2 Replies Latest reply on Apr 12, 2016 9:41 AM by Jon Boone

    Standard Error in Regression Seems Off

    Jon Boone

      I am running regressions for a few different stores and am looking at values such as the r-squared to see the fit of the line. I know an important value to use is the standard error of the mean, which is listed right below the r-squared value when you click on "describe the trend line". For some of my regressions, my standard error is believable, but for others, it is way too low. A normal standard error for my analysis should be around 8 or 9, however in one of my graphs a trend line has a standard error of only .17. I know this shouldn't be the case because the r-squared value is low and the residual plot does not look that good.

       

      Does anyone have an idea of what is throwing off my standard error value?

       

      Thanks in advance.