Transcript Chapters 7-8 Short Version

MARKETING RESEARCH
CHAPTERS
7: Descriptive Research Design: Survey
and Observation
8 Experimentation
Descriptive Research Design:
Survey
• The survey method is a structured questionnaire
given to a sample of a population designed to
elicit specific information from respondents.
• The survey has the following advantages: ease,
reliability, and simplicity (the ease of removing
the information from the structured questionnaire
and putting the information into computerized
form).
• The survey has the following disadvantages:
respondents may be unable or unwilling to
provide the information and it may be difficult to
get a sampling frame to draw a representative
sample.
Descriptive Research Design:
Survey: Modes
• Telephone: Traditional and computerassisted telephone interviewing
• Personal: in home, mall intercept,
computer assisted personal interviewing
• Mail: mail panel, mail/fax interview
• Electronic: email or internet
Advantages and Disadvantages of
Telephone Interviewing
•
•
•
•
•
•
•
•
•
Interviews completed quickly
No travel time
Limited labor costs
Sample control high
Supervision from central location
Response rate good
Personal rapport difficult to achieve
Respondents can escape easily
Less tolerance for lengthy interviews
Advantages and Disadvantages of
Personal Interviews
• In-home:
Provides clarification to respondent
Can use visual aids
Easy to sample homes
Respondents tolerate longer interviews
Response rate good
However people home now so difficult to
find respondent at home
Interviewer control difficult
More social desirability problems
Interviewers can influence answers-interviewer bias
Advantages and Disadvantages of
Personal Interviews
• Mall Intercept:
Respondent is already present in mall
Research facilities in many malls
Appropriate when there is a need to
handle product
Response rate good
Social desirability issues
Interviewer bias
Limited data can be collected as people are in
a hurry
High cost
Advantages and Disadvantages of
Personal Interviews
• Computer Assisted Personal Interview
Holds respondents interest
Can ask completed set of questions
as computer automatically carries
out skip pattern
Computer administers interview
Interviewer bias limited
Useful when physical stimuli needed
Good control of data collection
Good response rate
Social desirability issues
Limited data
High cost
Advantages and Disadvantages of
Mail Methods
•
•
•
•
•
•
•
•
Economical
Can reach lots of people
No personal contact so no interviewer bias
Moderate to large amounts of data
People can easily throw away the survey
Low sample control
Response rates very low
Can result in non-response bias
Improving Survey Response Rates
•
•
•
•
Prior notification
Incentives (monetary and non-monetary)
Follow-up
Other facilitators of response such as
reminders
Observation Methods
• Personal observation
• Mechanical observation
Advantages and Disadvantages of
Observation
• Do not require conscious respondent input
• Interviewer bias control since observer only puts down
what is observed
• Easier to use with pets and children
• Best applied to phenomenon that occur frequently or
are of short duration
• We do not know why something is happening –
subject to subjective analysis
• Not good for infrequent behavior or behavior of long
duration (although anthropologists have used this
method effectively living with people for long periods
of time)
Causal Research Design:
Experimentation
• What is the meaning of causality?
• Causality: when the occurrence of variable
X increases the probability of the
occurrence of Y
• To have causality: if the two events occur
together one might cause the other, but
one must actually precede the other, and
there must be no other event that could
cause the second event
Experimentation
• To actually show cause and effect, we must set
up an experiment.
• An experiment is the process of manipulating
one or more independent variables and
measuring their effect on one or more
dependent variables while controlling for
extraneous variables.
• An independent variable is the cause of
something and usually manipulated by the
researcher
• The dependent variable is the effect of changes
in the independent variable
• Experiments use test units: anything sampled or
selected for an experiment.
Experimentation Continued
• We control for extraneous variables or other things than
the independent variable that might affect the response
of test units.
• The most usual way of controlling for extraneous
variables is to use random assignment of test units to the
experimental groups.
• There are often two types of groups in an experiment:
control and experimental groups.
• The control group is the comparison group and the
experimental group you will change in some way. You
will then compare results between the two groups.
• Normally, you will DESIGN your experiments
Experimental Design
• We can use specific symbols help organize how
we think about experimental design:
• X = The exposure of a group to some kind of
change often called a treatment
• O= The process of observation or measurement
of the dependent variable on the test units.
• R= The random assignment of test units to
separate treatments or groups.
Experimental Design Continued
• Designs can be pre-experimental, true experimental,
quasi-experimental, and what your book calls statistical
designs although with the other designs it is possible to
use statistics as well.
• Pre-experimental designs do not use randomization.
One shot case study: X O1
One-group pretest-posttest: O1 X O2
Static Group Design: EG: X O1
CG O2
Experimental Design Continued
• True experimental designs use
randomization which is one method for
controlling for extraneous variables.
• Randomization means that you know the
probability of selecting each test unit for a
particular group.
Experimental Design Continued
• Pretest-posttest control group design:
EG: R O1 X O2
CG: R O3
O4
The treatment effect is measured as the difference: (O2-O1)-(O4-O3).
There may be a problem with this design although it is very common:
interactive testing effects
• Posttest-Only control group design:
EG: R X O1
CG: R
O2
The treatment effect is: O1-O2
Validity in Experimentation
• Internal validity refers to whether the manipulation
of the independent variables or treatments actually
caused the observed effects on the dependent
variables. Control of extraneous variables is a
necessary condition for establishing internal validity.
• External validity refers to whether the cause-andeffect relationships found in the experiment can be
generalized. To what populations, settings, times,
independent variables, and dependent variables can
the results be projected?
Extraneous Variables
• History refers to specific events that are external to
the experiment but occur at the same time as the
experiment.
• Maturation (MA) refers to changes in the test units
themselves that occur with the passage of time.
• Testing effects are caused by the process of
experimentation. Typically, these are the effects on
the experiment of taking a measure on the dependent
variable before and after the presentation of the
treatment.
Extraneous Variables
• Instrumentation (I) refers to changes in the
measuring instrument, in the observers, or in the
scores themselves.
• Statistical regression effects (SR) occur when test
units with extreme scores move closer to the average
score during the course of the experiment.
• Selection bias (SB) refers to the improper
assignment of test units to treatment conditions.
• Mortality (MO) refers to the loss of test units while
the experiment is in progress.
Experimental Design Continued
• Quasi-experimental designs include time
series designs where you look at changes
in a variable over time.
Experimental Design Continued
• Statistical designs consists of a series of basic
experiments that allow for statistical control and
analysis of external variables.
• They offer advantages in that the effects of more
than one independent variable can be
measured, extraneous variables can be
controlled, and designs can be used where each
subject is measured more than once.
• Generally, these designs are called Analysis of
Variance Designs. One type of this design is
called factorial which we will discuss when we
look at ANOVA (analysis of variance designs).