Research Replication In Health Care
Why Replication is Needed
(a) scientific merit is established
(b) Type I and Type II errors are decreased
(c) construct validity is increased
(d) support for theory development is provided
(e) acceptance of erroneous results is prevented. Replication studies are essential for developing a scientific knowledge base in nursing.
Limitation of Replication in nursing Research
Incorporating research findings into nursing practice has been seriously hampered by the limited number of replication studies. Clarification of replication terminology can assist in advancing replication research. Three of the most often cited classifications of replication research have been developed by Finifter (1975), LaSorte (1972), and Lykken (1968).
Methods of Replication
Lykken (1968) identified three methods of replication: literal, operational, and constructive. Literal replication is an exact duplication of the original researcher's sampling, procedure, experimental treatment, data collection techniques, and data analysis.
Operational replication involves an
exact duplication of only the sampling and experimental procedures in the
original research to check whether the original design when used by other leads
to the same results. In constructive replication duplicate methods are
purposely avoided.
Replication Strategist
Finifter (1975) listed four replication strategies: identical, virtual, systematic, and pseudo. Identical replication involves a one-to-one duplication of the original study's procedures and conditions. In virtual replication the methods of the original study are re-created in varying degrees.
In systematic replication neither the methods nor the substance of the
original study is duplicated. Pseudo-replication is similar to identical and
virtual replication; however, data for pseudo replication are collected at the
same time as those for the original study. The simultaneous confirmation of the
study is built into the original design.
Retest Replication
LaSorte (1972) described retest, internal, independent, and theoretical replication. Retest replication involves repeating an original study with few, if any, significant changes in the research design. Internal replication is incorporated into the original study. Data for both the original study and its replicated study are collected simultaneously to provide a cross check for the reliability of the original results.
In independent
replication, significant modifications in the design of the original study are
made to verify the empirical generalization. In theoretical replication the
inductive process is used to examine the feasibility of fitting the empirical
findings into a general theoretical framework.
Replication Strategies Comparison
A comparison of these three replication classifications reveals that Finifter's (1975) identical replication is similar to Lykken's (1968) literal replication. The three classifications include strategies to approximate the original research design. Finifter calls this virtual replication, Lykken labels it operational, and LaSorte (1972) describes it as retest replication.
The purpose of choosing this type of replication is to
determine if the original findings can be confirmed when modest changes in the
research conditions have been made. When original findings are replicated,
confidence in the reliability of these results is enhanced.
All three classifications include an approach to increase empirical generalization by significantly modifying the original design. Finifter (1975) labels this systematic replication, Lykken (1968) describes it as constructive replication, and LaSorte (1972) calls it independent replication.
This type of
replication strategy is used when the researcher not only wants to validate
earlier work but also wants to extend the results and determine the degree of
generalizability.
Both Finifter (1975) and LaSorte (1972) specifically identify types
of replications where data for both the original and replication studies are
collected at the same time. According to Finifter , this is pseudo replication,
and to LaSorte it is internal replication. This type of replication provides
additional data, which is used as a cross check of the data's reliability.
LaSorte's classification is the only one that includes a replication strategy
to develop and verify theory.
Needs of Replication
Replication studies conducted in nursing have addressed topics such as nursing education, nurses' characteristics, perioperative nursing, body image during pregnancy, cardiac care, fetal monitoring, and time perception. When publishing replicated studies nurse researchers should include the following information:
(a) identification of the specific type of replication that is conducted
(b) provision of specific information on how a replicated study is the same as and different from the original study
(c) explanation of what is replicated and how.
This information will help readers to more clearly understand how the researchers methodically revised previous studies in a progressive manner. When publishing original studies, researchers should also explicitly detail the important points of their sampling and data collection techniques and their research design to aid replication of their work.
Authors
must be more diligent in identifying the minimum essential conditions and
controls necessary for producing findings because replication is crucial for the
further development of nursing knowledge.
Give your opinion if have any.