The author administered standardised personality tests to people tested as fluently bi-lingual in English and German. Each of the four possibilities was tested. Everyone got both tests, some tested first in English, the others first in German. Each group had roughly equal numbers of those claiming English as their first language and those claiming German as their first language.
The data showed two distinct types of behaviour. Some of those tested showed no personality shift at all, while others showed an unmistakable change. There were none who changed a little - either they did, or they didn't.
It's been suggested that those who didn't change were, for example, compound bilinguals, whereas those who did were coordinates. Or vice versa.
The conclusion was that a personality shift for some types of bilinguals is clear and unavoidable.
I really hope I can get this material onto the web.
ADDED IN EDIT: Separating the effects of cultural influences from those of the language itself is difficult, and the author of the thesis went to great lengths to look for changes based on the language. that's what makes this research so important (in my opinion). My next step is to contact the supervisor(s) of the candidate, but that's awkward. Working on it.