there will be an associated change in the other. Evaluating statistical data pro-
vides information about the strength of the relationship, the direction of the
relationship (positive or negative), and whether the relationship is statistically
significant. There are three types of correlational designs commonly reported
in the literature: (1) descriptive correlational, (2) predictive correlational, and
(3) model-testing.
Descriptive Correlational Designs
Descriptive correlational designs build on comparative descriptive designs.
Comparative descriptive designs simply describe the phenomena as they occur
in two or more groups or among two or more variables associated with the
phenomena. Descriptive correlational designs are used when researchers are
interested in explaining the degree and characteristics of relationships that exist
among the variables or groups. For example, in the case of nurses’ willingness to
discuss sexual concerns with their patients, the researcher can use a descriptive
correlational design to determine the extent to which level of education or years
of experience is related to the nurses’ comfort with discussing sexual issues.
Predictive Correlational Designs
Sometimes there is insufficient empirical or theoretical literature for research-
ers to assume the degree or direction of the relationships among the variables.
Researchers use descriptive correlational designs to test nondirectional hypoth-
eses. Like descriptive studies, there is no IV or DV, and the variables are simply
referred to as “research” variables. In other situations, researchers have sufficient
evidence to predict the expected direction of relationships that will be found
among the variables. When this is the case, a predictive correlational design
is used. Researchers hypothesize which variables are predictors and which are
outcomes. Although researchers may be able to predict a statistically significant
relationship, the design is still correlational and causality cannot be assumed.
Therefore, even though it is common to see the terms IV and DV used with this
type of design, this is not technically correct. The variables should be referred
to as predictor and outcome.
There are two major aims of predictive designs. First, and most commonly
seen in nursing, researchers attempt to determine the amount of variance in
an outcome variable that can be explained by multiple predictor variables. For
example, a researcher is interested in determining which factors are most likely
to predict the quality of life (QOL) in patients receiving dialysis treatments.
The researcher could conduct a study to determine the relationship between
the hypothesized predictor variables of emotional distress, functional status,
marital status, employment status, age, and length of time on dialysis. The
researcher could then use statistical tests to determine the amount of variance
KEY TERMS
descriptive
correlational
designs:
Correlational design
type used to explain
the relationship
among the variables
or groups using
a nondirectional
hypothesis
predictive
correlational
design:
Correlational design
when researchers
hypothesize
which variables
are predictors or
outcomes7.3 Nonexperimental Designs 183