I used a 3x2 between-subject ANOVA to examine the effect of colour (three levels: colourless, light red, and dark red) and odour (odourless or not) on perceived sweetness. The results indicated significant main effects for colour and odour and also a significant interaction of these two factors. I don't know what to do next.

  • What post-hoc tests can I run to understand this interaction?
  • How can I do it in SPSS?
  • 2
    $\begingroup$ This is more difficult to read than is necessary. You may find it quicker to ignore upper case and be casual about punctuation when you type, but when you are asking a favour, it's a better tactic to write with more attention to the usual conventions. $\endgroup$
    – Nick Cox
    May 14, 2013 at 18:45
  • $\begingroup$ you are totally rght but im in a very bad situation because tommorow i have to submit and i dont know what to to do. thank you again $\endgroup$
    – user25650
    May 14, 2013 at 18:49
  • 3
    $\begingroup$ You don't have to do anything. If you are lost and in a hurry and since the design is not very complicated and all effects are significant, just interpret the pattern of means directly. $\endgroup$
    – Gala
    May 14, 2013 at 19:02
  • $\begingroup$ Thank you Gahel Laurans but my supervisor told me its not ok that and i have to run independent or paired t tests to see the differences from the interaction.to see exactly if e.g dark red colour with strawberry odour enhanced more the perceived sweetness than did the light red colour with strawberry odour and if for example no colour with do odour did not have any significance difference with the scores of no colour with strawbeery odour. is there any other solution? than you in advanceagain $\endgroup$
    – user25650
    May 14, 2013 at 19:10
  • $\begingroup$ hello, i would like to ask how i check the simple main effects ? i used 3x2 anova and i find significant interaction? whixh post hoc test to use ? thank you in advance $\endgroup$
    – user25650
    May 14, 2013 at 21:00

1 Answer 1


This comes up frequently on the site. Odour has an effect, colour has an effect, and they interact. Therefore, the degree of odour's effect varies across colour. You know this now. Look at the means and describe the amount of effects.

Let's say that you checked the effect of odour for light red and it was significant but it wasn't for colourless. What does that tell you? It actually doesn't tell you anything meaningful. It doesn't tell you there's a difference between those conditions because you didn't test that. And, you already knew there was a main effect of odour, did it suddenly go away? What does tell you there's a difference between the conditions is the interaction. What happens if all of your simple effect tests are significant? Does that help you interpret your interaction? No. You can't infer anything across your conditions from those simple effects.

Please add a table of your 6 means to your question. It is possible that you might need to test some kind of subset of the data but simple effects aren't going to get you meaningful info. If you need anything, more than likely you'd need a 2x2 interaction or a direct test of the differences among effects. But it's doubtful you even need that.

  • $\begingroup$ thank you very much. so, pairwise comparisons i run with syntax and gave me mean differences among the subtypes of factors in different conditions it is ok ?? $\endgroup$
    – user25650
    May 14, 2013 at 21:56
  • $\begingroup$ It doesn't really tell you anything. You should really post the means you have in your question. $\endgroup$
    – John
    May 15, 2013 at 0:55

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.