Launch StatClinic →
Statistical Tests

Statistical Assumptions in Medical Research: What Researchers Need to Check Before Analysis

📖 18 min read 🗓 July 2026 ✓ Updated July 2026
S
StatClinic Editorial Team Statistical content for medical researchers and clinicians
Every statistical test — every t-test, every ANOVA, every regression model, every chi-square table — is built on a set of mathematical conditions it assumes your data already satisfies. The test does not check these conditions for you, and it will not warn you when they are broken. It will simply run, produce a p-value, and hand you a result that looks exactly as confident whether the underlying assumptions held or collapsed entirely. This guide walks through the assumptions that matter most in medical research — independence, normality, homogeneity of variance, linearity, outliers, multicollinearity, sample size and expected cell counts, and the proportional hazards assumption — with a clinical example for each, so you know exactly what to check before you trust your own output.
Key Takeaways
  • A statistical test's output is only as trustworthy as the assumptions behind it — the software will never warn you when an assumption is violated; it will simply produce a number.
  • Independence of observations must be verified before any other assumption — repeated measures, matched pairs, and clustered data (patients within hospitals) all violate it and require different methods entirely.
  • Normality applies to the continuous outcome (or regression residuals), not to every variable in your dataset, and matters most in smaller samples.
  • Different tests carry different assumption profiles — Cox regression needs the proportional hazards assumption checked; chi-square needs expected cell counts checked; neither of these applies to a t-test.
  • Checking assumptions is part of running the analysis, not an optional validation step performed only if a reviewer asks.

Why Checking Assumptions Matters Before You Trust a P-Value

Every parametric statistical test is derived mathematically under a specific set of conditions — a particular distribution shape, a particular independence structure, a particular relationship between variables. The formulas that produce your p-value, your confidence interval, and your effect estimate are only valid proofs under those conditions. When the conditions don't hold, the formula still runs to completion and still returns a number, but that number is no longer the quantity the test was designed to estimate.

This is different from most errors in a data pipeline, which tend to announce themselves — a missing file throws an error, a wrong data type throws a warning. A violated statistical assumption does neither. SPSS, R, and every other statistical package will compute a t-test on badly non-independent data and hand back a p-value formatted identically to a p-value computed on perfectly valid data. The distinction between a trustworthy and an untrustworthy result exists only in whether the researcher checked the assumptions first — nothing about the output itself reveals which one you're looking at.

This is also why assumption-checking sits logically after study design and variable classification, but strictly before interpreting any test result — it is one of the last checks standing between your analysis and a conclusion you can actually defend to a thesis committee or peer reviewer.

Take-Home Points A statistical test cannot tell you whether its own assumptions are satisfied — that verification is entirely the researcher's responsibility, and it has to happen before the p-value is interpreted, not after.

Independence of Observations

Independence means that each observation in your dataset provides genuinely separate information — knowing one patient's value should tell you nothing about another patient's value. Most standard tests (independent t-test, one-way ANOVA, chi-square, standard linear and logistic regression) assume this by default, and it is arguably the single most consequential assumption to check first, because violating it doesn't just bias one number — it invalidates the entire mathematical basis for the standard error calculation.

Independence is most commonly broken by repeated measurements on the same patient (baseline, 4 weeks, 8 weeks), by matched or paired designs, and by clustered data, where patients treated by the same surgeon, admitted to the same ward, or enrolled at the same trial site tend to be more similar to each other than to patients elsewhere — a structure common enough in multi-center trials that it has its own name, the design effect.

Clinical Example

Measuring pain scores in 30 patients before and after a nerve block and analyzing all 60 values with an independent t-test, as if they came from 60 separate patients, ignores that each pair of measurements comes from the same person — a paired t-test is required instead.

Take-Home Points When observations are clustered or repeated, use a method built for that structure — paired tests, repeated measures ANOVA, mixed-effects models, or GEE — rather than a standard test that silently assumes independence it doesn't have.

Normality

The normality assumption applies to the distribution of a continuous outcome variable (or, in regression, to the model's residuals) and underlies parametric tests including the t-test, ANOVA, Pearson correlation, and linear regression. It is formally checked with the Shapiro-Wilk test in small-to-moderate samples, and assessed visually with a histogram or Q-Q plot, as covered in full in our normality testing guide.

Many medical measurements are naturally skewed rather than normal — length of hospital stay, C-reactive protein, and most biomarker concentrations tend to have a long right tail, with a small number of very high values pulling the mean upward. Reporting mean ± SD and running a t-test on data like this can be misleading; median and interquartile range, alongside a non-parametric test, is usually the more honest summary.

Clinical Example

Comparing ICU length of stay between two treatment groups: a handful of very long stays make the distribution right-skewed, so the median (IQR) and a Mann-Whitney U test better represent the typical patient than the mean and a t-test would.

Homogeneity of Variance (Homoscedasticity)

Homogeneity of variance means the spread (variance) of the outcome is similar across the groups being compared, or, in regression, similar across the range of predicted values. It applies to the independent t-test, one-way ANOVA, and the residuals of a linear regression model, and is checked formally with Levene's test or visually with a residual-versus-fitted plot.

Unequal variances distort the standard error used to calculate the p-value, and the distortion is worse when group sample sizes are also unequal — a common scenario in medical research when a rare-exposure or rare-outcome group is naturally much smaller than the comparison group.

Clinical Example

Comparing a biomarker's variability between healthy controls (a tight, consistent range) and a disease group (a much wider range, reflecting disease heterogeneity) — Levene's test flags unequal variances, and Welch's t-test, which does not assume equal variances, is the more defensible choice.

Take-Home Points Most statistical software reports both the standard and the variance-corrected (Welch) version of a t-test or ANOVA by default — checking Levene's test result tells you which row of output to actually report.

Linearity

Linearity is the assumption that the relationship between a continuous predictor and the outcome follows a straight line — for Pearson correlation and linear regression, this means the outcome itself; for logistic regression, it means the log-odds of the outcome; for Cox regression, it means the log-hazard. It is checked with a scatterplot of predictor against outcome, or with a residual-versus-predictor plot in a fitted model.

Many physiological relationships are not linear across their full range. Age and a given biomarker, for example, may rise, plateau, and fall again across a lifespan, producing a curved or even U-shaped true relationship. Forcing a straight line through a genuinely curved relationship can produce a near-zero, non-significant slope even when a real, strong, non-linear association exists — a false negative created entirely by an unchecked assumption, not by an absence of true effect.

Clinical Example

Modeling the relationship between maternal age and risk of chromosomal abnormality as a straight line would badly misrepresent a relationship that is actually roughly flat through the 20s and 30s before rising sharply after 35 — a quadratic term or a spline captures this shape far better than a single linear slope.

Outliers and Influential Points

An outlier is a data point with an extreme value relative to the rest of the dataset. An influential point is a data point whose removal would meaningfully change the model's results — not every outlier is influential, and not every influential point looks extreme at first glance; formal detection in regression uses measures like Cook's distance rather than eyeballing the raw values alone.

The instinct to simply delete outliers is usually the wrong first move. An extreme value can be a genuine, clinically important observation (a real, unusually severe case), a data entry error (a decimal point in the wrong place), or evidence the point belongs to a different population entirely (a lab value from a sample that was actually hemolyzed). The correct first step is always to investigate the cause, not to remove the point and move on.

Clinical Example

One patient in a chronic kidney disease cohort has a creatinine level far above the rest of the group due to a documented episode of acute kidney injury — this is a genuine, clinically meaningful value, not an error, and the appropriate response is a sensitivity analysis reporting results with and without that patient, rather than silent deletion.

Caution Never remove an outlier purely because it makes your p-value non-significant. Document the investigation, report results both with and without the point if its status is genuinely ambiguous, and let a reviewer see the reasoning.

Multicollinearity

Multicollinearity occurs in multiple regression (linear or logistic) when two or more predictor variables are highly correlated with each other. The overall model can still fit reasonably well, but the individual coefficient estimates become unstable — standard errors inflate, confidence intervals widen, and which of the correlated predictors appears "significant" can flip unpredictably with small changes to the data or model specification.

Multicollinearity is detected using the Variance Inflation Factor (VIF) for each predictor; a VIF near 1 indicates no problematic correlation, while a VIF above 5, and especially above 10, signals a predictor whose effect estimate can no longer be trusted in isolation. A simple correlation matrix among candidate predictors before modeling is a useful, faster first screen.

Clinical Example

Including both BMI and waist circumference as separate predictors of cardiovascular risk in the same regression model — the two measures are highly correlated, and their individual coefficients become unreliable; choosing one, or combining them into a single composite measure, produces a more interpretable model.

Adequate Sample Size and Expected Cell Counts

Beyond the assumptions specific to individual tests, every statistical method also assumes the sample is large enough for its underlying approximation to hold. An underpowered study risks a Type II error regardless of how carefully every other assumption was checked — a topic covered fully in our sample size and power guide.

The chi-square test has its own specific version of this assumption: no more than 20% of cells in the contingency table should have an expected count below 5, and no cell should have an expected count below 1. When this is violated — common with rare outcomes or small subgroups — Fisher's Exact test should be used instead, since it calculates exact probabilities rather than relying on chi-square's large-sample approximation. Multivariable logistic regression carries a related rule of thumb, commonly cited as at least 10 outcome events per predictor variable included in the model.

Clinical Example

A 2×2 table comparing a rare surgical complication (only 3 events in one arm) between two procedures has expected cell counts below 5 — chi-square is not appropriate here; Fisher's Exact test is the correct choice.

The Proportional Hazards Assumption

The proportional hazards assumption is specific to Cox regression: it requires that the hazard ratio comparing any two groups (or any two levels of a covariate) stays constant across the entire follow-up period. Visually, this means the survival curves for compared groups should never cross, and log-minus-log survival plots for each group should run roughly parallel throughout follow-up.

The formal check uses a test based on Schoenfeld residuals, which examines whether each covariate's residuals show a significant relationship with time — a significant result flags a violation. When the assumption is violated, a single hazard ratio no longer meaningfully summarizes the relationship, since the true risk difference between groups is actually changing across follow-up; the fix is a stratified Cox model, a covariate-by-time interaction term, or reporting time-varying coefficients instead of a single pooled hazard ratio.

Clinical Example

Comparing surgical versus medical management for a condition: surgery may carry a higher early risk from procedural complications but a lower risk later from durable treatment benefit — the hazard curves cross, proportional hazards is violated, and a single reported "HR = 0.8" would misrepresent a relationship that actually reverses direction over time.

Take-Home Points Different tests carry entirely different assumption profiles — proportional hazards has no meaning for a t-test, and normality has no meaning for a chi-square table. Match the checklist to the specific test you're running, not a generic list applied to every analysis.

Which Assumptions Apply to Which Test

Not every assumption applies to every method — this table maps the assumptions covered above to the tests they actually govern, as a quick reference before you run your analysis.

AssumptionApplies To
Independence of observationsNearly all standard tests — t-test, ANOVA, chi-square, standard regression
Normalityt-test, ANOVA, Pearson correlation, linear regression (residuals)
Homogeneity of varianceIndependent t-test, one-way ANOVA, linear regression
LinearityPearson correlation, linear regression, logistic regression (logit scale), Cox regression (log-hazard scale)
Outliers / influential pointsCorrelation, regression, and any mean-based comparison
MulticollinearityMultiple linear regression, multiple logistic regression
Adequate sample size / expected cell countsChi-square test, logistic regression
Proportional hazardsCox regression
❌ Incorrect

Checking only normality for every analysis, regardless of which test is actually being run, and assuming a "normal-looking" histogram means the analysis is assumption-safe.

✅ Correct

Identify the specific test first, then check only the assumptions that test actually depends on — a chi-square table needs expected cell counts checked, not normality; a Cox model needs proportional hazards checked, not homogeneity of variance.

Common Mistakes When Checking Assumptions

Mistake 1: Skipping Normality Because "n Is Large Enough"

Treating the Central Limit Theorem as a blanket excuse to skip checking normality entirely, even when a subgroup within the larger sample is small or the data are severely skewed.

✓ Fix: Test formally regardless of sample size, and pay particular attention to smaller subgroups within a larger dataset.

Mistake 2: Ignoring Clustering or Repeated Measures

Analyzing repeated or clustered observations as if each row in the dataset were an independent patient, artificially shrinking standard errors and inflating false-positive risk.

✓ Fix: Identify the dependency structure before choosing a test — paired, repeated-measures, mixed-effects, or GEE methods exist specifically for this.

Mistake 3: Deleting Outliers Without Investigation

Removing an extreme value simply because it looks unusual or because it changes the p-value, without first checking whether it's a data error, a genuine case, or a different population.

✓ Fix: Investigate the cause first; use sensitivity analysis (with and without the point) if its status remains genuinely ambiguous.

Mistake 4: Reporting a Single Hazard Ratio Without Checking Proportional Hazards

Running a Cox model and reporting the hazard ratio without ever testing whether it's actually constant across follow-up time.

✓ Fix: Check Schoenfeld residuals or log-minus-log plots before finalizing any Cox regression result.

Mistake 5: Running Chi-Square on Small Expected Cell Counts

Applying a standard chi-square test to a contingency table with several cells expected to contain fewer than 5 observations, silently violating the test's large-sample approximation.

✓ Fix: Check expected cell counts before running the test; switch to Fisher's Exact test when the 20% rule is violated.

Mistake 6: Interpreting Individual Coefficients Under Severe Multicollinearity

Drawing conclusions about which of two highly correlated predictors is "the real driver" of an outcome, when multicollinearity has made both coefficients individually unstable.

✓ Fix: Check VIF before interpreting individual coefficients in any multivariable model; address high VIF before drawing conclusions about specific predictors.

Practical Workflow for Checking Assumptions Before Analysis

1. Are your observations independent — no repeated measures, matching, or clustering?
NO → Use a paired, repeated-measures, mixed-effects, or GEE method built for the dependency.
YES → Continue to Question 2.
2. Is your outcome continuous?
NO → Skip normality/variance checks; verify expected cell counts (categorical) or events-per-predictor (logistic/Cox) instead.
YES → Continue to Question 3.
3. Is the outcome normally distributed (tested formally, not by eye alone)?
NO → Use a non-parametric test, or transform the variable (e.g., log transform for right-skewed data).
YES → Continue to Question 4.
4. Are variances roughly equal across the groups being compared?
NO → Use Welch's t-test or Welch's ANOVA instead of the standard version.
YES → Proceed with the standard parametric test.
Finally, check any test-specific assumption — linearity for regression, multicollinearity for multivariable models, proportional hazards for Cox regression — before interpreting the result.

Final Checklist Before Running Your Analysis

1

Confirm observations are independent

Identify repeated measures, matching, or clustering before choosing a test.

2

Test normality of the continuous outcome (or model residuals)

Shapiro-Wilk plus a visual check (histogram or Q-Q plot).

3

Check homogeneity of variance for group comparisons

Levene's test; use Welch's correction if variances differ.

4

Verify linearity for correlation and regression

Scatterplot or residual-versus-predictor plot before fitting the final model.

5

Screen for outliers and influential points

Investigate the cause before deciding whether to keep, transform, or flag a point.

6

Check multicollinearity in any multivariable model

VIF for every predictor before interpreting individual coefficients.

7

Confirm adequate sample size and expected cell counts

The 20% rule for chi-square; events-per-predictor for logistic regression.

8

Test proportional hazards if running Cox regression

Schoenfeld residuals or log-minus-log plots before reporting a hazard ratio.

Further Reading

For deeper methodological background on statistical assumptions and reporting standards in medical research, these external resources are widely regarded as authoritative:

Frequently Asked Questions

What happens if I ignore a violated statistical assumption? +
The specific consequence depends on the assumption, but the general result is the same: your p-value, confidence interval, or effect estimate is no longer trustworthy, even though the software still returns a number with full confidence. Violated independence commonly deflates standard errors and inflates false positives; violated normality distorts small-sample p-values; violated proportional hazards makes a single hazard ratio misleading. Nothing in the output warns you which situation you're looking at.
Do I need to test normality for large samples? +
The Central Limit Theorem makes parametric tests reasonably robust to non-normality in large samples (commonly around 30+ per group), but this robustness applies to the sampling distribution of the mean, not the raw data. It doesn't fully excuse severe skewness, heavy outliers, or small subgroups within a larger dataset. Testing formally, even in large samples, remains good practice.
What is the difference between an outlier and an influential point? +
An outlier is a data point with an extreme value relative to the rest of the dataset. An influential point is one whose removal would substantially change the model's results, identified formally with measures like Cook's distance. A point can be an outlier without being influential, and vice versa.
How do I fix unequal variances between groups? +
Use Welch's t-test instead of the standard independent t-test, or Welch's ANOVA instead of standard one-way ANOVA — both adjust degrees of freedom for unequal spread. Most software reports both versions by default. A log transform can also stabilize variance for right-skewed measurements.
What is an acceptable VIF value for multicollinearity? +
A VIF near 1 indicates no problematic correlation; values up to about 2.5 are generally unproblematic. VIF above 5 warrants closer inspection, and above 10 is widely treated as a serious multicollinearity problem requiring corrective action before interpreting individual coefficients.
How do I check the proportional hazards assumption in Cox regression? +
Use a formal test based on Schoenfeld residuals, available in essentially all statistical software, which flags a violation if residuals show a significant relationship with time. Visually, log-minus-log survival plots that fail to run parallel across groups suggest the same problem.
What do I do if my chi-square table has small expected cell counts? +
If more than 20% of cells have an expected count below 5, or any cell is below 1, use Fisher's Exact test instead — it calculates exact probabilities rather than relying on chi-square's large-sample approximation. Most software will flag this violation or let you request Fisher's Exact directly.
Does logistic regression require normally distributed predictors? +
No. Logistic regression doesn't assume normally distributed predictors or outcome — the outcome is binary by definition. It does assume linearity between continuous predictors and the log-odds, independence of observations, limited multicollinearity, and an adequate number of events per predictor (commonly at least 10).
Can I test assumptions after running the analysis, or must it be before? +
Some assumptions, like independence and adequate sample size, depend on study design and must be considered before data collection. Others, like normality and proportional hazards, are checked after data collection but strictly before interpreting the main result — checking them is part of running the analysis, not a follow-up step performed only if asked.

Once your assumptions are confirmed, these guides take you through the next steps:

Ready to Run a Trustworthy Analysis?

Now that your assumptions are checked, let StatClinic's AI Statistical Assistant confirm the right test and run your analysis. Free, no registration required.

Try StatClinic Free →