zepid.calc.utils.s_value¶

zepid.calc.utils.
s_value
(pvalue)¶ Function to calculate the Svalue. The ‘S’ stands for Shannon information or surprisal values. The name comes from Claude Shannon for this work to information theory. Svalues are calculated from pvalues using the following transformation
\[s = \log(p)\]The Svalue transformation allows a more intuitive explanation of what pvalues tell us about the null hypothesis and alternative hypothesis compatibility. The Svalue tells us how many ‘bits’ of information exist against the null hypothesis. For an example, a Svalue of 5.1 is no more surprising than seeing heads for 5 fair coin tosses. The Svalue should be rounded down in the interpretation
Note
Svalues do NOT have a significant cutpoint. Rather this transformation is to help build intuition what information a pvalues is providing and the corresponding ‘surprisal’ of a result
Parameters: pvalue (float, container) – Pvalue (or array of pvalues) to convert into a Svalue(s) Returns: NumPy array of calculated Svalues Return type: array Examples
>>> from zepid.calc import s_value >>> s_value(pvalue=0.05)
References
Greenland S. (2019). Valid Pvalues behave exactly as they should: Some misleading criticisms of Pvalues and their resolution with Svalues. The American Statistician, 73(sup1), 106114.
Amrhein V, Trafimow D, & Greenland S. (2018). Inferential Statistics as Descriptive Statistics: There is No Replication Crisis if We Don’t Expect Replication. The American Statistician.