I got strange results in spss and I don't know how to explain them statistically. I made correlations between:

1) personality and emotional inteligence (EI)

2) personality and emotional competence (EC)

Correlations (r-Pearson) showed that personality has many connections with EC, but almost none with EI. Everything was OK until I made intergroup comparisons (ANOVA). I made 4 groups of people:

  • with high EC and high EI
  • with high EC and low EI
  • with low EC and high EI
  • with low EC and low EI

I checked if there are significant differences between these gropus, in personality profiles.

And results showed that:

  • If two groups have the same level of EI they don't differ with personality (even if they have different level of EC).

  • And when groups differ with EI, they also differ with personality (even if EC are the same).

These results are unexplainable for me because correlation showed that personality is connected with EC , NOT with EI!

What do you think... What does it mean? I've checked if I did all analysis in proper way.


1 Answer 1


There is an interaction effect between the variables that you have considered. Correlation takes the entire set of values and then finds the relationship between the groups. So this in effect can mask the internal variation arising out of the interaction. Taking intergroup correlations made you find out that indeed interactions exist. This wiki page gives you explanation about interaction effects.

There are many ways to deal with interaction in your analysis, you can use a regression model that takes into account the interaction element Y ~ intercept +A + B + A*B where A*B is the interaction term. Or you could use a two way anova in your case


Your Answer

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

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