>>Return to Tell Me About Statistics!
Spearman’s rank correlation coefficient is an “analytical method for understanding the correlation of ordinal scales,” such as rank data and 5-level rating data.
For example, an employee satisfaction survey comprises data that evaluates satisfaction with “salary” and “company” on a five-point scale. The Spearman’s rank correlation coefficient was used to examine whether satisfaction levels were correlated.
Spearman’s rank correlation coefficient takes values from -1 to 1. As the value approaches ±1, the correlation becomes stronger, and conversely, as the value approaches 0, it becomes weaker. There is no correlation when it is 0, and there is a weak correlation even when it is 0.05. Therefore, a correlation was observed in most cases, though the strengths and weaknesses varied.
A strong correlation is important; however, no statistical standard states “the correlation is strong if there is more than one number.” Analysts determine standards based on their own empirical judgment.
The table below shows the general criteria used in this study. If the value is negative, the absolute value is used.
[Table] Criteria for Spearman’s rank correlation coefficient
![](https://www.carenet.co.jp/english/ensys/wp-content/uploads/2024/01/vol54_table1.png)
We have summarized the common boundaries that have appeared so far.
・Boundary of Cramer’s coefficient of association: 0.1
・Boundary of correlation ratio: 0.1
・Boundary of simple correlation coefficient: 0.3
・Boundary of Spearman’s rank correlation coefficient: 0.3
>>Return to Tell Me About Statistics!
![](https://www-s.carenet.co.jp/english/ensys/wp-content/uploads/2023/02/LI-In-Bug2.png)
![](https://www-s.carenet.co.jp/english/ensys/wp-content/uploads/2023/02/icon-mail-コピー.png)
Comments are closed