A nutrition researcher in Mumbai is analyzing data from a randomized controlled trial comparing the mean hemoglobin levels (in g/dL) of 45 anemic women receiving iron supplementation vs. 45 women receiving placebo. Both groups had baseline hemoglobin measured and post-intervention hemoglobin measured after 3 months. The researcher wants to determine if there is a significant difference in the change in hemoglobin between the two groups. What is the most appropriate next step in statistical analysis?

Explanation

## Comparing Mean Changes Between Independent Groups **Key Point:** When comparing the mean change (or difference) in a continuous variable between two independent groups, an independent samples t-test is the appropriate choice. ### Study Design Analysis | Feature | Details | | --- | --- | | **Outcome variable** | Hemoglobin change (continuous: g/dL) | | **Independent variable** | Group (categorical: iron vs. placebo) | | **Number of groups** | 2 (independent samples) | | **Sample size per group** | 45 (adequate for t-test) | | **Measurement** | Pre-post within each subject, but comparison is between groups | ### Calculation Approach 1. Calculate change score for each subject: Δ Hb = Post Hb − Baseline Hb 2. Compare mean Δ Hb between iron group and placebo group using independent samples t-test: $$t = \frac{\bar{x}_1 - \bar{x}_2}{SE_{difference}}$$ where $SE_{difference} = \sqrt{\frac{s_1^2}{n_1} + \frac{s_2^2}{n_2}}$ **High-Yield:** The key distinction is that you are comparing **between groups** (iron vs. placebo), not within a single group over time. This mandates independent samples t-test. **Mnemonic:** **ICOT** — **I**ndependent groups → **C**ompare **O**utcome (continuous) **T**-test. ### Assumptions - Continuous outcome (hemoglobin in g/dL) ✓ - Two independent groups ✓ - Approximately normal distribution of change scores (n = 45 allows CLT) ✓ - Homogeneity of variance (can be tested; Levene's test) ✓