In previous iterations of this series, I’ve given advice about writing the results and discussion sections of your manuscript for submission to the Journal of Public Health Management and Practice (JPHMP). Arguably, I’ve saved the most important section for the third in the series, as the methods section of your manuscript contains the material that most people will use to judge the quality of your article. As an editor, I typically skip the introduction (initially) when screening articles and go to the methods. If the methods are questionable, I may not even bother to read the results, as I don’t want to get excited about the results of a study that were derived from methods of poor quality. Those types of studies tend to generate a lot of headlines and debate but can be the target of skeptics who pore over the methods (and disclosures) carefully. But I digress.
Fortunately, there are a number of guidelines available to authors to assist in deciding what information is relevant to the reader and how it should be presented. The Equator Network presents a comprehensive list of reporting guidelines that cover most of the article types that JPHMP publishes (though our format for Case Studies more resembles the business case study format than the clinical case report format). Specifically, CONSORT, STROBE, PRISMA, and SRQR inform reporting of randomized trials, observational studies, systematic reviews, and qualitative research studies, respectively. CONSORT, STROBE, and PRISMA have a number of useful “extensions” that address specific types of articles that may deviate from the standard format. For this post, I’ll discuss the methods section more generally (and focus on quantitative methods), but I strongly recommend spending some time on each of the reporting guideline websites to explore the resources available since many of the available checklists can be amazingly helpful to writers of any skill level. Below, I briefly describe relevant components of a methods section with special relevance for authors submitting to the JPHMP.
- Study design – Early in the Methods, you should specify the design of the study as clearly as possible with the appreciation that much practice-based work doesn’t fall into neat categories such as randomized-controlled trial, cohort, case-control studies, time-series, or cross-sectional designs. For complex or non-traditional designs, you might consider supplemental figures or tables that detail who and when participants were recruited, measured, and intervened upon (if relevant). This is a good place to present information about ethical oversight of the study (ie, IRB review
- Setting – Basically, the setting is where the study occurred. If space allows, it’s good practice to describe the location(s) as they related to the generalizability of the results. For example, it’s one thing to say that an intervention was conducted in a middle school in the southwestern United States. It’s another to say that it was conducted in a middle school in the southwestern United States that served predominately low-income students from a rural community whose student body was 53% American Indian.
- Participants – Here you want to describe the eligible population, eligibility criteria, and methods for selecting the sample. You’ll want to be careful not to use words like “a random sample” when you really have “a convenience sample.” Indicate any exclusion criteria that impact the final sample size (to be reported in the results section).
- Data source or measurement – Clearly indicate which data were collected for the current study and how those data were captured. Questionnaires or scales should be clearly described with a brief explanation of the available evidence for their validity and reliability or a citation that presents that information if available. Consider making the scale or data extraction instrument available as supplemental content if the psychometric properties haven’t been previously established.
- Quantification of variables – It’s important that you clearly define all exposures (in intervention studies), predictors, control variables, and potential confounders. Ideally, you should operationalize your variables here. For example, don’t simply say that you “controlled for physical activity” when you report collecting data on numerous physical activity related behaviors. It would be more helpful to explicitly state how you’ve operationalized the variable for the analysis: “we controlled for physical activity by including a variable indicating those currently active (1 = Active) as defined as reporting five or more days of 30 minutes of moderate-to-vigorous physical activity in the last seven days.”
- Statistical methods – When presenting statistical methods, it can be a challenge to balance between presenting enough information to ensure that your analyses could be reproduced and not presenting it in a manner that confuses the non-statistically oriented reader. Typically for JPHMP, we suggest avoiding the presentation of formulas unless a new analytical technique is being presented. However, it’s important that the information necessary for the interpretation of the results is presented. Generally speaking, authors should present the power analysis (if applicable), data cleaning, handling of missing data, data transformations (if any), statistical techniques, and post-hoc comparisons to be presented.
About the Author
- Justin B. Moore, PhD, MS, FACSM, is a Professor and Vice Chair in the Department of Implementation Science in the Division of Public Health Sciences at Wake Forest University School of Medicine. He serves as the Editor-in-Chief of the Journal of Public Health Management and Practice.
Latest entries
