## Correct Answer: B. Unpaired t-test The unpaired t-test is the gold standard for comparing the **mean of a continuous variable between two independent groups**. Here, hemoglobin is a continuous quantitative variable (measured in g/dL), and the two groups are independent—meaning observations in one group are not related to observations in the other. This is the classic scenario for unpaired t-test application. The test assumes both groups are normally distributed (or sample size ≥30 by Central Limit Theorem) and have approximately equal variances (tested by Levene's test). In Indian clinical practice, this is routinely used to compare Hb levels between, for example, anemic vs. non-anemic populations, or pre- and post-supplementation cohorts in separate patient groups. The t-statistic is calculated as the difference in means divided by the pooled standard error, and compared against a t-distribution with (n₁ + n₂ − 2) degrees of freedom. This test is parametric and more powerful than non-parametric alternatives when assumptions are met. ## Why the other options are wrong **A. ANOVA** — ANOVA (Analysis of Variance) is used to compare means across **three or more independent groups**, not two. While ANOVA would technically work for two groups, it is unnecessarily complex and not the most appropriate choice. The question explicitly states two groups, making unpaired t-test the simpler, more direct test. ANOVA is reserved for multi-group comparisons (e.g., comparing Hb across three socioeconomic strata). **C. Paired t-test** — Paired t-test is used when the two groups are **dependent or matched**—such as before-and-after measurements in the same individuals, or matched case-control pairs. The question specifies independent groups, ruling out paired t-test. This is a common NBE trap: students who confuse 'two groups' with 'paired data' fall here. Paired t-test uses differences within pairs, not group means. **D. Chi-square test** — Chi-square test is used for **categorical variables** (e.g., presence/absence of anemia, gender distribution), not continuous variables like hemoglobin levels. Hemoglobin is measured on a continuous scale, not as categories. Using chi-square here would require inappropriate dichotomization of Hb data, losing statistical power and violating the test's assumptions. This option tests whether students confuse variable types. ## High-Yield Facts - **Unpaired t-test**: used for comparing means of a continuous variable between exactly **two independent groups**. - **Assumptions**: normality (or n ≥ 30), homogeneity of variance (Levene's test), and independence of observations. - **Degrees of freedom** for unpaired t-test = n₁ + n₂ − 2 (total sample size minus 2). - **ANOVA** is for ≥3 groups; **paired t-test** is for dependent/matched pairs; **chi-square** is for categorical data. - In Indian clinical audits, unpaired t-test commonly compares Hb between rural vs. urban populations, or supplemented vs. non-supplemented cohorts. ## Mnemonics **2 Groups → Unpaired t** **2** independent groups → **t-test (unpaired)**. **3+** groups → **ANOVA**. **Same subjects twice** → **paired t**. **Categories** → **chi-square**. **CAIT Rule** **C**ontinuous + **A**independent + **I**ndependent (2 groups) + **T**-test = Unpaired t-test. Helps distinguish from paired t (dependent) and chi-square (categorical). ## NBE Trap NBE pairs "two groups" with ANOVA to catch students who memorize "ANOVA compares groups" without noting the three-or-more requirement. Similarly, "independent groups" is paired with paired t-test to trap those who confuse "two groups" with "paired data." ## Clinical Pearl In Indian anemia surveillance studies (e.g., NFHS data collection), when comparing mean Hb between two independent district populations, unpaired t-test is the standard analytical tool. This is also used in hospital-based audits comparing Hb outcomes between two treatment protocols in separate patient cohorts. _Reference: Park's Textbook of Preventive and Social Medicine, Ch. 10 (Biostatistics); Guyton & Hall Textbook of Medical Physiology (statistical methods appendix)_
Sign up free to access AI-powered MCQ practice with detailed explanations and adaptive learning.