0
$\begingroup$

I'm doing a random effects model. I have a dummy per treatment, which would be "Message", "MessageTax", "Tax" and "Donation" which refers to whether or not the experiment included a donation. The data comes from 2 similar experiments, but one was with a donation and the other was not, so the Message, Tax and MessageTax treatments are in both experiments. This appears in Stata when I run this regression:

xtreg q Mensaje MensajeImpuesto Impuesto Donacion Donacion##Mensaje Donacion##MensajeImpuesto Donacion##Impuesto,re

How can I fix it? enter image description here [![enter image description here][2]][2]

Improve this question
$\endgroup$
5
  • $\begingroup$ It might help if you also add a translation of the error into your question. $\endgroup$
    – Shawn Hemelstrand
    Commented Apr 16, 2023 at 14:59
  • 2
    $\begingroup$ This type of omission of coefficients for variables happens when there are redundant predictors in the model. I don't use Stata, so I might be missing some subtlety, but this question seems very closely related to your question, with 2 answers that might point you in the correct direction. If that's the case, there's nothing to "fix"; your data simply contain multiple predictors that are linearly dependent. $\endgroup$
    – EdM
    Commented Apr 16, 2023 at 15:32
  • 1
    $\begingroup$ You have left out the first part of the output which contains crucial details about sample size. $\endgroup$
    – Nick Cox
    Commented Apr 17, 2023 at 20:35
  • $\begingroup$ Thanks for the feedback. I just added the full image in case you can see something else @NickCox $\endgroup$
    – Carolina
    Commented Apr 23, 2023 at 11:44
  • 1
    $\begingroup$ The message is clear and explicit. For more information, search our site for "collinear*". $\endgroup$
    – whuber ♦
    Commented Apr 23, 2023 at 15:43

0

Reset to default

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.