How to Choose Statistical Test in Clinical Research: T-test, Mann-Whitney U / Ranksum, ANOVA, Kruskal-Wallis, Paired t-test, Wilcoxon Signed-Rank, Chi-square, Fisher’s Exact, Log-rank, Cox regression

---

Step 1 Identify the Dependent Variable Y

Y (Outcome) Example	Data Type	Typical Scale / Notes
Blood-loss volume (mL), Hb level (g/dL)	Continuous	Numeric, theoretically infinite decimals
Gender (male / female)	Binary	2 ordered or unordered categories
Mortality (yes / no)	Binary	—
Pain score (1–10)	Ordinal	Ranked but distances unequal
Satisfaction (low / med / high)	Ordinal	—
Survival time (days)	Time-to-Event	Event plus censoring
Blood group (A / B / AB / O)	Categorical > 2 (Nominal)	Unordered, >2
Seizure count per month	Count	Non-negative integers, often skewed
Proportion of wound infections (%)	Proportion / Rate	Bounded 0–1; often with different denominators

---

Step 2 Define the Independent Groups X

1. How many groups? (2 vs > 2)

2. Structure:

   • Independent groups (different patients / subjects)

   • Paired / repeated (same subject measured twice or more)

(If X itself is continuous—e.g., dosage in mg—see “When to use regression” below.)

---

Step 3 If Y Is Continuous → Check Normality

stata: hist postopbleeding, normal swilk postopbleeding

• p > 0.05 → “looks normal” → parametric tests

• p < 0.05 → not normal → non-parametric tests (or transform / use GLM)

---

Step 4 Master Test-Selection Table

Dependent Y	# Groups	Structure	If Normal / Large-Sample	If Not Normal / Small n / Ordinal
Continuous	2	Independent	Student t-test	Mann-Whitney U / Wilcoxon rank-sum
	> 2	Independent	One-way ANOVA	Kruskal-Wallis
	2	Paired	Paired t-test	Wilcoxon signed-rank
	> 2	Repeated	Repeated-measures ANOVA	Friedman test
Binary	2 or > 2	Independent	χ² test of independence	Fisher exact (if any cell < 5)
Categorical > 2 (Nominal)	2 or > 2	Independent	χ² test (RxC)	Fisher exact (if sparse)
Ordinal	2	Independent	—	Mann-Whitney / Wilcoxon rank-sum
	> 2	Independent	—	Kruskal-Wallis
Count	2 or > 2	Independent	—	Poisson test or negative-binomial
Proportion / Rate	2	Independent	z-test for two proportions	Fisher exact
Time-to-Event	2 or > 2	Independent	—	Log-rank test / Cox proportional-hazards

---

Step 5 When to Use Regression Instead of a Simple Test

If you need …	Regression of choice
Adjust for ≥1 covariate (age, sex, baseline Hb, …)	Linear (continuous Y), logistic (binary Y), multinomial logistic (nominal Y > 2), ordinal logistic, Poisson/negative-binomial (counts), Cox (time-to-event)
Predict risk / odds / mean rather than test difference	Same as above
Model continuous X (e.g., dose in mg)	Include X as continuous term or use spline
Handle interaction terms (e.g., Tx × sex)	Any generalized linear model (GLM)

---

Step 6 Example Walk-Throughs

1. Post-op bleeding (mL) across 3 TXA dose groups

   • Y: Continuous

   • 2 independent groups

   • Normality p < 0.05 → Not normal

   • → Kruskal-Wallis

   stata: kwallis postopbleeding, by(txagroup)

   
   

2. Infection type (A/B/C) by TXA vs no TXA

   • Y: Categorical (>2)

   • 2 independent groups

   • → χ² test (2×3 table)

   stata: tabi 20 5 3 \ 15 12 8 , chi2

   
   

3. Seizure counts pre- vs post-drug in same patients

   • Y: Count, paired

   • Small n, skewed

   • → Wilcoxon signed-rank or Poisson GLMM if covariates needed
     

4. 30-day survival across 4 surgical centers

   • Y: Time-to-event

   • 2 independent groups

   • → Log-rank test; follow-up Cox model for adjustment

   stata: sts test center stcox i.center age sex

   
   

---

Quick Recap

1. Start with Y: Identify its scale.

2. Count & classify X groups: 2 vs > 2, independent vs paired.

3. If Y is continuous: Check normality.

4. Plug into the table to pick the test.

5. Need adjustment or continuous X? → Move to regression.