Textbook Notes (280,000)
CA (170,000)
U of G (10,000)
MCS (700)
Chapter 9

Chapter 9 Research Methods Notes.docx

Marketing and Consumer Studies
Course Code
MCS 3030
Tanya Mark

of 2
Chapter 9 Research Methods Notes
Quasi-Experimental Design
Quasi-experimental Design: Have several of the key features of randomized experimental designs such as pre-
post measurement & treatment control group comparisons by lack random assignment to a treatment group.
- Two types to be focused on are Nonequivalent groups design (NEGD) and the regression-discontinuity design
- Probably the most frequently used design in social research b/c its very intuitively sensible
- Has a pre-post measure: measure treatment group before and after treatment is given
- Has a control group that doesn't get the treatment
- Major challenge stems from the term nonequivalent,… if your comparison group is really similar to the program
group in all respects (except for receiving the program) this is a good design.
How do you assure groups are equivalent?
9-1 A The Basic Design
- Structured like a pretest-posttest randomized experiment, but lacks random assignment
- Most often use the intact groups that you think are similar to the control and treatment groups
- B/c its often likely the groups aren’t equivalent = why its called nonequivalent
- This design very susceptible to the internal validity threat of selection
- Seen as a useful method for determining whether a program or treatment is effective
- RD Design is a pretest-posttest program-comparison group strategy
- Participants are assigned to program or comparison groups solely on the basis of a cutoff score on a preprogram
- Cutoff criteration implies a major advantage of RD designs, they are appropriate when you want to target a
program or treatment to those who most need or deserve it
9-2 A The Basic RD Design
- RD design doesn’t require that pre and post measures be the same
- A single pretest-cutoff score is used to assign participants to either the program ro comparison group
C = indicates the groups are assigned by means of a cutoff score
O = the administration of a measure to a group
X = Depicts the implementation of a program
- Each group is described on a single line
The Role of the Comparison Group in RD Designs
- In experimental or other quasi-experimental designs, you either assume or try to provide evidence that the
program and comparison groups are equivalent prior to the program so that post-program differences can be
attributed to manipulation
RD Design doesn't assume this
- RD Design assumes that w/o the treatment, both pre-post relationship would be equivalent for the 2 groups
The Internal Validity of RD Design
- Only factors that would naturally induce a discontinuity in the pre-post relationship could be considered threats
to the internal validity of inferences from the RD design
9-2 B Statistical Power & The RD Design
- RD design has strong internal validity… stronger then the NEGD design… almost as strong as randomized
experimental design
- To achieve level of statistical accuracy, an RD design needs 2.75 times the # of participants as randomized
9-2 C Ethics & The RD Design
- RD design allows you to assign the treatment program to those who need or deserve it… therefore the attraction
is ethical
9-3 A The Proxy-Pretest Design
- Should never be selected by choice, but might find self in situation where you have to evaluate the program that
ahs already begun.
- Looks like a standard pre-post design with an important difference…. The pretest in this design is collected after
the program is given!
You use a proxy variable to estimate where the groups would have been on the pretest
- You can ask participants to estimate where their pre-test level would have been (recollection proxy-pretest
design) This would be helpful for assessing participant’s perceived change/gain.
- Or can use archived records to stand in for the pretest (Archived proxy-pretest design)
9-3 B The Separate Pre-Post Samples Design
Separate pre-post samples: A design in which the people who receive the pretest are not the same as the people
who take the post-test
- This isn’t strong b/c you cant match individual participant responses from pre to post; you can only look ar the
change in average
9-3 C The Double-Pretest Design
- a strong quasi-experimental design w/ respect to internal validity
- Includes 2 measures prior to the program
- Consequently, if the program and comparison group are maturing at different rates, you should detect this as a
change from pretest 1 to pretest 2 (controls for selection-maturation threats)
9-3 D The Switching Replications Design
- Also strong w/ respect to internal validity, & b/c it allows for 2 independent implementations lf the program, it
may enhance external validity or generalizability
- Has 2 groups and 3 waves of measurement
- In 1st phase f the design, both groups are given or pretests, one is given program, and both are post tested
- In 2nd phase the original comparison group is given the program while the original program group serves as the
9-3 E The Nonequivalent Dependent Variables (NEDV) Design
- Its an extremely weak design in terms of internal validity
- Potential extremely powerful causal assessment possibility
- You have a program designed go change a specific outcome
- Key is that the control variable has to be similar enough to the target variable to be affected in the same way by
history, maturation, and the other single group internal validity threats, but not so similar it is affected by the
The Pattern-Matching NEV Design
- Can make stronger by adding multiple outcome variables
- You need many outcome variables and a theory that tells how affected each variable will be by the program
9-3 F The Regression Point Displacement(RPD) Design
- A simple quasi-experimental strategy that has important implications especially for community-based research
- You look at pre-post indicators of the program community and see whether there is a change.
If you are relatively enlightened, you seek out another similar community and use it as a
- This design aims to enhance he single program unit situation by comparing the performance on rhat single unit
w. the performance of a large set of comparison units