Forum Discussion
Get p value from Independent Sample T-Test
Like this thread: https://community.powerbi.com/t5/Desktop/How-to-replicate-T-TEST-excel-function-in-powerBI/m-p/2866324 I am also trying to replicate Excel's t.test using DAX.
One of the info in that thread is this: https://datakuity.com/2022/02/07/paired-t-test-in-power-bi-using-dax/#P-value. Unfortunately, it is a bit too complex for me to understand. I'm also not good at statistics.
I did manage to replicate this: https://www.statisticshowto.com/probability-and-statistics/t-distribution/independent-samples-t-test/#ISTTbyhand. With the t value here, is it correct to use T.DIST.2T(t value, deg of freedom) to get the p value?
I've done some testing using Scipy in Python to check the results. T.DIST.2T does appear to give the p value if you can provide the t statistic. Why there is a T.DIST.2T but not a T.TEST in DAX is puzzling. It strongly implies the T.DIST.2T in DAX does something very different. I'll mark this as the "Solution" until I figure how what exactly T.DIST.2T does, why T.TEST is missing, and why I am forced to manually calculate the t-statistic.
4 Replies
Greg_Deckler Did you ever get around to add that to your conversion list?
Hi Greg_Deckler , Did you find anything on calculating the p-value in Power BI?
I've done some testing using Scipy in Python to check the results. T.DIST.2T does appear to give the p value if you can provide the t statistic. Why there is a T.DIST.2T but not a T.TEST in DAX is puzzling. It strongly implies the T.DIST.2T in DAX does something very different. I'll mark this as the "Solution" until I figure how what exactly T.DIST.2T does, why T.TEST is missing, and why I am forced to manually calculate the t-statistic.
