Although experimental designs are considered to provide the strongest quantita-
tive evidence, that is not the only type of evidence needed for nursing practice.
Nurses are concerned with the whole person in interaction with the environ-
ment, and as such, they need information that is not always available from
experimental studies. Information about personal phenomena such as patients’
thoughts, beliefs, and subjective experiences, especially related to health care,
is often most appropriately obtained using either qualitative or quantitative
nonexperimental designs, which are descriptive in nature (Campbell & Stanley,
1966; Cook & Campbell, 1979). Nonexperimental designs also are important
when there is little information known about a particular phenomenon, when
it would be unethical to manipulate the independent variable, or when it is not
practical to conduct an experiment (e.g., lack of resources or excessive burden
to the subjects). The primary difference between nonexperimental and other
quantitative designs is that researchers do not actively manipulate the IV.
Even though the IV is not being manipulated, it is important for quantitative
researchers to exert as much control as possible to avoid threats to internal and
external validity. Researchers still need to protect against bias when conducting
these quantitative studies. Bias can be controlled with carefully selecting the
sample, having an adequate sample size to ensure sufficient statistical power,
clearly defining conceptual and operational variables, using reliable and valid
instruments to measure the variables of interest, and exerting control over the
conditions under which the data will be collected.
Nonexperimental designs can be used for the purposes of: (1) describing
a phenomenon in detail, (2) explaining relationships and differences among
variables, and (3) predicting relationships and differences among variables.
Because the IV is not manipulated, researchers using quantitative non-
experimental designs cannot make claims about cause and effect. Although
nonexperimental designs can be categorized in many ways, they fall into two
general categories: descriptive and correlational.
Descriptive Designs
The purpose of descriptive designs is to describe in detail a phenomenon of
interest. Descriptive designs provide a picture of a situation as it is naturally
happening without manipulation of any of the variables. This type of design
allows researchers to identify and document the different
characteristics of phenomena and describe the frequency
with which they occur. Because there is no manipulation
of variables and no attempt to establish causality, the terms
IV and DV should not be used with descriptive designs.
Instead, the term research variable is more appropriate.
Researchers use descriptive designs for a variety of
KEY TERMS
nonexperimental
designs: Research
designs that lack
manipulation of
the independent
variable and
random assignment
descriptive designs:
Designs that
provide a picture
of a situation
as it is naturally
happening without
manipulation of any
of the variablesFYI
Nonexperimental designs are descriptive
in nature and can be used to describe
phenomena in detail, explain differences
among variables, and predict relationships
and differences among variables.7.3 Nonexperimental Designs 181