SAMPLING Census -- the entire population population (e.g., 10% or 25%)

advertisement
SAMPLING
• Census -- the entire population
– most useful is the population ("n") is small
– or the cost of making an error are high
• Sample -- contacting a portion of the
population (e.g., 10% or 25%)
– best with a very large population (n)
– easiest with a homogeneous population
7/17/2016
Marketing Research
1
Sampling -- graphically
Population
7/17/2016
Marketing Research
Sample
2
The sampling process
1. Determine the target population
who are the people you want information on?
age, gender, product use
2. Determine the sampling frame
how will you get the names, phone numbers or
addresses?
– existing lists, phone book, random digit dialing
7/17/2016
Marketing Research
3
The sampling process
3. Select a sampling procedure:
A. Probability (random) sample
– equal chance of being included in the sample
– random number table, even-odd, etc.
B. Stratified
– equalizing "important" variables
• year in school, geographic area, product use, etc.
7/17/2016
Marketing Research
4
The sampling process
C. Nonprobability sampling
– convenience sample
• people in my classes
– "snowball" sample
• friends of friends
– "quota" sampling
• 50 women, 50 men
– mall intercepts
• Park Place Mall
7/17/2016
Marketing Research
5
Generalization
• You can only generalize to the
population from which you sampled
– U of L students not college students
• geographic, different majors, different jobs, etc.
– College students not Canadian population
• younger, poorer, etc.
– Canadians not people everywhere
• less traditional, more affluent, etc.
7/17/2016
Marketing Research
6
Instead of this (good)…
Population
7/17/2016
Marketing Research
Sample
7
…this (bad)…
Sample
Population
7/17/2016
Marketing Research
8
…or this (VERY bad)…
Sample
Population
7/17/2016
Marketing Research
9
Drawing inferences from samples
• Population estimates
– % who smoke, buy your product, etc
• 25% of sample
• what % of population?
– very dangerous with a non-representative
sample or with low response rates
7/17/2016
Marketing Research
10
Drawing inferences from samples
• Relationships
– e.g., exposure to ads and liking for the product
– relationships (qualitatively different)
– some danger with a non-representative sample
• less problematic in experiments
– (other factors are controlled)
7/17/2016
Marketing Research
11
Sampling messages
• To be able to draw inferences about
a sample of messages
– (e.g., "funny" ads vs. "informational" ads)
• you should also sample from the
universe of those messages.
7/17/2016
Marketing Research
12
Instead of this (bad)….
Messages
1
7/17/2016
Marketing Research
13
…this (good)
Messages
7/17/2016
Sample
Marketing Research
14
The End
7/17/2016
Marketing Research
15
Download