目次

S値


$$ \log_{\frac{1}{2} } \left( \text{P-value} \right) $$


具体的なP値との対応

S値の解釈

利点と限界

Greenlandらの論文より抜粋

At best, a 95% confidence interval roughly indicates an entire region of high compatibility between the data and possible parameter values within a given model, as judged for example by having less than 4.3 bits of information against the values inside the interval (Using S-values, a simpler and arguably better range for claiming “high compatibility” would be a 5-bit interval (≈ 97% “confidence”.).

95%信頼区間は,せいぜいデータとモデル内のパラメータ値の適合性が高い領域全体を大まかに示すものである.例えば,区間内の値に対する情報が4.3ビット以下であることで判断される.S値を使えば「適合性が高い」と主張するより単純で良さそうな範囲は,5ビット区間(≒97%「信頼」区間に該当)であろう

1)
Sander Greenland (2019) Valid P-Values Behave Exactly as They Should: Some Misleading Criticisms of P-Values and Their Resolution With S-Values, The American Statistician, 73:sup1, 106-114 DOI