Transcript Document

Experimental Research
Jeremy Kees, Ph.D.
Evidence that supports a causal
inference…
• Concomitant variation--evidence of the extent
to which X and Y occur together or vary
together in the way predicted by the
hypothesis
• Time order of occurrence of variables-evidence that shows X occurs before Y
• Elimination of other possible causal factors—
evidence that allows the elimination of factors
other than X as the cause of Y
– X -- the presumed cause
– Y -- the presumed effect
Types of Experiments
Laboratory Experiment
Experiment
Scientific investigation in which
an investigator manipulates
and controls one or more
independent variables and
observes the dependent
variable for variation
concomitant to the
manipulation of the
independent variables
Research investigation in which
investigator creates a situation
with exact conditions, so as to
control some, and manipulate other,
variables
Field Experiment
Research study in a realistic situation
in which one or more independent
variables are manipulated by the
experimenter under as carefully
controlled conditions as the situation
will permit
3
Definitions and Concepts
• Independent variables (IV) are variables or alternatives
that are manipulated and whose effects are measured and
compared, e.g., price levels.
• Test units are individuals, organizations, or other entities
whose response to the independent variables or treatments
is being examined, e.g., consumers or stores.
• Dependent variables (DV) are the variables which
measure the effect of the independent variables on the test
units, e.g., sales, profits, and market shares.
• Extraneous variables are all variables other than the
independent variables that affect the response of the test
units, e.g., store size, store location, and competitive effort.
– Covariates
Validity
• 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 causeand-effect relationships found in the experiment
can be generalized. To what populations,
settings, times, independent variables and
dependent variables can the results be projected?
Causal Research (Experimental
Design)
• Internal Validity
Causal Research (Experimental
Design)
• External Validity
Threats to Validity
• History--Specific events external to an
experiment, but occurring at the same time,
which may affect the criterion or response
variable
• Maturation--Processes operating within the test
units in an experiment as a function of the
passage of time per se
• Testing--Contaminating effect in an experiment
due to the fact that the process of
experimentation itself affected the observed
response
Threats to Validity
• Instrument Variation--Any and all changes in the
measuring device used in an experiment that might account
for differences in two or more measurements
• Statistical Regression--Tendency of extreme cases of a
phenomenon to move toward a more central position
during the course of an experiment
• Selection Bias--Contaminating influence in an experiment
occurring when there is no way of certifying that groups of
test units were equivalent at some prior time
• Experimental Mortality--Experimental condition in
which test units are lost during the course of an experiment
Controlling Extraneous Variables
• Randomization refers to the random assignment
of treatment conditions to experimental groups
by using random numbers. This is the key to
internal validity (extraneous variables are equal
across groups due to random assignment).
Assumed to produce ‘balancing’ across groups -• Comparable groups due to randomness of
assignment; ‘average’ participant is the same
across groups for non-manipulated variables
(e.g., Distribution of extraneous variance and variables are
constant across groups)
– if 65% female in one group, about same in others
Controlling Extraneous Variables
(less effective ways)
– Matching involves comparing test units on a
set of key background variables before
assigning them to the treatment conditions.
– Statistical control involves measuring the
extraneous variables and adjusting for their
effects through statistical analysis.
– Design control involves the use of
experiments designed to control specific
extraneous variables.
Characteristics of “Good”
Experiments
• Random assignment
• Comparison group/control group
• As a source of comparison
• As a control for rival hypotheses
• Generalizability/external validity
• Random selection
Limitations of Experimentation
• Experiments can be time consuming, particularly if
the researcher is interested in measuring the longterm effects.
• Experiments are often expensive. The requirements of
experimental group, control group, and multiple
measurements significantly add to the cost of
research.
• Experiments can be difficult to administer. It may be
impossible to control for the effects of the extraneous
variables, particularly in a field environment.
• Competitors may deliberately contaminate the results
of a field experiment.
Causal Research
“Who Can Resist an Oreo? Choice Behavior and Gender Differences
when Body Image Anxiety is made Salient,” Presented at the
Marketing and Public Policy Conference (2005).
Lit Review
• The majority of women and men are unhappy with their
appearance (Warner 2002)
• Media images may result in body dissatisfaction (Shaw 1995),
decreased perception of personal attractiveness (Odgen and
Mundray 1996) and body image anxiety (Richins 1991)
• Media images are a primary factor that leads to body image
anxiety (Richins 1991, 1995)
• Social Comparison (Festinger 1954)
• Differences in self-monitoring determine to what extent people
use internal versus external info to guide decisions and behaviors
(Snyder 1980)
Causal Research
“Who Can Resist an Oreo? Choice Behavior and Gender Differences when
Body Image Anxiety is made Salient,” Presented at the Marketing and
Public Policy Conference (2005).
Research Questions
• Can exposure to models in advertisements affect body esteem and
choice behavior in addition to body image anxiety?
• Do individual differences in self-monitoring a) impact BIA, body
esteem, and behavior and b) moderate the impact of model
exposure?
• Are there differences in how women versus men are impacted by
model ideals in ads?
Hypos
H1: Exposure to a model in an advertisement will result in (a) higher levels
of body image anxiety and (b) lower levels of body esteem.
H3: Consumers with higher tendencies to self-monitor will have stronger
reactions to the exposure to a model in an advertisement than those
consumers who do not self-monitor.
Causal Research
“Who Can Resist an Oreo? Choice Behavior and Gender Differences
when Body Image Anxiety is made Salient,” Presented at the
Marketing and Public Policy Conference (2005).
Research Design
• 3 (model exposure) X 2 (self-monitoring) for both males
and females
• N = 240 Undergraduates
• Dependent Variables
Females
Males
Body Image Anxiety
Physical Condition (10 items, α = .85)
Upper Body Strength (6 items, α = .81)
Physical Attractiveness (3 items, α = .73)
Cookie Choice
Body Image Anxiety
Physical Condition (7 items, α = .96)
Body Weight (9 items, α = .98)
Sexual Attractiveness (2 items, α = .72)
Cookie Choice
Causal Research
“Who Can Resist an Oreo? Choice Behavior and Gender Differences when
Body Image Anxiety is made Salient,” Presented at the Marketing and
Public Policy Conference (2005).
Results
Effects of Model Prime and Self Monitoring on Dependent Variables
MANOVA Results
Univariate F Values
Wilks’
Lambda
F-Value
Body Image
Anxiety
Physical
Condition
Body
Weight
Sexual
Attractiveness
Model Prime (MP)
.57
13.57***
42.77***
39.5***
38.03***
46.76***
Self Monitoring (SM)
.86
3.09**
0.10
2.53
6.55**
5.42**
MP X SM
.87
2.77**
0.56***
5.63**
4.91**
5.42**
Independent Variables
***p < .01
**p < .05
*p < .10
Causal Research
“Who Can Resist an Oreo? Choice Behavior and Gender Differences when
Body Image Anxiety is made Salient,” Presented at the Marketing and
Public Policy Conference (2005).
Results
Self Monitoring
5.50
low
Satisfaction with Body Weight
high
5.00
4.50
4.00
3.50
no model
female model
Prime
Causal Research
“Who Can Resist an Oreo? Choice Behavior and Gender Differences when
Body Image Anxiety is made Salient,” Presented at the Marketing and
Public Policy Conference (2005).
Results
Regular Oreo
Reduced Fat Oreo
14
No Oreo
14
prime
14
no model
12
12
10
10
10
8
Count
12
Count
Count
female model
8
8
6
6
6
4
4
4
2
2
2
0
0
low
high
Self Monitoring
0
low
high
Self Monitoring
low
high
Self Monitoring
To conclude…
• Experiments are the only way to show
causation
– But often take a back seat to descriptive
studies due to time, cost, and control issues
• Exploratory and descriptive studies are
useful, but be careful not to infer too
much
– Correlation is not causation
• Again, let your research questions dictate
your design!