The critical insights of this chapter revolve around aspects of constructing and maintaining the statistical and practical significance of the study and the importance of meta-analytical designs. Researchers often employ analysis methods that demonstrate the statistical significance of policy, yet not every study is able to demonstrate it. For instance, the use of percentile difference is not always a valid measure of the significance of change as not every measure can have a true zero. Thus, in order to accurately present and demonstrate the study, researchers are required to carefully plan the measures and coefficients they are going to use. In this endeavor, an effect size statistic could help rate the results that lack a true zero, thus helping to ascertain the program’s magnitude (Rossi, Lipsey, & Freeman, 2003). An odds ratio or standardized mean difference is also a crucial tool that assists in assessing and demonstrating statistical significance by allowing to compare results through min-max spread or the change in the chance of an occurrence.
A key insight about statistical significance gained from this chapter is that it cannot be used as a measure of importance. It is, however, useful in establishing the fact of correlation or dependency in variables, thus eliminating the notion of chance. Conversely, if statistical significance is not registered, it is not indicative of the program as not effective. It is merely a demonstration of a study’s design weakness (Rossi et al., 2003). Thus, the successful establishment of this indicator could be relevant mostly as a premise for further discussion of the program’s significance in relation to measured outcomes. The practical importance of this complex of assessments is strong as it allows for finding more correlations between variables and using them to construct more knowledge about the program, uncovering possibly unexplored areas relative to its effects on the studied population.
Practical significance is seldom uncovered in studies but is more often indicated indirectly. This factor of research is especially relevant in health care or law enforcement policies, as the allocation of financing needs to be guided by the practical use of a new procedure or method for the society since in both areas, lives are often at stake (Kang, Yeon, & Han, 2015). According to Rossi et al. (2003), the practical relevance of the uncovered data is substantially dependent on the careful interpretation of results. In light of this, they suggest that to ascertain such relevance, it is often required to cross-reference the data with certain external measures of quality or standards of a higher level, such as national norms or averages.
One more critical insight of the chapter is connected to meta-analytical studies. They are mostly used to summarize the previous knowledge and derive the cumulative wisdom acquired over the years of research (Hedges & Olkin, 2014). Nowadays, major medical policies need to be warranted by sufficient evidence, which is established through meta-analytical studies. A crucial point to consider in such a case is the choice of studies which requires the meticulous planning of selection criteria.
References
Rossi, P. H., Lipsey, M. W., & Freeman, H. E. (2003). Evaluation: A systematic approach (7th ed.). Thousand Oaks, CA: SAGE Publications, Inc.
Hedges, L. V., & Olkin, I. (1985). Statistical methods for meta-analysis. New York, NY: Academic Press.
Kang, H., Yeon, K., & Han, S.-T. (2015). A review on the use of effect size in nursing research. Journal of Korean Academy of Nursing, 45(5), 641-649.