Sampling

1
Sampling
2
Why Sample?

• Time, cost
• Accuracy representativeness
• time-sensitive issues

3
What is a sample? Key Ideas Basic Terminology

• Sampling Guide (general introduction) in Reading
  Folder
• Population, target population
• the universe of phenomena we want to study
• Can be people, things, practices
• Sampling Frame (conceptual operational issues)
• how can we locate the population we wish to
  study? Examples
• Residents of a city? Telephone book, voters lists
• Newsbroadcasts? Broadcast corporation archives?
• Telecommunications technologies?....
• Homeless teenagers?
• ethnic media providers in BC (print, broadcast)

4
Diagram of key ideas terms
5
Target Population

• Target Population--Conceptual definition
• the entire group about which the researcher
  wishes to draw conclusions.
• Example Suppose we want to study homeless men
  aged 35-40 who live in the downtown east side and
  are HIV positive.
• The purpose of this study could be to compare the
  effectiveness of two AIDs prevention campaigns,
  one that encourages the men to seek access to
  care at drop-in clinics and the other that
  involves distribution of information and supplies
  by community health workers at shelters and on
  the street.
• The target population here would be all men
  meeting the same general conditions as those
  actually included in the sample drawn for the
  study.
• What sampling frames could we use to draw our
  samples?

6
Bad sampling frame

• parameters do not accurately represent target
  population
• e.g., a list of people in the phone directory
  does not reflect all the people in a town because
  not everyone has a phone or is listed in the
  directory.

7
Recall Videoclip from Ask a Silly Question
(play videoclip)

• Ice Storm, electricity disruption, telephone
  survey
• Target Population Hydro company users
• Sampling frame unclear, probably phonebook or
  phone numbers of subscribers
• Problem people with no electricity not at home
  but in shelters
• Famous examples from the past Polls of voters
  before election (people with phones or car owners
  not representative of total voters, or opinions
  not yet formed)

8
More Basic Terminology

• Sampling element (recall unit of analysis)
• e.g., person, group, city block, news broadcast,
  advertisement, etc

9
Recall Units of Analysis (Individuals)
10
Recall Units of Analysis (Families)
11
( Households)
12
Recall Importance of Choosing Appropriate Unit
of Analysis for Research

• Recall example Ecological Fallacy (cheating)
• Unit of analysis here is a class of students.
  Classes with more males had more cheating

13
What happens if we compare number and gender of
cheaters? (unit of analysis students)

• Do males cheat more than females?
• Same absolute number of male and female cheaters
  in each class

14
Comparison of and of cheaters by gender
15
Recall Ecological Fallacy Reductionism
ecological fallacy--wrong unit of analysis
(too high) reductionism--wrong unit of
analysis (too low)
reductionism--wrong unit of analysis (too low)
16
More Basic Terminology

• Sampling ratio
• a proportion of a population
• e.g., 3 out of 100 people
• e.g., 3 of the universe

17
Factors Influencing Choice of Sampling Technique

• Speed
• Cost
• Accuracy
• Assumptions about distribution of characteristics
  of population
• link to stats Can site http//www.statcan.ca/engl
  ish/edu/power/ch13/non_probability/non_probability
  .htm
• Availability of means of access (sampling frame)
• Nature of research question(s) objectives

18
Some types of Non-probability Sampling

• 1. Haphazard, accidental, convenience(ex.
  Person on the street interview)
• 2. Quota (predetermined groups)
• 3. Purposive or Judgemental
• Deviant case (type of purposive sampling)
• 4. Snowball (network, chain, referral,
  reputation) volunteer
• Also--multi-stage sampling designs

19
Non-probability Sampling1. Haphazard,
accidental, convenience(ex. Person on the
street interview)
Babbie (1995 192)
20
Non-probability Sampling 2. Quota (predetermined
groups)
Neuman (2000 197)
21
Why have quotas?

• Ex. populations with unequal representation of
  groups under study
• Comparative studies of minority groups with
  majority or groups that are not equally
  represented in population
• Study of different experiences of hospital staff
  with technological change (nurses, nurses aids,
  doctors, pharmacistsdifferent sizes of staff,
  different numbers)

22
Non-probability Sampling 3. Purposive or
Judgemental

• Unique/singular/particular cases
• Hard-to-find groups
• Leaders (success stories)
• Range of different types

23
Non-probability Sampling 4. Snowball (network,
chain, referral, reputational)
Sociogram of Friendship Relations
Neuman (2000 199)
24
Issues in Non-probability sampling

• Bias?
• Is the sample representative?
• Types of sampling problems
• Alpha find a trend in the sample that does not
  exist in the population
• Beta do not find a trend in the sample that
  exists in the population

25
Types of Probability Sampling

• 1. Simple Random Sample
• 2. Systematic Sample
• 3. Stratified Sampling
• 4. Cluster Sampling
• See Statistics Canada site
• http//www.statcan.ca/english/edu/power/ch13/proba
  bility/probability.htm

26
Simple Random Sample

• With/without replacement?
• Must take into account characteristics of
  population sampling frame
• Develop a sampling frame Number sampling frame
  units
• Select elements using mathematically random
  procedure
• Table of random numbers
• random number generator
• Other statistical software
• Link How to use a table of random numbers

27
Principles of Probability Sampling

• each member of the population an equal chance of
  being chosen within specified parameters
• Advantages
• ideal for statistical purposes
• Disadvantages
• hard to achieve in practice
• requires an accurate list (sampling frame or
  operational definition) of the whole population
• expensive

28
How to Do a Simple Random Sample

• Develop sampling frame
• Locate and identify selected element
• Link to helpful website

29
2. Systematic Sample (every nth person) With
Random Start
Babbie (1995 211)
30
Problems with Systematic Sampling

• Biases or regularities in some types of
  sampling frames (ex. Property owners names of
  heterosexual couples listed with mans name
  first, etc)
• Urban studies example)

31
Other Types

• Stratified

Neuman (2000 209)
32
Stratified SamplingSampling Disproportionately
and Weighting
Babbie (1995 222)
33
Stratified Sampling

• Used when information is needed about subgroups
• Divide population into subgroups before using
  random sampling technique

34
Other Types

• Cluster
• When is it used?
• lack good sampling frame or cost too high

Singleton, et al (1993 156)
35
Other Sampling Techniques (contd)

• Probability Proportionate to Size (PPS)
• Random Digit Dialing

36
New Technologies Data Mining the Blogosphere

• Jan. 3, 2007 image with Boingboing as largest
  node (source http//datamining.typepad.com/data_
  mining/2007/01/the_blogosphere.html)

37
Sample Size?

• Statistical methods to estimate confidence
  intervals
• Past experience (rule of thumb)
• Smaller populations, larger sampling ratios
• Other factors
• goals of study
• number of variables and type of analysis
• features of populations
• In qualitative methods notion of Saturation
  (Bertaux)

38
Examples of sampling issues techniques

• Survey about football (soccer) market
• Rural poverty project and sampling issues

39
Issues/notions in Probability Sampling

• Assessing Equal chance of being chosen
• Standard deviation
• Sampling error
• Sampling distribution
• Central limit theorem
• Confidence intervals (margin of error)

40
Techniques for Assessing Probability Sampling

• Standard deviation
• Sampling error
• Sampling distribution
• Central limit theorem
• Confidence intervals (margin of error)

41
Inferences (Logic of Sampling)

• Use data collected about probabilistic samples to
  make statistical inferences about target
  population
• Note inferences made about the probability
  (likelihood) that the observations were or were
  not due to chance