I'd be very conservative in interpreting these results. The questionnaire's internal fidelity is either under research basic acceptable levels, or barely above. Their decision to take less items that measure different aspects of a concept is misguided, as it exposes they questionnaire to too much error.
They also use and defend pragmatic validity, which is limited in it's ability to resist scrutiny in most psychometrics manuals.
The questionnaire is really short but has many sub-scales. It's a problem because you might miss a lot of what a concept could need to be included to be fully measured. And for what it's worth, they basically made caricatures of old moral categories and added scales with them.
The statistical models they use are really sophisticated, but you can also link certain genes with the skills to eat with chopsticks, so i'm not impressed there either.
The arcticle ahs been cited a lot, and JPSP is huge in social psychology.
I could go on about methodological limits of this study, The conclusions are interesting, but the method is lacking.