I fit a model with transformed data with log10 transformation. I would like to report the results using the original scale and not the transformed means. I know you can back transform the log means by taking 10 and raising it to the power of the log-mean. I would also like to have the standard error back-transformed. Is this possible and can you do it the same way as for means? What about confidence intervals?
Asked
Active
Viewed 27 times
0
-
Does this answer your question? Calculating standard error after a log-transform This question deals more directly with the confidence interval issue, and this question provides more links and discusses alternative approaches to log transformation. – EdM Oct 20 '22 at 20:10
-
1Also, note that if you "back transform the log means by taking 10 and raising it to the power of the log-mean" you are estimating the geometric mean of the original data, not the usual arithmetic mean. – EdM Oct 20 '22 at 20:19