I want to use chi-square to see if (1) "gender" and (2) "satisfaction with a study abroad program" are independent. Trouble is, the data sets are small, and I get some low expected values. I'd like to avoid this, and I can, provided I am able to combine reported gender-satisfaction results from two different years (winter of 2006 and 2007) (same type of students, same study abroad program with same features and characteristics). So my question is if I can do that. If it is permissible. Thanks a lot!