#MINITAB LLC SOFTWARE#
Usability of the software tools plays an essential role in the ability of researchers to plan studies with sufficient power to be informative. Thus, statistical software tools or packages programmed for estimating power are resources that can make power analysis easier to perform ( Thomas & Krebs, 1997). Estimating a priori power can be challenging even for experienced researchers ( Bakker, Hartgerink, Wicherts, & van der Maas, 2016). However, for more complex tests and most non-parametric statistics, such tables are not available (Thomas & Juanes, 1996). For very basic statistical tests, power can be calculated directly, and results can be presented in tables (e.g., Cohen, 1988). Many methods exist for calculating statistical power. Power analysis has a number of useful applications, such as optimizing limited resources when conducting research ( Lane & Hennes, 2019). Power analysis, then, broadly refers to a process of determining design characteristics (often sample size) for a planned study to optimize the likelihood of correctly rejecting the null hypothesis in that study, given a particular statistical model. Power is a function of the magnitude of an effect, variability, sample size, and Type I error level ( Cohen, 1992). One key strategy for increasing the likelihood of correctly rejecting a false null (i.e., power) is to perform an a priori power analysis. We illustrate use of the decision-aiding process with three example scenarios and lay out implications for researchers.Ī major aim of empirical research methodology is to design studies that are likely to reject the null hypothesis if it is indeed false. These criteria and heuristics are translated into questions that researchers should go through during the decision process. This process consists of five user-centered considerations – two functional criteria (software input software output) and three usability heuristics (learnability ease of use accessibility) – which are organized into a workflow.
Centering on users’ wants and needs, we propose a decision-aiding process to assist researchers in identifying appropriate power analysis tools. We review the emergence and current state of power analysis software tools, and use the concept of User-Centered Design as an analytical lens for examining the relationship between researchers’ needs and power analysis tools’ characteristics.
Given the complexity of conducting such analyses, software tools are essential. A priori power analyses have become increasingly popular in scientific communities, but the practice has not been widely discussed by HFE researchers.
0 kommentar(er)
