PSYCH 213W: Chapter 13 Quasi Experiments
What is a Quasi Experiment?
§ Quasi Experiment
§ Research procedure in which the scientist must select subjects for different
conditions from pre-existing groups.
§ Also known as ex post facto designs, or after the fact, because the experiment is
conducted after the groups have been formed.
§ IE: in the case of a quasi experiment with the subject variable of sex as the IV, the
experiment takes place long after the participants become males or females.
§ If you performed an experiment using prexisiting classes of students, the two
classes would be ex post facto variables because the classes were formed before
you did the experiment.
§ In addition to the experimenter not being able to control the who in a quasi experiment,
in some other instances the experimenter may not me able to also control the what,
when, where, and how.
§ NOTE: the boundaries between true and quasi experiments are not always distinct.
§ If the experimenter has good control over all aspects of the experiment, we can call it a
true experiment.
§ The enough compromise of experimental control takes place, the research is then
considered a quasi experiment.
§ A True Experiment permits the experimenter the greatest degree of control in ruling out
alternative hypotheses, or alternative independent variables, as being the cause of the
difference between two groups or conditions.
§ Thus, a true experiment permits the most powerful control for confounding because all
other potential independent variables have been eliminated by randomly assigning
assigning subjects to conditions.
§ A quasi experiment leaves open the possibility that other differences exist between the
experimental and control conditions and thus presents other potential differences to
remain.
§ It is possible to have one experimental variable, and one quasi experimental variable in
an experiment.
§ IE: studying effects of two different teaching methods on classroom learning.
§ Our interest may be whether slow learners differ from fast learners in
their response to the teaching methods.
§ Experimental Variable: two teaching methods (assuming we assigned
students to sections), and
§ Quasi-Experimental Variable: classification of slow and fast learners.
True VS. Quasi: What is Better?
§ The degree of control is less in quasi experiments, however, the fact that we discussed
true experiments first does not mean that true experiments are necessarily better than
other types of investigations.
, § Rather, discussing true experiments first reflects the fact that as the degree of control
that the researcher can exercise decreases, the threats to validity of the conclusion
increase.
§ Of all things equal, one would choose a true experiment over a quasi experiment, over a
nonexperimental method.
§ However, many social-psychological phenomena are difficult to bring into the
laboratory in in a realistic fashion.
§ Therefore, a field study may be preferable to an experiment because the
advantage of realism and ecological validity outweighs the loss of control.
§ Nevertheless, you should first try to design a true experimental study and use the
other designs only when you believe that the gain in validity will be worth the
loss of control.
§ If a quasi experiment best answers your question, it is the best method.
Nonequivalent Control-Group Design
§ Nonequivalent Control-Group Design
§ Research design having both an experimental and a control group wherein
subjects are not randomly assigned to groups.
§ *Most typical design in the quasi experimental design realm.
§ Quasi experiments that employ Nonequivalent control-groups with pretest and posttest
may be interpretable or not.
§ Whether they can be interpreted depends on whether the pattern of results obtained
can be accounted for by possible differences in the groups or by something else in the
experiment.
Designs Without Control Groups
§ Interrupted Time-Series Design: research design that allows the same group to be
compared over time by considering the trend of the data before and after the
experimental manipulation.
§ In Chapter 10 we said to avoid the measurement of a single group before and after the
manipulation.
§ One way to improve on the One-group before-after design is to consider the trend of the
data before and after the manipulation, rather than simply comparing the average data
before and after.
§ The ideal situation would be a flat and stable baseline before the change,
followed by either an abrupt change to a new level, or a gradual change to a new
level.
§ However, a baseline that is changing in the opposite direction from the predicted
influence of the manipulation would still produce a valid comparison.
§ Repeated-Treatment Design: design in which a treatment is withdrawn and then
presented a second time.
What is a Quasi Experiment?
§ Quasi Experiment
§ Research procedure in which the scientist must select subjects for different
conditions from pre-existing groups.
§ Also known as ex post facto designs, or after the fact, because the experiment is
conducted after the groups have been formed.
§ IE: in the case of a quasi experiment with the subject variable of sex as the IV, the
experiment takes place long after the participants become males or females.
§ If you performed an experiment using prexisiting classes of students, the two
classes would be ex post facto variables because the classes were formed before
you did the experiment.
§ In addition to the experimenter not being able to control the who in a quasi experiment,
in some other instances the experimenter may not me able to also control the what,
when, where, and how.
§ NOTE: the boundaries between true and quasi experiments are not always distinct.
§ If the experimenter has good control over all aspects of the experiment, we can call it a
true experiment.
§ The enough compromise of experimental control takes place, the research is then
considered a quasi experiment.
§ A True Experiment permits the experimenter the greatest degree of control in ruling out
alternative hypotheses, or alternative independent variables, as being the cause of the
difference between two groups or conditions.
§ Thus, a true experiment permits the most powerful control for confounding because all
other potential independent variables have been eliminated by randomly assigning
assigning subjects to conditions.
§ A quasi experiment leaves open the possibility that other differences exist between the
experimental and control conditions and thus presents other potential differences to
remain.
§ It is possible to have one experimental variable, and one quasi experimental variable in
an experiment.
§ IE: studying effects of two different teaching methods on classroom learning.
§ Our interest may be whether slow learners differ from fast learners in
their response to the teaching methods.
§ Experimental Variable: two teaching methods (assuming we assigned
students to sections), and
§ Quasi-Experimental Variable: classification of slow and fast learners.
True VS. Quasi: What is Better?
§ The degree of control is less in quasi experiments, however, the fact that we discussed
true experiments first does not mean that true experiments are necessarily better than
other types of investigations.
, § Rather, discussing true experiments first reflects the fact that as the degree of control
that the researcher can exercise decreases, the threats to validity of the conclusion
increase.
§ Of all things equal, one would choose a true experiment over a quasi experiment, over a
nonexperimental method.
§ However, many social-psychological phenomena are difficult to bring into the
laboratory in in a realistic fashion.
§ Therefore, a field study may be preferable to an experiment because the
advantage of realism and ecological validity outweighs the loss of control.
§ Nevertheless, you should first try to design a true experimental study and use the
other designs only when you believe that the gain in validity will be worth the
loss of control.
§ If a quasi experiment best answers your question, it is the best method.
Nonequivalent Control-Group Design
§ Nonequivalent Control-Group Design
§ Research design having both an experimental and a control group wherein
subjects are not randomly assigned to groups.
§ *Most typical design in the quasi experimental design realm.
§ Quasi experiments that employ Nonequivalent control-groups with pretest and posttest
may be interpretable or not.
§ Whether they can be interpreted depends on whether the pattern of results obtained
can be accounted for by possible differences in the groups or by something else in the
experiment.
Designs Without Control Groups
§ Interrupted Time-Series Design: research design that allows the same group to be
compared over time by considering the trend of the data before and after the
experimental manipulation.
§ In Chapter 10 we said to avoid the measurement of a single group before and after the
manipulation.
§ One way to improve on the One-group before-after design is to consider the trend of the
data before and after the manipulation, rather than simply comparing the average data
before and after.
§ The ideal situation would be a flat and stable baseline before the change,
followed by either an abrupt change to a new level, or a gradual change to a new
level.
§ However, a baseline that is changing in the opposite direction from the predicted
influence of the manipulation would still produce a valid comparison.
§ Repeated-Treatment Design: design in which a treatment is withdrawn and then
presented a second time.
