0

How can I force a predictor to stay in the model and select others using a stepwise method in a multiple linear regression in SPSS? I would like to force one predictor to appear in my final model regardless of significance.

mariam alemairi
  • 11
  • 1
    It is possible. But off-topic here. Please ask this question in one of SPSS communities. A list of those are found on the info page of the tag spss. – ttnphns Oct 08 '22 at 18:36
  • 4
    There is a statistical aspect to this question that would be on-topic here: whether you should do this stepwise selection at all. See this thread among many others on this site. Even if this were on-topic, I'd be reluctant to provide advice on how to do something that's generally inadvisable. – EdM Oct 08 '22 at 18:44
  • just a curiosity: why your predictor must stay in the selected model? – utobi Oct 08 '22 at 18:52
  • 1
    There's another statistical aspect: given you want to perform some kind of variable selection procedure (in any software) in a linear regression and you wish to keep a subset of variables in the final model, how can you proceed? The solution is to regress all explanatory variables against the subset and apply your procedure to the residuals. Then include your subset with the variables that the procedure identifies. See https://stats.stackexchange.com/a/46508/919 for an extended explanation. – whuber Oct 09 '22 at 12:51

0 Answers0