📊 統計・データ

相関係数

そうかんけいすう

2 つの変数の間の線形関係の強さと方向を -1 から 1 の範囲で表す指標。因果関係を意味しない点に注意が必要。

1 分で読める

Definition

The correlation coefficient quantifies the strength and direction of a linear relationship between two variables on a scale from -1 to 1. The Pearson product-moment correlation is most widely used, where values near 1 indicate strong positive correlation, near -1 indicate strong negative correlation, and near 0 indicate no linear relationship.

Interpretation Caveats

Correlation does not imply causation. A positive correlation between GDP and life expectancy does not prove that higher GDP causes longer lives. A third variable such as healthcare infrastructure or education may drive both.

Additionally, the correlation coefficient only captures linear relationships. U-shaped or exponential patterns will not be detected. Visual inspection through scatter plots is always recommended alongside numerical analysis.

Applications in Ranking Data

In country-level ranking data, correlation coefficients can quantify relationships between income and health indicators, education and life satisfaction, and other paired metrics. Spearman's rank correlation is suited for ordinal data and is less sensitive to outliers.

Use in MyRank

Understanding correlations among the multiple indicators MyRank provides (income, BMI, sleep duration, etc.) helps you build a multidimensional picture of your global standing. Recognizing relationships between metrics is a first step toward data literacy.

関連用語

関連記事

この用語解説は役に立ちましたか?