Causal Research Design and Techniques

1
Causal Research Design and Techniques

• Week 11

2
Study Materials for Module 9

• Module 9 of study book
• Cavana et al. - Ch 12
• Jennings - Ch 1 pp. 18-19
• Ch 8 pp. 259
• Handout Cavana et al. Ch 12
• Coakes and Steed Ch 6, 7, 19
• Ch 6 t-tests one-sample t-test repeated
  measures t-test (parametric)
• Ch 19 Wilcoxon signed rank test p.202
  (non-parametric)
• Ch 19 Kruskal-Wallis test p.204
  (non-parametric)
• Ch 7 One-way between groups ANOVA
  (parametric)

3
Lecture overview

• Causal research design
• Experiments what are they?
• Experimental design
• Quasi-experimental designs
• True or Good experimental designs
• Validity internal and external
• Types of Experiments
• Ethical issues in experimental research
• Summary
• Tutorial

4
Research ProcessCausal Research Design
Exploratory
Descriptive
Causal
5
Causal research design

• What is it
• Research in which the major objective is to
  obtain evidence regarding cause-and-effect
  relationships
• How one variable (independent) effects, or is
  responsible for changes in another variable
  (dependent).
• If x, then y if then statements
• Example hypothesis
• If Yellow Page ads are in colour, then consumers
  attitudes towards the product will be higher than
  if the ad is in black and white

6
Experiments

• Causal relationship are determined by using the
  technique experiments
• Experiment - manipulation of an independent
  variable to see how it affects a dependent
  variable, while controlling the effects of
  additional extraneous variables
• Independent variable researcher has control
  manipulate the variable e.g. level of advertising
  expenditure, display location, price, type of
  product
• Dependent variable - we have little or no direct
  control over yet we have a strong interest in
  manipulating e.g. market share, sales, level of
  customer satisfaction

7
Experiments

• Extraneous variables are those that may have
  some effect on a dependent variable but yet are
  not independent variables
• e.g. brand of fuel effects fuel consumption per
  Km in cars what are the extraneous variables?
• e.g. type of display in a supermarket effects
  sales of apples - what are the extraneous
  variables?

8
Experimental Designs

• is a procedure for devising an experimental
  setting such that a change in a dependent
  variable may be attributed solely to the change
  in an independent variable
• Symbols
• O - measure of a dependent variable
• X manipulation, or change, of an independent
  variable
• R random assignment of subjects (consumers,
  stores, etc) to experimental control
  groups
• E experimental effect change of the dependent
  variable due to the independent variable

9
Experimental Designs cont

• Experimental group group of subjects exposed to
  the experimental treatment
• Control group group of subjects that are not
  exposed to the experimental treatment
• Treatment factor experimental factor to which
  the subjects are exposed i.e. advertising copy,
  graphic design, price
• Treatment level groupings within the
  independent variable i.e price 1.29 1.69
  1.99
• (basic experimental design has only one level of
  the treatment)

10
Experimental Designs cont

• Pretest measure the dependent variable prior to
  changing the independent variable
• Posttest measurement of the dependent variable
  is taken after changing the independent variable
• true or good experimental design isolates
  the effects of the independent variable on the
  dependent variable while controlling for effects
  of any extraneous variables
• Quasi experimental design
• Research design that lacks adequate control of
  extraneous variables - some design elements not
  included i.e. pretest, control groups

11
Quasi-experimental designs

• After-only design changing the independent
  variable (ID) and, after some period of time ,
  measuring the dependent variable (D)
• One group, before-after design (pre-test and
  post-test design) measure dependent variable
  pre and post treatment
• After-only with control (post-test only with
  experimental control group) two groups
  experimental control. Treatment applied to E
  and after sometime posttest measure taken of D
  for both groups

• Pretest Treatment Posttest
• EG X O1
• EG O1 X O2
• EG (R) X O1
• C (R) O2
• X - treatment
• EG - experimental group
• C control group
• R random assignment of subjects

12
True or Good experimental designs

• Before and after with control (pre-test
  post-test experimental control group designs)
• Solomon four group design (more rigorous
  design)
• Many more advanced experimental designs

• Pretest Treatment Posttest
• EG (R) O1 X O2
• C (R) O3 O4
• E (O2 - O1) (O4 O3)
• (O2 - O1) effect of treatment extraneous
  variables
• (O4 O3) change in influence of extraneous
  variable
• EG - experimental group
• C control group
• X - treatment
• R random assignment of subjects
• E - experimental effect (treatment effect)

13
Validity internal and external

• An experiment is valid if
• The observed changes in the dependent variable
  is, in fact, due to the independent variable,
  refers to - Interval validity
• Threats
• Selection bias experimental control group to
  be equivalent i.e. age, other relevant
  characteristics (otherwise introduce extraneous
  factors)
• Instrument - pre post to be the same same
  conditions
• History effects changes in the environment
  after pretest
• Maturation effect due to time between pre
  post test changes occur to the subject ie
  bored, tired etc
• Mortality when one or more subject (units) drop
  out after pretest
• Others - refer to study notes
• and

14
Validity internal and external cont

• If the results of the experiment apply to the
  real world outside the experimental setting.
  refers to - External validity
• Threats
• Realism was experiments context conduct
  artificial
• Selection of subject representative sample of
  the population
• Generalisability other factors that could cast
  doubt on the experiments results being relevant
  to the real world

15
Types of Experiments

• Laboratory contrived, artificial setting for
  the purpose of controlling the many possible
  extraneous variables that may affect the
  dependent variable.
• Achieves high level of internal validity
• Allows for control of extraneous variables
• Conducted quickly less expensively than field
  exp.
• Disadvantage conducted in an artificial setting
• Field measurements of the dependent variable
  are made on subjects (test units) in their
  natural setting
• Conducted in natural settings
• Disadvantages expensive timing consuming
• Much of the experimentation in marketing,
  conducted as field experiments, is known as test
  marketing

16
Decision points for embarking on an experimental
design (fig 12.6, p. 307
17
Ethical issues in experimental research

• Cant pressure people to participate
• Keep subjects safe - free from physical mental
  stress
• Must explain the procedures
• Must not deceive subjects as to the true purpose
  of the research
• Allow subjects to withdraw
• Confidentiality and privacy
• Debrief subject fully and accurately after the
  experiment

18
Summary

• Causal research cause effect relationship
• Experiment - manipulation of an independent
  variable to see how it affects a dependent
  variable, while controlling the effects of
  additional extraneous variables
• Experimental Design
• Quasi experimental design lacks adequate
  control of extraneous variables some design
  elements not included - pretest, control groups
• True or Good experimental designs - as per
  definition above
• Validity internal external main threats
• Types of Experiments lab field
• Ethical issues specific to experimental research

19
Next Topic

• Sampling process

20
Tutorial

• Presentation
• Class exercises (bring text/Ch 12 Cavana
  et.al to tut)
• SPSS prac session
• Coakes and Steed Ch 6, 7, 19
• Ch 6 t-tests one-sample t-test repeated
  measures t-test (parametric)
• Ch 19 Wilcoxon signed rank test p.202
  (non-parametric)
• Ch 19 Kruskal-Wallis test p.204
  (non-parametric)
• Ch 7 One-way between groups ANOVA
  (parametric)