QuickStats Healthan Ayom Healthcare venture
Research Methodology

Common Mistakes in Medical Research Methodology

Ten frequent methodology mistakes in student medical research, from vague objectives to biased sampling, and how to avoid each one before you collect data.

By the QuickStats Health editorial teamUpdated 10 November 20264 min readLevel: Intermediate

Statistical analysis cannot rescue a study with weak methods. Most problems examiners and reviewers find were built in at the planning stage. Here are ten common methodology mistakes, and how to avoid them while you can still change your plan.

1. Objectives that cannot be measured

"To study the clinical profile of patients" does not tell anyone what you will measure. Write objectives with a clear outcome: "To estimate the proportion of patients with diabetic retinopathy among people with type 2 diabetes attending the medicine OPD."

2. A design that does not match the objective

Incidence needs follow-up (a cohort). Prevalence fits a cross-sectional design. Rare outcomes suit case-control studies. Check that your design can actually answer each objective. Our guide to observational study designs explains the differences.

3. No justified sample size

A number like "100 patients" with no formula or reference is a common reason for revisions. Calculate the sample size from a published estimate, state the formula and add an allowance for non-response.

4. Convenience sampling described as random

Taking every eligible patient who comes to the OPD is consecutive sampling, not random sampling. Describe what you actually do. Consecutive sampling is acceptable for many hospital studies, but say so honestly and discuss its limits.

5. Vague inclusion and exclusion criteria

"Patients with hypertension" needs a definition: which blood pressure cut-off, on how many readings, with or without treatment? Exclusions such as "serious illness" should also be defined.

6. Unvalidated or home-made tools

Wherever possible, use a validated questionnaire or a standard clinical definition. If you must create a tool, describe how it was developed, have experts review it and pilot it on a small group first.

7. No plan to reduce bias

Think about selection bias (who gets into your study), information bias (how you measure) and confounding (other factors that explain an association). Simple steps help: standard measurement methods, trained observers and recording known confounders such as age and sex.

8. Ignoring missing data

Decide in advance how you will handle participants who drop out or leave questions blank. Report how many were missing rather than quietly removing them.

9. A vague statistical analysis plan

"Appropriate tests will be applied" is not a plan. Name the test you expect to use for each objective, the software and your significance level.

10. Ethics as an afterthought

Informed consent, confidentiality and IEC approval must be in place before data collection. Changing the protocol after approval usually needs another submission to the ethics committee.

Summary table

Mistake Simple fix
Unmeasurable objectives Name the outcome and how it is measured
Design–objective mismatch Choose the design from the objective
Unjustified sample size Formula + published estimate + non-response allowance
Mislabelled sampling Describe exactly how participants are selected
Vague criteria Define every inclusion and exclusion term
Unvalidated tools Use validated tools or pilot your own
Bias not considered Standardise measurement, record confounders
Missing data ignored Plan and report how missing data are handled
Vague analysis plan Name a test for each objective
Late ethics Get IEC approval before data collection

Key takeaways

  • Methodology problems are cheapest to fix before data collection.
  • Every objective should map to a design, a measurement and a planned analysis.
  • Be honest about sampling and limitations; examiners respect clarity.
  • Reporting guidelines such as STROBE and CONSORT are useful checklists while planning, not only while writing.

References

  1. Grimes DA, Schulz KF. An overview of clinical research: the lay of the land. Lancet. 2002;359(9300):57–61.
  2. Vandenbroucke JP, von Elm E, Altman DG, et al. Strengthening the Reporting of Observational Studies in Epidemiology (STROBE): explanation and elaboration. PLoS Med. 2007;4(10):e297.
  3. Hopewell S, Chan AW, Collins GS, et al. CONSORT 2025 statement: updated guideline for reporting randomised trials. BMJ. 2025. PubMed
  4. Hulley SB, Cummings SR, Browner WS, Grady DG, Newman TB. Designing Clinical Research. 4th ed. Philadelphia: Lippincott Williams & Wilkins; 2013.

This article is for education. Always follow the requirements of your university, guide, ethics committee and target journal.