Evidence-Based Practice for Nurses

self-care behaviors of people with high blood pressure. The researcher hypoth-

esizes that home visits and an interactive website will be effective in lowering

blood pressures. Two experiments could be used to test each intervention sepa-

rately, or a factorial design could be used to test both interventions separately,

while testing the effects of a combined intervention for controlling high blood

pressures. Table 7-2 depicts the 2 × 2 factorial design for this example. Note

that there are three intervention groups (home visit alone, website alone, and

combined home visit and website) and one control group that receives neither

the home visits nor the interactive website.

Because of its complex nature and sophisticated data analysis, more

experienced researchers tend to use this type of design. Like Solomon four-

group designs, sample sizes need to be large because there are a minimum

of four groups. However, time and effort are saved by not having to conduct

multiple studies.

Crossover Designs

In the discussion of longitudinal designs, crossover designs were mentioned.

Although only one group of subjects is used, crossover designs include the

three essential components of an experiment (see Table 7-1). Independent

variables are manipulated by the researcher, and the order in which the

interventions are administered to subjects is randomized. Because subjects

serve as their own control group, the criterion of control is met because it

is assumed that each subject will remain stable on the extraneous variables

during the study.

Smaller sample sizes are needed because only one group of subjects is re-

quired, and subjects serve as their own controls. A disadvantage is that subjects

may continue to engage in the first intervention after moving to the second.

Website Intervention

YES NO

Home Visit

Intervention

YES Both home visit

and website

Home visit alone

NO Website alone No intervention

control group

TABLE 7-2 Factorial Design

KEY TERM

crossover designs:

Experimental

designs that

use two or more

treatments; subjects

receive treatments

in a random order

176 CHAPTER 7 Quantitative Designs: Using Numbers to Provide Evidence