I officially feel like a PhD student after undertaking this course. The "big picture" is starting to come together, if you will. A one-word summary of the course would be "bias." All research has bias, but the goal is to minimize sources of bias. I am more confident in my ability to critically review studies and identify sources of bias, and I now find myself asking important questions when reading journals or other sources of research outcomes. I feel confident in my ability to design a research study, and in fact had to do this for the final paper of the course. Here are some highlights of the course with some references.
What is the research question? The research question should be clearly defined and, at a minimum, identify the target population, intervention, comparison, and outcome (PICO). Clearly identifying inclusion and exclusion criteria, providing operational definitions of all variables, and clarifying the research hypothesis are important steps of this process.
What type of study? There is a continuum of study designs from descriptive to explanatory, and each study type has its own pros and cons, as well as statistical tools required. The course text is quite useful: Foundations of Clinical Research, 4th Ed. (Portney, 2020). Here are some helpful links I have found in addition to the course text that provide insight to various study designs:
Study Designs with more detail
Randomly controlled trials (RCT), or experimental research designs, are the foundation of explanatory research, and will most likely be the design of my dissertation. Several decisions must be made during the design phase: What are the variables, what will be measured and how, how will confounding be controlled? How will the population be recruited and sampled, and how will groups be assigned? How valid and reliable are the tools being used to measure outcomes, and who will be the testers? How will the data be managed and anonymity retained? Always look for sources of bias or error, and mitigate as much as possible.
How many participants will be required? A review of the literature can help identify common effect sizes in similar studies, but if none are available the standard from Cohen (1992) are: 0.2 - small, 0.5 - medium, 0.8 - large. Using the desired effect size and appropriate statistical analysis, an a priori analysis with G*Power can provide the necessary number of participants (Faul et al., 2009). A quick YouTube search of "G*Power examples" will return several helpful videos, including this primary course-used lecture G*Power How To, and the software is a free download here: G*Power Download. A useful paper on choosing a sample size can be found here as well: Sample Sizes in Health Research
How will the data be analyzed? This will be decided by the study type, type of data, number of groups, and number of variables. It is a topic unto its own in coursework, the next course I am taking this summer in fact. In the meantime, here is an informative article on choosing the appropriate statistical test: Choosing the Correct Statistical Test
How will the research be reported? Various tools for analyzing research reports were introduced in the course. One that stood out was the Consolidated Standards of Reporting Trials (CONSORT), which has been recently updated in May 2025. The CONSORT offers a checklist for publishing research with a primary goal of being transparent and thorough in reporting experimental methodology and results. Here is the updated link: CONSORT 2025. Other tools that can be used depending on the type of research are:
1. Grading of Recommendations Assessment, Development and Evaluation (GRADE)
2. Critical Appraisal Skills Programme (CASP)
3. JBI Critical Appraisal Tools
4. Physiotherapy Evidence Database Scale (PEDro Scale)
5. Tool for the Assessment of Study Quality and Reporting in Exercise (TESTEX)
What's next? I have courses in quantitative and qualitative analysis upcoming. My background in mathematics has prepared me quite well for the statistical components, and I'm really looking forward to clarifying the use of proper statistical measures in context of applied research. I am also adding to my list of research interests (see previous post) as ideas arise during coursework or in my practice of coaching.
References
Cohen, J.
(1992). A power primer. Psychological Bulletin, 112(1), 155–159.
Faul, F., Erdfelder, E., Buchner, A., & Lang, A.-G. (2009). Statistical power analyses using G*Power 3.1: Tests for correlation and regression analyses. Behavior Research Methods, 41, 1149-1160.
Portney,
L.G. (2020). Foundations of clinical research (4th ed). FA Davis.
No comments:
Post a Comment