Sampling Techniques

1
Sampling Techniques
2

• Governments, companies, and news agencies often
  want to know the publics opinion on pertinent
  questions.
• Elections offer an excellent example of sampling
  and bias.

3
Suppose you want to know who is going to win the
next election?
4

• Clearly it is not feasible to ask every person in
  the country directly.
• You can probably get an idea of the results by
  asking only a certain number of people
• The question is, how many?

5
A marketing research firm (Ipsos-Reid or Ekos or
Decima) would be hired by a news agency (CBC) to
poll the publicRecord the final results of our
last federal election by clicking belowExamine
the following
6
Check the following website to see how the polls
were able to track and predict the resultsThe
dates of each collection are on the x axis
results
7

• A private company must be efficient to stay in
  business.

8

• If a company asks too many people,
• they are wasting time and money
• If a company asks too few people,
• the results will not be valid.
• Determining the right number of respondents is a
  major challenge to these companies

9
Canadas population is about 32.5 million

• There are about 22.5 million registered voters
• Approximately 60 of the registered voters
  actually vote
• About 13.5 million people vote

10
Canadas population is about 32.5 million

• There are about 22.5 millions registered voters
• Approximately 60 of the population actually
  votes
• About 13.5 million people vote
• SES polls tracks 1200 voters
• 0.0089 of the population !!!!!!

11
Population

• All individuals in the group being studied

Sample

• A subset of the population

12
To see some examples of samples taken from
populations, check out the website below
samples
13
There are a number of different ways populations
can be sampled.
14
Simple Random Sample

• All selections must be independent of one another
  and equally likely
• Use a random number generator, dice, or a hat
  draw to ensure the data is randomly sampled.

15
Systematic Random Sample

• Used when you are sampling a fixed percent of the
  population.
• A random starting point is chosen, and then you
  select every nth individual, where n is the
  sampling interval.

16
For example

• You want to determine the height of 25 of the
  students in this class. (9 out of 36)

36
4
9
The sampling interval would be 4
17

• Randomly select the first person to measure (from
  1 to 4), then measure every 4th person after them.

18
Stratified Random Sampling

• The population is divided into different groups
  called strata (ex. geographic areas, gender,age).
• A simple random sample of the members in each
  stratum is taken.
• The size of the sample is proportional to the
  stratums size. (a consistent percent)

19
Other sampling techniques

• Make a note of the sampling techniques discussed
  on page 116 in the text.

20
Sampling Summary Chart
21
Simple Random Sample Every member of the population has an equal and independent chance of being selected
Systematic Sample Select the members at regular intervals starting from a random spot
Stratified Sample Divide the population into strata that have something in common (age, province). Select a SRS from each strata
22
Cluster Sample Certain groups can be sampled if they represent the entire population. All the employees at a single McDonalds.
Multi-Stage Sample Two or more SRSs. Cities, then subdivisions, then houses.
Voluntary Response Collect data on a voluntary basis. ie call in show or mail in survey
23
Convenience Sample The sample is selected because it is easily accessible. Not as random as other techniques.
24

• Page 117
• 1,2,4,8,9
• Plus examples on pg 116