EXPERIMENTAL AND QUASI-EXPERIMENTAL DESIGNS FOR RESEARCH

women, we recoil from the with more numerous treatments

implication that our harsh in
the
Fisher
factorial
course
of
training
is experiment tradition represent
debeautifying.
and
instead important
elaborations
point to the hazards in the way tangential to the main thread
of a beautiful girls finishing of this chapter and are
college before getting married. discussed at the end of this
Such an effect is classified here section, subsequent to Design
as experimental mortality. (Of 6. But this perspective can
course, if we consider the serve to remind us at this point
same girls when they are that the comparison of X with
freshmen and seniors, this no X is an oversimplification.
problem disappears, and we The comparison is actually
have Design 2.)
with the specific activities of
the control group which have
THREE TRUE
filled
the
time
period
EXPERIMENTAL DESIGNS corresponding to that in which
experimental
group
The three basic designs to be the
treated in this section are the receives the X. Thus the commight
better
be
currently recommended designs parison
in
the
methodological between Xi and Xc, or between
literature. They will also turn X\ and Xo, or Xi and Xi. That
out to be the most strongly
recommended designs of this
presentation, even though this
endorsement is subject to many
specific qualifications regarding
usual practice and to some
minus signs in Table 1 under
external validity. Design 4 is
the most used of the three, and
for this reason we allow its
presentation
to
be
disproportionately extended and
to become the locus of
discussions more generally applicable. Note that all three of
these designs are presented in
terms of a single X being
compared with no X. Designs
these control group activities
are often unspecified adds an
undesirable ambiguity to the
interpretation
of
the
contribution of X. Bearing
these comments in mind, we
will continue in this section
the graphic convention of
presenting no X in the control
group.

EXPERIMENTAL AND QUASI-EXPERIMENTAL DESIGNS FOR RESEARCH

4. THE FRETEST-POSTTEST
CONTROL GROUP DESIGN

R Oi X O2
ROs

O4

Because the design so neatly

Controls for Internal Validity
controls for all of the seven
One or another of the above rival hypotheses described so
considerations
led far, the presentations of it have
psychological and educational usually not made explicit the
researchers between 1900 and control needs which it met. In
1920 to add a control group to the tradition of learning
Design 2, creating the presently research, the practice effects of
orthodox control group design. testing seem to provide the
McCall
(1923),
Solomon first recognition of the need for
(1949), and Boring (1954) a control group. Maturation
have given us some of this was a frequent critical focus in
history, and a scanning of the experimental
studies
in
Teachers College Record for education, as well
as in the
that period implies still more, nature-nurture problem
in the
for as early as 1912 control child development area.
groups were being referred to research on attitude change, In
as
without need of explanation in the early studies on the
(e.g., Pearson, 1912). The effects of motion pictures,
control group designs thus history may have been the
introduced are classified in this main
necessitating
chapter under two heads: the consideration. In any
event, it
present Design 4 in which
equivalent groups as achieved seems desirable here to discuss
by
randomization
are briefly the way in which, or
employed, and the quasi- the conditions under which,
experimen- tal Design 10 in these factors are controlled.
which extant intact comparison History is controlled insofar
groups
of
unassured as general historical events that
equivalence are employed. might have produced an Oi
O2 difference would also
Design 4 takes this form:
produce an O3Oi difference.
Note, however, that