I have missing values (MCAR) for which I used EM to fill in those values. Some of the imputed values are negative integers or zero. I am using a likert scale to measure responses, and thus i need the missing values to equal between 1 and 5.

To deal with this problem i re-coded my data table in SPSS so that likert scale runs from 0 to 4, this way if the obtained missing value equals zero I can use it. However this does not solve the negative integers. Is it possible to put a restriction on the results so that they don't come up as negative?

Any suggestions, or references to a book/article about this?


Update: to anyone sharing this problem, this may be helpful:


  • $\begingroup$ EM should return a maximum argument for the imputed values, argument taken over the range of possible values, hence you can restrict the imputed values to be in any set you want them to be. $\endgroup$ – Xi'an Nov 24 '16 at 19:37
  • $\begingroup$ Thank you for responding. Do you have advice as to how I can do this in SPSS? $\endgroup$ – Northerntravel Nov 26 '16 at 16:49
  • $\begingroup$ Not at all. This is not the primary purpose of this forum. $\endgroup$ – Xi'an Nov 26 '16 at 17:01

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.