In the Estimates of Fixed Effects SPSS-table, I don't understand why certain parameters are set to zero an just a dotted line appears (because they are redundant according to SPSS). Moreover, I don't get what the t-tests in this table mean. What is being tested here?

Parameter Estimate Std. Error df t Sig. 95% Confidence Interval

[treatment=,00] * [stimuluscode=-1,00] * QuestionnaireScore ,371900 ,133225 1462,000 2,792 ,005 ,110567 ,633233

[treatment=,00] * [stimuluscode=,00] * QuestionnaireScore ,379735 ,133225 1462,000 2,850 ,004 ,118402 ,641068

[treatment=,00] * [stimuluscode=1,00] * QuestionnaireScore 0b 0 . . . . .

[treatment=1,00] * [stimuluscode=-1,00] * QuestionnaireScore 0b 0 . . . . .

[treatment=1,00] * [stimuluscode=,00] * QuestionnaireScore 0b 0 . . . . .

[treatment=1,00] * [stimuluscode=1,00] * QuestionnaireScore 0b 0 . . . . .

a. Dependent Variable: inclusion. b. This parameter is set to zero because it is redundant.

Any help or comments are very much appreciated!

best wishes, Mariska


The way I investigate 3 way interactions is to condition on one of the variables and then create 2-way interaction plots for the other 2 variables. Creating 2-way interaction plots with continuous variables is a bit trickier but can be done. I'm not sure how this is done in SPSS however.

  • $\begingroup$ Thank you Glen, this is very helpful. I plotted the 2-way interaction per group which makes it very clear. To test the significance, do you think I should simply run the analysis again in each group separately? And than, if significant in one group, investigate the 2-way interaction between the questionnaire score and the stimulus condition further per stimulus condition? $\endgroup$ – user9203 Feb 22 '12 at 16:44
  • $\begingroup$ I'm not sure what you mean, you found the the 3 way interaction to be significant and now are trying to interpret what it means. I don't think any more significance testing needs to be done. $\endgroup$ – Glen Feb 23 '12 at 1:45
  • $\begingroup$ Do you have a reference for that? A reference that says that if I find the 3-way interaction, and I create 2-way interaction plots, that no further testing is required? By the way, I found the SPSS code for generalized: stats.stackexchange.com/questions/40932/… $\endgroup$ – user9203 Nov 2 '12 at 8:33

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.