Enhancing Nurse Practitioner Research through Better Derived Effect Sizes

The strength of the results and implications of nursing research reside initially in the design of the research plan. Providing inadequate attention to rigor in the design phase of research can easily lead to a poor design, and therefore results which may not be that useful. Nurse practitioner faculty that embark on quantitative research activities must pay careful attention to effect sizes in the planning phase of their research.  However statistical software options and how to use existing literature for health promotion and health outcomes research has led to confusion over determination of appropriate effect sizes. The purpose of this session is to discuss timely issues surrounding the use of effect sizes.

The dependence on probability statements in academic psychology has hampered the field’s ability to make progress. The reasons for this are many, but the underlying problem is that probability statements are constructed from two sources of information about an experiment or study—the effect size and the sample size.  As both effect sizes and sample sizes increase, the probability that the result is due to chance diminishes.  Of these two pieces of information, the one that has the most impact on how well a patient does is the effect size. 

This session will discuss how effect sizes are defined. It will also address how effect sizes are currently used in research for estimating statistical power and conducting meta-analyses. The current use of effect sizes for estimating power is a relatively straightforward process, but the use of effect sizes in other contexts can be more complex.  The basis on how statistics are used with effect sizes and how they should be applied to addressing research issues will be given. 

A theoretical model for how effect sizes can be used in nursing research to develop new and more effective interventions, better describe the cost and benefits associated with current interventions and treatments, as well as can be used in a Bayesian decision-making context to generate optimal patient-centered treatments, will be presented. By addressing effect sizes, the impact of important research problems that nurse practitioner faculty study, can be optimized.