Spearman Rank Correlation

Spearman rank correlation is a nonparametric measure of the strength and direction of a monotonic relationship between two variables, computed as the Pearson correlation of their ranks.

스피어만 순위상관은 두 변수의 순위에 대해 피어슨 상관을 계산하여 단조적 관계의 강도와 방향을 나타내는 비모수 상관 측도이다.

How do I measure the correlation between two variables that are ordinal or not normally distributed?

서열척도이거나 정규분포가 아닌 두 변수의 상관은 어떻게 구하나요?

Spearman's rank correlation (rho) measures how consistently one variable increases or decreases as the other increases, using ranks instead of raw values. It fits ordinal data, skewed variables, outliers, or relationships that are monotonic but not linear.

When to use Spearman Rank Correlation

• Two ordinal variables such as rankings or Likert-type scores
• Continuous variables with skew or outliers that distort Pearson's r
• A relationship that is monotonic but curved

When to use something else

• Both variables are continuous, approximately bivariate normal and linearly related: use Pearson correlation
• Many tied ranks in small samples: consider Kendall's tau

Assumptions and requirements

• Paired observations that are independent across cases
• Variables are at least ordinal
• The relationship is monotonic

What the analysis reports

• Spearman's rho for each pair of variables
• p-value and sample size
• Correlation matrix when several variables are analyzed
• Scatterplot of ranks or raw values

How to report it

There was a significant positive monotonic relationship between [variable 1] and [variable 2], rs([df]) = [value], p = [value].

Related methods

• Pearson Correlation
• Kendall's Tau
• Partial Correlation

Run Spearman Rank Correlation on Quorant