Analysis – As with most of UX, the answer is “it depends.” It depends on the goal of the question, the type of data, and the number of responses. There are no special methods for analyzing survey data, so you would apply standard analysis approaches, whether you’re working with qualitative or quantitative data. You should determine the kinds of statistics you need to answer your research questions, make sure you’re getting the data to support those analyses, and analyze accordingly. For example, you may want means and standard deviations, or you may get more useful data looking at counts or distributions. International Surveys - I couldn’t find much on this topic, and the people I asked didn’t know. However, there may be some work out of Europe, where surveys might be more likely to be administered in multiple languages with multiple cultures. Incentives - There’s been a fair amount of work done in this area, and in general, incentives do help, but they can be tricky to administer (or expensive if you need a lot of responses). Jeff Sauro has a nice article on the topic that sums up the research and provides some guidance. Sample size – Again, it depends. Survey Monkey has a sample size calculator that may be a good starting point if for quantitative data. But other factors may influence the desired sample size. For example, if your sample is fairly homogeneous, you may not need as many people.