Evidence-Based Practice for Nurses

Patients in a diabetic clinic are randomly assigned to either the computerized

learning module group or the standard of care group. Knowledge about diabetes

is measured before and after teaching for all subjects. One way the researcher

can determine the efficacy of the intervention is to compare the pretest and

posttest scores of subjects in the intervention group to see if learning occurred.

Another way to evaluate the intervention would be to compare posttest scores

from the two groups to determine whether one teaching method increased

posttest scores more than another did.

A potential disadvantage of the two-group pretest-posttest design is that

threats to internal validity are introduced from repeated testing. In the previ-

ous example, the threat of testing could be reduced by providing subjects with

alternate forms of the diabetic knowledge test. Mortality is another threat

because subjects are measured more than once and some subjects may drop

out before the study is completed. Also, if the intervention or data collection

is burdensome, subjects may be more likely to withdraw from studies.

Two-Group Posttest-Only Designs

In experimental two-group posttest-only designs, researchers randomly assign

subjects to either an intervention group or a control group. In nursing studies,

the intervention is conducted with the experimental group (see Table 7-1), while

the control group receives the usual standard of care. After the intervention is

completed, the DV is measured at the same point in time in both groups. This

type of design is used when it is not possible or practical to measure the DV

before the intervention is implemented. For example, a researcher is interested

in examining the effect of music therapy on pain levels of patients undergoing

invasive procedures. Because patients have no pain prior to invasive procedures,

it would be impossible to determine whether the intervention decreased levels of

pain. Therefore, the only way a researcher can determine whether music therapy

effectively reduced pain would be to use a between-subjects design. Pain levels

from subjects who used music therapy would be compared to the pain levels of

subjects who did not. Because subjects are measured only once, threats of testing

and mortality are minimized. A disadvantage of the two-group posttest-only

design is that it is susceptible to the threats of selection bias. One cannot assume

that the two groups are equivalent, because there was no measurement at baseline.

Selection bias should be minimized because subjects are randomly assigned to

groups. The characteristics of subjects that might affect the DV should be equally

distributed between the two groups through the use of random assignment.

Solomon Four-Group Designs

As the name implies, there are four groups of subjects involved in the Solomon

four-group design. Two groups of subjects receive the intervention, and two

KEY TERMS

two-group posttest-

only designs:

Experimental

designs when

subjects are

randomly assigned

to an experimental

or control group

and measured after

the intervention

Solomon four-

group design:

An experimental

design with four

groups—some

receive the

intervention,

others serve as

controls; some are

measured before

and after, others are

measured only after

the intervention

174 CHAPTER 7 Quantitative Designs: Using Numbers to Provide Evidence