6 Evidence Appraisal: Research 101
Meta-Analysis
Meta-analysis offers the advantage of objectivity because the decisions made by
the study reviewers are explicit, and the integration of studies uses statistics ap-
plied to the data set (Polit & Beck, 2017). By combining the results of multiple
studies in a meta-analysis, the number of study participants from each individual
study are combined. This increases the statistical power and the probability of
detecting a true relationship between the independent and dependent variables
(Polit & Beck, 2017). As mentioned earlier, the common metric called effect
size (ES), a measure of the strength of the relationship between two variables,
is developed for each of the primary studies. A positive ES indicates a positive
relationship (as one variable increases, the second variable increases); a negative
ES indicates a negative relationship (as one variable increases or decreases, the
second variable moves in the opposite direction). By combining results across a
number of smaller studies, the researcher can increase the power, or the probabil-
ity, of detecting a true relationship between the intervention and the outcomes of
the intervention (Polit & Beck, 2017). When combining studies for meta-analysis,
the researcher can statistically analyze only those interventions (independent
variables) and outcomes (dependent variables) that the studies have in common.
The results of the meta-analysis are often displayed in a forest plot (see Figure
6.1). A forest plot, which is a graphical representation of a meta-analysis, is usu-
ally accompanied by a table listing references (author and date) of the studies in-
cluded in the meta-analysis and the statistical results (Centre for Evidence-Based
Intervention, n.d.).
An overall summary statistic combines and averages ESs across studies. An inves-
tigator should describe the method that determined the ES and should help the
reader interpret the statistic. Cohen’s (1988) methodology for determining ESs
includes the following strength of correlation ratings: trivial (ES = 0.01–0.09),
low to moderate (0.10–0.29), moderate to substantial (0.30–0.49), substantial to
very strong (0.50–0.69), very strong (0.70–0.89), and almost perfect (0.90–0.99).
