You can do a power analysis using the program g*power, but I am almost certain that it will show you that your sample is too small to detect anything but the largest of effect sizes.
No - I would think too small a sample. Respondents are usually heterogenous (ie variable in their responses) and you have to be able to deal with that inherent variability.