Unbalanced sample - which test do I use?

#1
I have two samples - group 1 has 26 participants / group 2 has 44 participants

I was planning to use a two-way anova to analyze various aspects of my survey, but now am wondering if I need to use Friedman's two-way anova or something completely different. Any help would be greatly appreciated.
 
#2
Here is some additional information:

27 - Male
43 - Female

26 - Ensemble Participants (college freshmen who continued playing an instrument in college)
44 - Ensemble Nonparticipants (college freshmen who discontinued playing an instrument in college)

Thanks!
 

Miner

TS Contributor
#3
This sounds more like a 2-sample t-test or a 1-way ANOVA with 2 levels unless I am missing something.

Added: Just saw your additional information. A 2-way ANOVA will work. It is not necessary that the experiments be balanced.

Added later: This assumes that you have continuous response data. I saw that you mentioned a survey which often uses ordinal responses.
 
#4
Thanks. I am hoping to examine the relationship between and among these groups regarding a variety of influences, beliefs, and scores in other areas. My understanding was that two-way anova was required to compare in and between:


Gender vs. Participation and the relation to "Academic Major" or "Dropout Proneness (score on a test that measures likelihood of dropping out of college). Does that change your thoughts or confirm your original thought?
 
#5
My survey is all Likert scale "agreement" statements. "Strongly Agree" to "Strongly Disagree". The responses were then coded 5-1 (5 = Strongly Agree; 1 = Strongly Disagree) in order to make the "measurement" interval scale instead of ordinal or nominal. Does that make sense?

I also mentioned "dropout proneness" and "Academic Major". Dropout Proneness was measured 1-9 (1 = less likely to drop out) and Academic Major was (1 = Arts & Letters; 2 = STEM "Science, Tech, Engineering, Math).