I want to test a wound treatment or educational program in my clinical setting with patient
groups that are convenient or that already exist, How do I do it?
Implementing & evaluating wound research conducted using quasi-experimental designs
AAWC WocSpec Research Committee, Barbara M. Bates-Jensen, PhD, RN
ABSTRACT
PURPOSE: To provide information on how to evaluate and improve
methodology of wound research conducted using quasi-experimental designs.
BACKGROUND: Research methodology is concerned with how a study’s
design is implemented and how the research is conducted. Methodology
determines the quality of the data generated by the study and specifies: when
and how often data are collected, construction of data collection measures, who
data is collected from, how data is analyzed, and how findings are presented.
Study design is a part of research methodology but not the only part.
METHODS: Quasi-experimental studies include: pretest posttest nonequivalent
group design, time series design, and non-equivalent before-after design. The
purpose of quasi-experimental research is to approximate the conditions of the
true experiment in a setting which does not allow the control and/or
manipulation of all relevant variables. The researcher must clearly understand
what compromises exist in the internal and external validity of the design and
proceed within these limitations. We use quasi-experimental designs in the
clinical setting when control of all variables is not possible. Quasi-experimental
designs are better than pre-experimental studies in that they use a method of
comparing groups. They are limited in that they fail to randomize the groups.
This research is characterized by methods of partial control based on a careful
identification of factors influencing both internal and external validity. Methods of
increasing the strength of quasi-experimental wound studies are outlined.
CONCLUSIONS: Quasi-experimental designs follow true experimental steps
but typically do not control all variables and do not include randomization. The
aim of quasi-experimental design is to evaluate the influence of a variable on
one group and not on another. Limitations relate to the researcher’s attention to
internal and external validity. Methods to improve the quality of wound research
using quasi-experimental designs include attending to internal validity or did the
treatment make the difference in this specific instance rather than other factors?
And external validity or to what populations, settings, or treatment variables can
this observed effect be generalized?
BACKGROUND
Research methodology determines the quality of
the data collected in a study, it concerns how the
design is implemented and specifies:
 when and how often data are collected
 development of data collection measures &
instruments
 who data is collected from (the sample)
 how data are analyzed
 how findings are presented.
Whether or not study results are really the result of
the variable on the group under investigation in the
study (internal validity) and not some other variable
AND
Whether or not the study results can be expected
to be the same with other groups of people
(external validity) are issues that are addressed by
methodology.
Research design is an important part of
methodology. Study design is a part of research
methodology but not the only part.
QUASI-EXPERIMENTAL
METHODS
Nonequivalent Before-after
Designs
The purpose of quasi-experimental research is to
approximate the conditions of the true experiment in
a setting which does not allow the control and/or
manipulation of all relevant variables. Quasiexperimental research designs:
 answer questions using as much of an
experimental approach as possible in a clinical
setting
 follow basic experimental steps but do not
include randomization and may not control other
variables; uses existing or convenient groups
Quasi-experimental research designs include:
1. Pretest posttest nonequivalent group design
2. Time series design
3. Non-equivalent before-after design
Pretest Posttest Nonequivalent
Group Designs
Treatment Group
O1
Control Group
O1
X
O2
O2
O1 = Pre observation measurement or Pretest
X = Wound Treatment or Educational Program
O2 = Post observation measurement or Posttest
 A strong and widely used quasi-experimental design.
 Differs from true experiments because there is no
randomization, the groups are not equal.
 Comparing pretest results can show the level of
equivalence between the groups.
Time Series Designs
Treatment Group
O1
O2
O3
Control Group
O1
O2
O3
X
O4
O5
O6
O4
O5
O6
O1 –O3= Pre observation measurements or Pretests
X = Wound Treatment or Educational Program
O4 –O6 = Post observation measurement s or Posttests
 To determine the influence of a variable introduced
after a series of initial observations
 Differs from true experiments because there is no
randomization, the groups are not equal.
 Comparing pretest results can show the level of
equivalence between the groups.
Treatment Group 1
O1
X
O2
Treatment Group 2
O1
X
O2
O1 = Pre observation measurement or Pretest
X = Wound Treatment or Educational Program
O2 = Post observation measurement or Posttest
 Two groups with known differences before the study
begins are tested and compared.
 The same treatment is done for both groups and
differences in posttests are assumed to be related to the
difference in the two groups.
Improving Quasi-experimental
studies: Internal Validity
Internal Validity asks the question: did, in fact, the
experimental treatments make a difference in this specific
instance?
• 8 classes of extraneous variables which, if not
controlled, may produce effects that can be confused with
the effects of the experimental treatment:
1. History: Did something happen between the first and
second measurements in addition to the treatment
which might affect the measurement?
 Use a control group.
2. Maturation: the process of growing, developing,
aging or maturing which takes place in the individual
during the study period.
 Use a control group.
3. Pre-testing: the effect created on the second
measurement by having a measurement before the
experiment (the pre-test).
 Use additional groups that do not receive the
pretest or experimental treatment. These groups
are in addition to the regular control group.
4. Instrumentation: changes in measurement due to
changes in instrumentation calibration or changes in
the observers.
 Use a control group. Use quality control monitoring
during the study to help detect/control changes in
instrument use/measurement. Include reliability
testing throughout the study, include instrument
calibration and adequate training.
5.
Statistical Regression: occurs when groups have
been selected based on extreme scores.
 Use a control group and randomization . Select
groups that are not at the extreme of characteristics.
6. Selection: biases resulting from the differential
selection of subjects.
 Use of randomization procedures and pretesting can
help by determining the presence or knowledge that
exist before the treatment.
7. Experimental Mortality: the differential loss of
subjects from the comparison groups.
 Use pretesting as subjects can be removed from the
entire comparison by removing their pretest.
8. Selection-Maturation Interaction: interaction effects
between the selection and maturation which can be
mistaken for the effects of the experimental variable.
 Use a control group and randomization
Quasi-experimental studies:
External Validity
External Validity asks the question: to what
populations, settings, treatment, and measurement
variables can this effect be generalized?
• Just like internal validity, 4 classes of variables which, if
not controlled, may produce effects that can only be
attributed to the specific sample, setting, and procedures
in the current study. Thus, there is no way to use the
findings with other groups.
1. Interaction effects of Selection and the experimental
variable. Subjects selected from a small group or one
with particular characteristics can’t be generalized to
any larger group.
2. Reactive or Interaction effect of Pretesting- the
pretesting modifies the subjects in such a way that
they respond to the experimental treatment differently
than will un-pretested persons making them
unrepresentative of the population.
3. Reactive effects of experimental procedures arising
from the experimental setting which will not occur in
non-experimental settings.
4. Multiple-Treatment Interference effects due to
multiple treatments applied to the same subjects so
the findings can only be generalized to persons
exposed to the same treatments in the same order of
presentation.