Textbook Notes (290,000)
CA (170,000)
UTSC (20,000)
Psychology (10,000)
PSYB01H3 (600)
Chapter 8

PSYB01H3 Chapter Notes - Chapter 8: Functional Magnetic Resonance Imaging, Ex Post Facto Law, Random Assignment

Course Code
David Nussbaum

This preview shows pages 1-2. to view the full 7 pages of the document.
PSYB01 Chapter 8 – Quasiexperimental and Nonexperimental Designs
Quasiexperimental Design
- In many instances, we cannot directly manipulate an independent variable:
gender, race, age, ethnicity, SE, locale, diagnosis, personality traits, and
personal history are just some examples
- For these kinds of variables: experimenters must select participants that have
a certain characteristic or those who have been exposed to specified events
- Subject variables: The above listed treated as an independent variable
- Natural treatment: Naturally occurring event treated as a quasi-independent
variable which can be measured but not manipulated (levels may be exposure
or nonexposure)
- Quasiexperiment: investigates effects of a quasi-independent variable on a
dependent variable (quasi meaning “as if” or “to a degree”)
- True experiments would be first option but ethically and practically,
quasiexperiments are needed
- 2 types of quasiexperimental designs
oNonequivalent-control-group-designs: have experimental and
comparison groups designated before treatment occurs and are not
created by random assignment
oBefore-and-after designs: pretest and posttest but no comparison group,
they are their own controls
- Natural Treatments as Quasi-Independent Variables
oExample would be 9/11, considered a natural treatment because it is a
real world event that experiments have no control over
oSome natural treatments may be: war, presidential assassinations and
terrorist attacks, natural disasters, ordinary system-level practices and
policies such as new educational curricula, school closings, changes in
rules governing drug prescriptions, etc.
oEx post facto: “after the fact” analysis, and its effect on a particular
dependent variable such as memory, suicide rates or grades
- Subject Variables as Quasi-Independent Variables
oMixed factorial designs: combines both between- and within- subjects
- Nonequivalent-Control-Group Designs

Only pages 1-2 are available for preview. Some parts have been intentionally blurred.

PSYB01 Chapter 8 – Quasiexperimental and Nonexperimental Designs
oControl groups can never be considered equivalent to experimental
oHave experimental and control groups that have been predetermined or
predesignated by either an existing subject characteristic or an already-
occurred natural treatment
oRuth Wageman (1995): used quasiexperimental design to investigate
organization of work tasks and rewards (interdependence, autonomy or
oOften high in external validity and ecological validity (high in
generalizability and high in realism)
- Matching
oIndividual matching: individual cases in treatment group are matched
with similar individuals in control group
oAggregate matching: finding comparison group that has similar
distribution on key variables: same average age, same percentage
female, etc.
oProblems with matching
Regression artifact: threat to internal validity, when subjects are
chosen for a study because of extreme scores that become less
extreme on posttest due to natural cyclical or episodic change in
Regression to the mean: Trend for extreme score son a measure
to move closer to group average when retested (unreliable)
- Before-and-After Designs
oAbsence of comparison group
oSimplest type is fixed-sample panel design: one pretest and one posttest
oInterrupted-time-series design: single experimental group for which
we’ve obtained multiple observations before and after a naturally
occurring treatment (e.g. comparing accidents before and after a new
law goes into effect)
oE.g. Study by John Gibbons (2007): discouraging use of antidepressants
for children, coincided with large jump in adolescent suicide rates
You're Reading a Preview

Unlock to view full version