Wilcoxon Signed-Rank Test

Introduction: Wilcoxon Signed-Rank Test

The Wilcoxon Signed-Rank Test is a non-parametric statistical method used to compare median differences between paired observations. It serves as an alternative to the paired t-test when data does not meet the assumption of normality.

Background

The test was introduced by statistician Frank Wilcoxon in 1945 as part of early developments in non-parametric statistics. Unlike parametric tests, it does not rely on the assumption of normally distributed differences. Instead, it ranks the absolute differences between paired values and assigns a positive or negative sign depending on the direction of change. This ranking approach makes the test more sensitive than the simpler 1-Sample Sign Test.

Key Elements/Features

• Paired data: Compares values from the same subject before and after a change, or from matched pairs.
• Signed ranks: Each difference is ranked in absolute value and given a positive or negative sign.
• Test statistic: Based on the sum of the signed ranks, which is compared against critical values or converted into a z-score for large samples.
• Robustness: Performs well with skewed data or outliers, where parametric t-tests may give misleading results.

Formula

\(
T = \min(W^+, W^-)
\)

where 

\(
W^+ = \sum \text{positive signed ranks}, \quad W^- = \sum \text{negative signed ranks}
\)

For large samples, \(T\) can be approximated by a normal distribution with 

\[
z = \frac{T – \frac{n(n+1)}{4}}{\sqrt{\frac{n(n+1)(2n+1)}{24}}}
\]

where \(n\) is the number of non-zero differences.

Applications/Examples

The Wilcoxon Signed-Rank Test is widely applied in education, healthcare, and behavioural research.

• Education: Testing whether a new teaching method changes student test scores compared to the traditional approach.
• Healthcare: Evaluating whether a new drug significantly lowers patient blood pressure relative to baseline measurements.
• Psychology: Measuring changes in anxiety levels before and after therapy sessions.

Relevance/Impact

This test balances robustness and sensitivity, making it ideal for paired data that violates normality assumptions. It is more powerful than the 1-Sample Sign Test while avoiding strict distributional requirements. Its broad applicability ensures reliable results in fields such as medicine, education, and social sciences.

See also

• 1-Sample Sign Test
• Non-Parametric Methods
• Hypothesis Testing
• p-Value