Experimental Designs

1
(No Transcript)
2
Chapter 12 Experimental Designs
3
Chapter Objectives

• understand the role and scope of experimental
  research in business
• distinguish between causal and correlational
  analysis
• explain the difference between laboratory and
  field experiments
• explain the following terms extraneous
  variables, manipulation, experimental and control
  groups, treatment effect, matching and
  randomisation
• discuss the seven possible threats to internal
  validity in experimental designs
• describe the different types of experimental
  designs
• explain the role of simulation in experimental
  research
• describe the ethical issues involved in
  experimental research

4
Experimental Designs
Laboratory Experiment
Field Experiment

Cause - Effect relationships established
by 1. Manipulating treatments 2.
Controlling for external or exogenous
variables Manipulation of Treatment Example
Three different teaching methods given to three
different groups of students Straight lectures
to 10 students simulation only, to another 10
students Both lectures and simulations to 10
other students Assess which results in greatest
amount of learning
5

• Simulation alone is ineffective.
• Lectures are more effective than no treatment at
  all.
• Both lectures and simulation are extremely
  effective.
• Cause - Effect relationship can be established
  because of
• Controls for age, etc. through either
  randomisation or matching of groups
• Because of an additional control group

6

• Control of Exogenous Variables through
• Random assignment of members to various groups
• Matched groups
• Control groups
• Example Different treatments may have different
  effects on people with differing interests, ages,
  expertise,etc.
• So, a) randomly assign members to different
  treatment groups. The differences will be
  randomly distributed. Systematic bias will be
  reduced.
• b) match the different groups as closely as
  possible in terms of age, interest, expertise,
  etc.
• c) have an additional control group of students
  who ar not exposed to any of the three
  treatments, and see how they learn and compare.

7
Controlled Variables
Variables that might affect the Cause - Effect
relationship among the IVs and DV, and hence need
to be controlled. Example 1. Age 2. Education
levels 3. Length of Service in
Organisation Might affect the relationship
between job characteristics and job satisfaction
8
Uncontrolled Variables
Variables or phenomena that occur unexpectedly
and can confound the results. Example
Advertising
Purchasing
(IV)
(DV)

• Age
• Life style

Sudden Unemployment
(Uncontrolled Variable)
(Controlled Variables)
9

• Lab Experiements can have tight controls and
  hence the validity of cause Effect findings is
  high ie., they have high internal validity. But
  their generalisability to real life is low,
  because of their tight controls ie., their
  external validity is low.
• Field Experiments (eg, different incentive plans
  (treatment0 in work organisations for assessing
  effect on productivity, have high external
  validity or generalisability (because they
  represent the actual situations), but have low
  internal validity (ie., cause effect
  relationships are contaminated because of no
  controls.)

10
Cause and effect relationship after randomisation
11

• FACTORS AFFECTING INTERNAL VALIDITY
• HISTORY EFFECTS
• MATURATION EFFECTS
• TESTING EFFECTS
• INSTRUMENTATION EFFECTS
• SELECTION BIAS
• STATISTICAL REGRESSION
• MORTALITY

12
History effects inexperimental design
13
Maturation effects on the cause and effect
relationship
14
Pre-test and post-test experimental group design
15
Post-test only with experimental and control
groups
Treatment effect (O1 - O2)
16
Pre-test and post-test experimental and control
groups

Treatment effect (O2 - O1) (O4 - O3)
17
Solomon four-group design
Treatment effect (E) could be judged by E 1
(O2 - O1) E 2 (O2 - O4 ) E 3 (O5 - O6) E 4
(O5 - O3 ) E 5 (O2 - O1) (O4 - O3 ) If
all Es are similar, the cause and effect
relationship is highly valid.
18
Major threats to internal validity in different
experimental designs

19
Simulation as experimentation
20
Example of a managementflight simulator
21
Ethical Issues in Experimental Research

• The following practices are considered unethical
• pressuring individuals to participate in
  experiments through coercion or applying social
  pressure
• giving out menial tasks and asking demeaning
  questions that diminish the subjects
  self-respect
• deceiving subjects by deliberately misleading
  them as to the true purpose of the research
• exposing participants to physical or mental
  stress
• not allowing subjects to withdraw from the
  research when they want to

22
Ethical Issues in Experimental Research(contd)

• using the research results to disadvantage the
  participants, or for purposes that they would not
  like
• not explaining the procedures to be followed in
  the experiment
• exposing respondents to hazardous and unsafe
  environments
• not debriefing participants fully and accurately
  after the experiment is over
• not preserving the confidentiality of the
  information given by the participants
• withholding benefits from control groups

23
Decision points for embarking on an experimental
design
24
A completely randomised design
25
A randomised block design
Blocking factor residential areas
Note that the Xs above indicate only various
levels of the blocking factor, and the Os (the
number of passengers before and after each
treatment at each level) are not shown, although
these measures will be taken.
26
The Latin square design
Day of the week
27
A 3 3 factorial design

28
(No Transcript)