The problem I am having is that I have a distribution that I want to remain the same shape, but the mean and variance altered. When I use scale() in R, it can change to a standard normal, but I am having trouble to get other distributions (as when you change the mean, it also affects the standard deviation and vice versa).

  • 2
    $\begingroup$ When you add a constant to all values, that changes only the mean and not the SD. When you rescale all values, that changes all moments but in a predictable way. In particular, it will not change the mean if the mean is already zero. So, why not (a) shift the variable to make the mean zero, then (b) rescale it to make the SD equal to 10 (leaving the mean equal to zero), then (c) shift the mean to 5? $\endgroup$
    – whuber
    Commented May 21, 2013 at 14:56
  • $\begingroup$ That is indeed the answer. $\endgroup$ Commented May 21, 2013 at 15:35


Browse other questions tagged or ask your own question.