You have a research idea that survived the novelty check. The next question is whether you can actually pull it off. Feasibility is what separates students who publish from students who spend a year on a project that quietly stalls in data extraction. The good news is that feasibility is mostly arithmetic. The bad news is that nobody teaches the arithmetic explicitly. Here is the practical version.

Five dimensions of feasibility

1. Data availability

For a systematic review or meta-analysis, "data" means the right number of published studies that meet your inclusion criteria. For a primary study, "data" means access to patient records, biospecimens, or recruitable participants. Both fail in predictable ways.

For an SR/MA: Run your search across at least two databases as a scoping exercise. If you get fewer than 8 to 10 likely-includable studies, your meta-analysis will be statistically underpowered and your reviewers will say so. If you get more than 200, screening becomes an 8-week marathon and you may need to narrow your question. The sweet spot for a first review is 20 to 60 candidate studies after title/abstract screening, roughly 8 to 20 included.

For a primary chart review: Estimate by writing the inclusion criteria as ICD-10 codes and asking your hospital data team for a count of records meeting them in the last 5 years. If they say "350 patients," you have a feasible cohort. If they say "8 patients," you do not.

2. Sample size and power

For any primary study with a hypothesis test, calculate the required sample size before you start. The four inputs are: alpha (typically 0.05), power (typically 0.80), effect size (from prior literature or a clinically meaningful threshold), and the standard deviation or event rate.

Use G*Power for individual tests or the pwr package in R for trial-level calculations. If your calculation says you need 1,200 patients per arm and your hospital sees 40 patients with this condition per year, the study is not feasible in your timeframe. Pivot.

Common student mistake: using the effect size observed in a previous small study as if it were the true effect. Small studies systematically overestimate effects (the "winner's curse"). Plan for a smaller true effect than the literature suggests.

3. Time budget

The single biggest cause of stalled student projects is timeline underestimation. Realistic budgets:

  • Systematic review: 6 to 12 months end to end for a first-time team. Step-by-step breakdown here.
  • Retrospective chart review: 4 to 9 months from IRB approval to manuscript draft, depending on cohort size and data extraction complexity.
  • Cross-sectional survey: 3 to 8 months including survey design, IRB, recruitment, analysis, write-up. Slower if recruiting outside your institution.
  • Prospective interventional pilot: 12 to 24 months minimum.
  • Secondary analysis of public data (NHANES, MIMIC, SEER): 3 to 6 months. The fastest path to first authorship for a student with statistical skill.

If you have 5 months until residency applications close, a prospective interventional study is not going to be published in time. A secondary analysis or scoping review might.

4. IRB and ethical clearance

Three risk tiers at most US institutions:

  • Exempt (Category 4 typically): retrospective use of de-identified data, anonymous survey of non-vulnerable adults. Usually approved in 1 to 3 weeks.
  • Expedited: minimal-risk research with identifiable data, retrospective chart reviews with a HIPAA waiver, low-risk surveys. 2 to 6 weeks.
  • Full board: prospective interventional research, vulnerable populations (children, pregnant, prisoners, cognitively impaired), greater than minimal risk. 6 to 16 weeks.

Build your timeline around the IRB tier. If your idea requires full board review and you have not started the application, add 3 months to your earliest publication date.

SR/MA and existing-literature analyses generally do not require IRB review. Confirm with your institution.

5. Skill and resource budget

An honest skills inventory before you commit:

  • Can someone on the team run the analysis you plan? (R, Stata, SPSS, Python)
  • Do you have access to a statistician for at least a 1-hour consult? Most universities offer this free to students.
  • Do you have institutional access to the databases you need (Embase, Scopus)?
  • Do you have time to learn one new tool (Rayyan, Covidence, RevMan) without falling behind?
  • Is your supervisor responsive enough to give feedback within 2 weeks on drafts?

If three of these are "no," the project will stall. Either pivot to a project that fits your skills or invest in upskilling before you commit.

A feasibility worksheet

Fill in before committing:

  • Estimated number of studies/patients/records meeting inclusion: ___
  • Sample size required for adequate power: ___
  • Months available before submission target: ___
  • IRB tier: Exempt / Expedited / Full board
  • Estimated months to IRB approval: ___
  • Team member responsible for analysis: ___
  • Statistical method: ___
  • Database access confirmed: Y / N
  • Supervisor feedback cadence: ___
  • Three biggest risks to completion: ___

If the answer to any of the first six items is "I don't know," do the scoping research before you commit.

When to abandon

If three months in, your literature search has dried up, your IRB has rejected the protocol twice, and your supervisor stopped replying, the project is not feasible in its current form. Pivoting early is not failure — it is the only sensible response. The students who suffer most are the ones who push through a clearly infeasible project for sunk-cost reasons. Cut losses, salvage what you have learned, and pick a project sized to your real time and skill budget.

If you are using ResearchChecker, the Methodological Feasibility section gives you a per-design verdict (Feasible, Challenging, Unlikely, Ideal) based on the actual evidence landscape. Use it as a starting point, then run the arithmetic above for the specific design you pick.