Comments From the Editor: Nobody Cares How You Analyzed Your Data

A few statements that are probably not overheard in the hallways at your school or institution: “I ran a multiple linear regression analysis with standardized beta coefficients! Rock on!” “My analysis included a 12 × 12 repeated measures analysis of variance with a sample size of 878,000 soprano saxophonists! They were really in tune!” But wait, there’s more. “I ran a t-test that compared two samples.” Well, you may have actually heard the last of these statements, but that test does not seem very complicated. A t-test? Isn’t that one of the easier analyses to compute? Heck, I can compute that by hand or even have Microsoft Excel do that. What about those fancy statistical software programs that I spent forever learning? Well, I hate to break it to you, but nobody cares how you analyzed your data. (Well, some people such as editors and editorial board review members care, and you probably should, but that’s not the purpose of these comments.) Selecting the right statistical test or analysis approach is critical to correctly understanding and interpreting your data. According to Nayak and Hazra (2011), “it is important that the appropriate statistical analysis is decided before starting the study, at the stage of planning itself” (p. 85). Although inferential statistics are not typically used in qualitative, historical, or philosophical research, the same planning processes extend to the types of analyses undertaken with these research methodologies. There are several excellent resources available for how to choose the appropriate statistical test or qualitative analysis approach (e.g., McDonald, 2014; Padgett, 2012).1 But selecting the right analysis before you begin a study is only part of the planning process. A critical component that surprisingly often gets overlooked is refining important and interesting research questions. Using a spiffy test or analysis may be fun, but should not be the primary factor when determining what type of data to collect or the rationale for why you chose to conduct a research study. A framework that I had not previously considered for crafting a good research question uses the acronym FINER—Feasible, Interesting, Novel, Ethical, and Relevant (Hulley, 2007). I know that not all—if any—of my research questions over the years have necessarily reflected all of these criteria. Nonetheless, I do believe this gives authors something to consider as they formulate their research questions. For example, let’s take one of the research questions that appears in this issue of Update. In his article, “High School Band Directors’ Perceptions and Applications of Democratic Rehearsal Procedures in Concert Band Rehearsals,” Scherer posits the following: How important do high school band directors believe it is for students to engage in democratic rehearsal procedures? Once you read the study, you’ll know that this was feasible, as he distributed a 10-minute questionnaire to a national sample of high school band directors. The topic is interesting because it involves a paradigm shift from the conductor-dominated approaches that have been used primarily and historically in large ensembles. In addition, the question is novel because we have scant information from high school band directors concerning their beliefs about democratic rehearsal procedures. The research is ethical—the appropriate permissions were granted, and respondents’ data was anonymous. Finally, the question is relevant as directors grapple with how to make ensembles more inviting and participatory for all music students. Before excitedly telling your colleagues or posting to your Twitter account that you read a manuscript that included a multiple analysis of variance with 17 dependent variables, you might ask yourself a few questions: Was that analysis necessary? Could the data have been analyzed in a more efficient manner? And before you answer those, perhaps the most important question: Were the research questions insightful to begin with? The use of the FINER framework is only one method that could prove helpful. Although not every research question you construct may be feasible, interesting, novel, ethical, or relevant, I do hope that you will consider the importance of your research questions and how they should be the primary factor when designing your study and selecting the appropriate analysis procedure. 1012798 UPDXXX10.1177/87551233211012798Update: Applications of Research in Music EducationSilvey editorial2021