measuring entire
  population
• Sampling Error/Sampling Bias how well a
  statistic estimates a parameter

4
5
Types of Sampling

• Probability Sampling relies on random selection
• Simple random sampling
• Stratified random sampling
• Systematic random sampling
• Cluster sampling
• Nonprobability Sampling no random selection
• Convenience Sampling
• Quota Sampling
• Expert Sampling
• Network Sampling

5
6
Types of Probability Sampling

• Simple random sampling
• Simplest form of random sampling
• Select n units out of N (target population) such
  that each unit has equal chance of being selected
• Stratified random sampling
• Population is divided into homogeneous subgroups
  and a random sample is taken from each subgroup
• Systematic random sampling
• Sampling interval is calculated and used to
  select sampling elements
• Cluster Sampling
• Population is divided into clusters
• Clusters are randomly sampled
• All units within clusters are measured

6
7
Types of Nonprobability Sampling

• Convenience Sampling
• Sample selected because of easy accessibility
• Quota Sampling
• A quota for subgroup is set in advance
• Persons with right characteristics are selected
  nonrandomly until number is met
• Expert Sampling
• Sample of persons with experience or expertise is
  assembled
• Network Sampling
• Begin by identifying someone who meets the
  criteria for inclusion in you study
• Next ask them to recommend others whom they know
  who meet the criteria. So on, so forth
• Useful when trying to reach inaccessible
  populations

7
8
Sampling Biases

• Sampling biases are errors that arise due to
  method of sampling used
• There are 3 common types of sampling biases
• Self-Selection Bias
• Survivorship (Attrition) Bias
• Sampling Selection Bias

8
9
Self-Selection Bias

• People who self-select themselves for an
  experiment might not be representative of the
  target population
• respondents who signup for an online survey about
  computer use
• When people sign-up for an experiment without
  knowing what it is about and leave subsequently
  upon finding out
• Non-response bias when certain subjects are more
  likely to respond than others

9
10
Survivorship (Attrition) Bias

• When observations recorded at the end of an
  investigation are a non-random set of those
  present at the beginning of the investigation
• Measuring gains in stock funds when poorly
  performing funds are often eliminated
• Measuring effects of school feeding programs when
  underweight students often drop-out of school

10
11
Sampling Selection Bias

• Arises through researchers choice of sampling
  method or units sampled
• Nonprobability sampling
• Under-coverage bias
• Researcher samples too few observations from
  certain segments of the population
• U.S Elections Landon-Roosevelt telephone poll
  Roosevelt won despite polls in favor of Landon ?
  Poorer people who tend to vote for Roosevelt did
  not have a telephone

11
12
Effects of Biases

• Resulting sample becomes less representative of
  target population
• Results of study not as generalizable to the
  target population

12