Research processes: Question diagram

Question diagram template is an explicit structure that facilitates the design of research questions in a secondary data analysis study [1]. It is a structured template that is based on backward design approach [2]. It is usually prepared in a Google document in collaboration with research team members (principal investigator, statistician, data collection team, and data analysis team).

As a first step, research team members should brainstorm and identify a few impactful and innovative research questions (based on available data, methodological expertise and real world experience) that they could answer using available data. Identifying a central metaphor related to the research questions facilitates impactful communication of the research results later on.

The question diagram makes hypothesized conclusions explicitly clear and specific at the very beginning of the research project. These should then be evaluated by conducting literature search using FINER (feasibility, interesting, novelty, ethical, relevant) criteria [3]. PICO framework (population, intervention, comparison, outcome) [4], SMART goal framework (specific, measurable, achievable, relevant, time bound) [5] and FAST (frequently, discussed, ambitious, specific, transparent) goals framework [6] may also be used for this purpose. Documentation of this literature research is maintained in the literature matrix [1]. Notably, hypothesized conclusions should be dynamic - i.e. they should be a part of the research team's gaze rather than goals. Further hypothesized conclusions map to and provide content for the objective section of the final research manuscript.

As a next step, hypothesized conclusions are then mapped to corresponding variables (outcomes, predictors, confounders) that are essential to answer the research question. The real world workflow essential to specifically and accurately collect data related to these variables should be broadly defined. Determining the type of variables (categorical, continuous) facilitates the choice of applicable statistical tests. The variables, type of variable and statistical tests also map to and provide content for the methods section and data analysis sub- section of the final research manuscript.

Next, strata (homogeneous, non- overlapping sub-groups within larger population) are defined. They are used to analyze differences between subgroups which might otherwise be hidden in aggregate data. "By using strata, data analysis moves beyond surface level observations to uncover nuanced trends and makes findings more valid and generalizable to specific populations" (Strata, Google Gemini).

Next, the inclusion and exclusion criteria essential to collect data corresponding to the variables type identified earlier are explicitly defined and mapped.

Next, preparing mock tables and figures to be used in the final research manuscript facilitates visualization of the results. Aligning with prevailing visualization trends ensures that the research team is well equipped to convey their message impactfully to the end readers.

Finally, reviewing and verifying alignment of hypothesized conclusions, variables, and headers of mock tables/figures ensures appropriate and impactful communication of the research results. This final step is termed as Question diagram axis check. The QD axis and metaphors are the quantitative and qualitative themes of the final research manuscript.

References:

1. Pietrobon R, Guller U, Martins H, Menezes AP, Higgins LD, Jacobs DO. A suite of web applications to streamline the interdisciplinary collaboration in secondary data analyses. BMC Med Res Methodol. 2004 Dec 14;4(1):29. doi: 10.1186/1471-2288-4-29. PMID: 15596017; PMCID: PMC544191.

2. Wiggins, G., and McTighe, J. (1998). Understanding by Design. ASCD.

3. Hulley SB, Cummings SR, Browner WS, Grady DG, Newman TB. Designing clinical research. 3rd ed. Philadelphia (PA): Lippincott Williams and Wilkins; 2007.

4. Richardson WS, Wilson MC, Nishikawa J, Hayward RS. The well-built clinical question: a key to evidence-based decisions. ACP J Club. 1995 Nov-Dec;123(3):A12-3. PMID: 7582737.

5. Doran, G. T. (1981). "There's a S.M.A.R.T. Way to Write Management's Goals and Objectives", Management Review, Vol. 70, Issue 11, pp. 35-36.
6. Sull, D., & Sull, C. (2018). With Goals, FAST Beats SMART. MIT Sloan Management Review.