There is no mathematical problem with having only one item for each concept. In practice, it could be that if you did have several items that would allow you to detect a small effect, but this is a matter of improving the quality of your measures, rather than making any statistical "error."

One additional possibility would be to determine if your six motivations could all be combined into a single scale, which might show a difference. To assess this, you should start with coefficient alpha, which will determine whether or not the items form a reliable scale (i.e., whether they are all highly inter-related).

Hello Justyna,

You could use logistic regression, with sex as the DV. The IVs (the six item/motivator scores) could be treated as ordinal (in spss binary logistic subprogram, categorical rather than continuous) variables. This would help you to identify the "best" linear composite that differentiated sexes for your sample.

Good luck with your work.

David Morse If chi-square did not reveal any relationships, I don't see how logistic regressions with categorical IVs would either.

Hello David Morgan,

Dollars to doughnuts, she was looking at the variables (vs. sex) one at a time. That doesn't rule out the possibility that some ensemble of the variables could work.

Justyna Cuglewska If the Chi-sqr test did not return a significant difference, I would presume there is no gender difference. I would definitely check if I could combine the six motivators to create a composite scale (as suggested by David Morgan).

However, I see that David Morse is suggesting logistic regression (though my opinion coincides with David Morgan that it won't make difference if the Chi-sqr test returns insignificant relation). Please run this and share your experience with us here, so that we could also learn something.