Moving Nursing Beyond p < .05

Abstract

We write as a community of professional statisticians and quantitative methodologistswith extensive collective experience in nursing research. Our group includes faculty in academic nursing programs, statistics reviewers for nursing research journals, statistics educators who teach nursing students, and statistics collaborators on nursing research studies. Members of our group have participated in expert panel discussions and presentations at international statistics conferences about the use of statistics in nursing research and education (Hayat, Eckardt, Higgins, Kim, & Schmiege, 2013; Hayat, Higgins, Schwartz, & Staggs, 2014). Our efforts are further described in the opening editorial for a special issue of Nursing Research devoted to statistics in nursing (Hayat, 2012). Since 2011 we have continued to maintain an email listserv for statisticians in nursing. Over the years, members of our group have routinely encountered, and tried to address, misuses, and misunderstandings of p-values and significance testing. Fortunately, the American Statistical Association (ASA) recently launched a largescale effort aimed at “Moving to a World Beyond ‘p < .05”’ (Wasserstein, Schirm, & Lazar, 2019), publishing a 19-page editorial with this title in the ASA-sponsored journal The American Statistician, along with 43 thought-provoking papers from prominent statisticians and other experts on the topic. Wasserstein et al. (2019) call for abandoning the phrase “statistically significant” and discontinuing the practice of categorizing p-values based on an arbitrary threshold such as .05, noting that, “Regardless of whether it was ever useful, a declaration of‘statistical significance’ has today become meaningless.” It is unfortunate that the term “significant” was ever attached to “p < .05.” Webster’s dictionary defines the word “significance” as “importance.” Yet, statistical significance is not synonymous with importance. This misleading label has led to misinterpretations and poor decisions that have cost lives, money, and resources (Ziliak & McCloskey, 2008). Patients and healthcare providers have paid a heavy price for the science community’s reliance on statistical significance as a criterion for importance. Decisions should never be made based solely on a significance