Double-blind experimental designs have been used in health research to
control for threats of reactivity. In double-blind studies, neither subjects nor
individuals administering the treatments know whether subjects are receiving
experimental interventions or standard of care. This design is fairly common
in drug studies where a placebo pill can be manufactured to appear exactly
like the drug being studied.Effects of Selection
Because researchers most often study samples rather than an entire population,
representativeness of the sample is essential. If the sample does not represent
the population, effects of selection limit whether the study can be generalized
to the population. For example, a researcher is interviewing mothers. Because
interviews are being conducted only during the day and no child care is pro-
vided, individuals who work during the day or who have small children are
less likely to participate in this study. Consequently, researchers would not be
able to generalize to all mothers but only to mothers who do not work during
the day and mothers without young children.Interaction of Treatment with Selection
of Subjects
Because external validity is concerned with generalizing to other individuals,
it is only natural to ask, “Will this IV affect other people in the same way?” The
interaction of treatment with selection of subjects must be considered (Cook
& Campbell, 1979). This requires consideration of the difference between the
accessible population and the target population of interest (Kempthorne, 1961).
In the condom use example, the target population is all sexually active teens.
The accessible population is the group of teens from which the researcher
is actually able to obtain a sample of subjects. So, if the computer-based
intervention is found to be effective in increasing condom use in a sample
of students from a Midwestern suburban high school, how generalizable are
those findings to teens living in urban centers in the West or rural areas of
the South?Interaction of Treatment and Setting
Sometimes an interaction of treatment and setting can affect external valid-
ity. This interaction is concerned with whether results from an intervention
conducted in one setting can be generalized to another setting where the same
intervention is used. If the study on condom use was conducted with teens
waiting to be seen in a family planning clinic, can the findings from that study
be generalized to teens in a high school setting?KEY TERMS
double-blind
experimental
designs: Studies
in which subjects
and researchers
are unaware
whether subjects
are receiving
experimental
interventions or
standard of care
effects of selection:
Threats to external
validity when the
sample does not
represent the
population
interaction
of treatment
with selection
of subjects: A
threat to external
validity where
the independent
variable might not
affect individuals
the same way
interaction of
treatment and
setting: A threat
to external
validity when
an intervention
conducted in one
setting cannot be
generalized to a
different setting160 CHAPTER 6 Key Principles of Quantitative Designs