This is perfectly OK, I think, and I'm not sure there's a satisfactory alternative. A few examples from academic studies:

"Whilst four studies were able to show a significant difference for the pain outcomes in favour of relaxation for the pre- and post-treatment assessments, few statistically significant differences were reported in favour of relaxation when between treatment comparisons were used. Only three studies reported statistically significant differences in favour of relaxation (judged as a significant difference for at least 1 of the pain outcomes) compared to the other treatment groups."
http://europepmc.org/abstract/MED/9543032/reload=0;jsessioni...

"Analysis of the visual acuity scores (discriminatory ability) shows a sex difference in favour of the males (p> 0.01)."
https://dspace.lboro.ac.uk/dspace-jspui/handle/2134/10871

"Both skeletal symptoms (back pain) and fractures were uncommon during the trial period but both showed non-significant differences in favour of the MCHC group"
http://europepmc.org/articles/PMC2418401

--------------------------------------------------
Note added at 56 mins (2012-12-02 09:04:36 GMT)
--------------------------------------------------

Sorry, I posted confidence 5 by mistake; I meant to make it 4.

--------------------------------------------------
Note added at 8 hrs (2012-12-02 17:08:11 GMT)
--------------------------------------------------

Having seen other answers, I think it's important to make it clear that that "a favor de" is purely statistical/numerical here; it doesn't carry any implication of qualitative superiority. A difference in favour of X compared to Y simply means that the numerical result for X was greater than that of Y. This is the case in both languages (ie. with "difference in favour of" as well).