1
$\begingroup$

I have pre and post measures of the same participants, divided over a Treatment condition and a control condition (no treatment). I have calculated the diff-in-diff by hand and want to know if this is significant.

Therefore, I want to perform a regression with a Treatment dummy (1= treatment, 0=no treatment), a Time dummy (1=post test, 0= prestest) and and the interaction between both. My dependent variable is a math test score.

I know how to create the Treatment dummy. My question is about the Time dummy: Every participant is measured pre and post, so I have 2 columns of math score data in my spss file. I would like to know how to create the Time dummy, as every participant is measured at both times (pre and post).

(I know how to create a dummy variable in spss, but I dont know what the Time dummy should look like).

$\endgroup$
4
  • 2
    $\begingroup$ At the moment this question is quite unclear to me. Could you explain a little more about your set-up? Are you asking for help with how to create dummies in your statistical software (which would be off-topic here) or just a conceptual question about what the dummies should look like when set up? $\endgroup$
    – Silverfish
    Aug 6, 2016 at 11:04
  • $\begingroup$ Please register &/or merge your accounts (you can find information on how to do this in the My Account section of our help center), then you will be able to edit & comment on your own question. $\endgroup$
    – Silverfish
    Aug 8, 2016 at 11:12
  • $\begingroup$ Hi Silverfish, I have registered my Cinta account. Is my question still on hold? $\endgroup$
    – Cinta
    Aug 8, 2016 at 11:38
  • 2
    $\begingroup$ This question is now in the review queue so that people can see if it is sufficiently clear and on-topic. I think that it should be, now it's clear that it's the data format that is the issue rather than the specific software. $\endgroup$
    – Silverfish
    Aug 8, 2016 at 11:41

1 Answer 1

1
$\begingroup$

I have no idea how you do this in SPSS but you need to have a single column for your outcome with two columns for your predictors: one for treatment and one for time. You would also need a column to identify the person so you can specify the repeated measures nature of your data-set. It way well be that SPSS would let you do the analysis with the data-set in the format that you have but you would need to ask on an SPSS-specific forum.

$\endgroup$

Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.