https://communities.sas.com/t5/SAS-Programming/Sensitivity-for-over-sampled-data/m-p/679045#M205039 <P>If a rare event on a logistic regression is 3% and if I have a confusion matrix from an oversampled data - do I need to adjust sensitivity for population event probability. My understanding was sensitivity won't change for oversampling.&nbsp;</P> Tue, 25 Aug 2020 03:51:50 GMT subhadip_kumar 2020-08-25T03:51:50Z https://communities.sas.com/t5/SAS-Programming/Sensitivity-for-over-sampled-data/m-p/679045#M205039 <P>If a rare event on a logistic regression is 3% and if I have a confusion matrix from an oversampled data - do I need to adjust sensitivity for population event probability. My understanding was sensitivity won't change for oversampling.&nbsp;</P> Tue, 25 Aug 2020 03:51:50 GMT https://communities.sas.com/t5/SAS-Programming/Sensitivity-for-over-sampled-data/m-p/679045#M205039 subhadip_kumar 2020-08-25T03:51:50Z https://communities.sas.com/t5/SAS-Programming/Sensitivity-for-over-sampled-data/m-p/679131#M205069 I think it is yes.<BR />population event probability would affect the cutoff p-value <BR />and furthermore affect the predicted 0 or 1 . Tue, 25 Aug 2020 12:10:22 GMT https://communities.sas.com/t5/SAS-Programming/Sensitivity-for-over-sampled-data/m-p/679131#M205069 Ksharp 2020-08-25T12:10:22Z