Lecture Marketing research (12th edition) - Chapter 14: Sampling fundamentals

1

Marketing Research
Aaker, Kumar, 
Leone and Day 
Twelfth Edition
Instructor’s 


2

2

Chapter Fourteen

Sampling Fundamentals

Marketing Research 12th Edition


3

3

Sampling Fundamentals
When is census appropriate?
•
•

•
•

Population size is quite small
Information is needed from every individual in 
the population
Cost of making an incorrect decision is high 
Sampling errors are high

Marketing Research 12th Edition


4

4

Sampling Fundamentals (Contd.)
When is sample appropriate? 
•
•

•
•

•
•

Population size is large
Both cost and time associated with obtaining information 
from the population is high
Quick decision is needed
To increase response quality since more time can be spent 

on each interview
Population being dealt with is homogeneous
If census is impossible

Marketing Research 12th Edition


5

5

Error in Sampling
Total Error
▫ Difference between the true value and the observed 

value of a variable
•

Sampling Error
▫ Error is due to sampling

•

Non­sampling Error
▫ Error is observed in both census and sample

Marketing Research 12th Edition


6

6

Error in Sampling (contd.)

Marketing Research 12th Edition


7

7

Sampling Process
Determining Target Population
•
•
•
•
•
•
•

Well thought out research objectives 
Consider all alternatives
Know your market
Consider the appropriate sampling unit
Specify clearly what is excluded
Should be reproducible
Consider convenience

Marketing Research 12th Edition


8

8

Sampling Process (Contd.)
Determining Sampling Frame
•
•

List of population members used to obtain a sample
Issues:
▫

Obtaining appropriate lists

▫

Dealing with population sampling frame differences


Superset problem



Intersection problem

Selecting a Sampling Procedure

•
•

Choose between Bayesian and Traditional sampling  procedure
Decide whether to sample with or without replacement

Marketing Research 12th Edition


9

9

The Sampling Process

Marketing Research 12th Edition


10

10

Sampling Techniques

Marketing Research 12th Edition


11

11

Types of Stratified Sampling
Proportionate Stratified Sampling 
•

•

Number of objects/sampling units chosen from each group is 
proportional to number in population
Can be classified as directly proportional or indirectly 
proportional stratified sampling

Disproportionate Stratified Sampling
•

•

Sample size in each group is not proportional to the respective 
group sizes
Used when multiple groups are compared and respective group 
sizes are small

Marketing Research 12th Edition


12

12

Directly Proportional Stratified Sampling

Marketing Research 12th Edition


13

13

Inversely Proportional Stratified Sampling
q

q

q

Assume that among the 600 consumers in the population, 200 are heavy 
drinkers and 400 are light drinkers.  
If a research values the opinion of the heavy drinkers more than that of the 
light drinkers, more people will have to be sampled from the heavy 
drinkers group. 
If a sample size of 60 is desired, a 10 percent inversely proportional 
stratified sampling is employed.

The selection probabilities are computed as follows:
Denominator

600/200 + 600/400 = 3 + 1.5 = 4.5

Heavy Drinkers proportional and
sample size

3/ 4.5 = 0.667; 0.667 * 60 = 40

Light drinkers proportional and
sample size

1.5 / 4.5 = 0.333; 0.333 * 60 = 20

Marketing Research 12th Edition


14

14

Cluster Sampling
•
•

•
•

Involves dividing population into subgroups 
Random sample of subgroups/clusters is 
selected and all members of subgroups are 
interviewed 
Very cost effective
Useful when subgroups can be identified that 
are representative of entire population

Marketing Research 12th Edition


15

Comparison of Stratified & Cluster Sampling Processes

Stratified sampling

Cluster sampling

Homogeneity within group

Homogeneity between groups

Heterogeneity between groups

Heterogeneity within groups

All groups are included

Random selection of groups
Sampling efficiency improved by

Sampling efficiency improved by
decreasing cost at a faster rate than
increasing accuracy at a faster rate than
accuracy.
cost

Marketing Research 12th Edition


16

16

Systematic Sampling
•

•
•

Involves systematically spreading the sample 
through the list of population members
Commonly used in telephone surveys
Sampling efficiency depends on ordering of the 
list in the sampling frame

Marketing Research 12th Edition


17

17

Non Probability Sampling
•

•

Costs and trouble of developing sampling 
frame are eliminated
Results can contain hidden biases and 
uncertainties
Used in:
•
•
•
•
•

  The exploratory stages of a research project
  Pre­testing a questionnaire
  Dealing with a homogeneous population
  When a researcher lacks statistical knowledge
  When operational ease is required

Marketing Research 12th Edition


18

18

Types of Non Probability Sampling

Marketing Research 12th Edition

19

19

Quota Sampling ­ Example

Marketing Research 12th Edition


20

20

Non­Response Problems
•

Respondents may:
▫
▫
▫

•

•

•

Refuse to respond
Lack the ability to respond
Be inaccessible

Sample size has to be large enough to allow for non 
response
Those who respond may differ from non respondents in 
a meaningful way, creating biases
Seriousness of non­response bias depends on extent of 
non response

Marketing Research 12th Edition


21

21

Solutions to Non­response Problem
•

•

•

Improve research design to reduce the 
number of non­responses
Repeat the contact one or more times (call 
back) to try to reduce non­responses
Attempt to estimate the non­response bias

Marketing Research 12th Edition

22

22

Shopping Center Sampling
•

•

20% of all questionnaires completed or 
interviews granted are store­intercept 
interviews
Bias is introduced by methods used to select

Marketing Research 12th Edition


23

23

Shopping Center Sampling (Contd.)
Solutions to Bias
•

Shopping Center Bias
▫
▫

•

Use several shopping centers in different neighborhoods
Use several diverse cities

Sample Locations Within a Center
▫
▫
▫

Stratify by entrance location
Take separate sample from each entrance
To obtain overall average, strata averages should be combined by 
weighing them to reflect traffic that is associated with each 
entrance

Marketing Research 12th Edition


24

24

Solutions to Bias (contd.)
•

Time Sampling
▫ Stratify by time segments
▫ Interview during each segment
▫ Final counts should be weighed according to traffic 

counts
•

Sampling People versus Shopping Visits – 
Options:
▫ Ask respondents how many times they visited the 

shopping center during a specified time period, 
such as the last four weeks and weight results 
according to frequency

Marketing Research 12th Edition


25

25

Different Levels of Sampling 
Frames

Marketing Research 12th Edition