Surveys are often used to measure high-value information. When you think of a survey, think of a set of variables. It's really that easy. For some problems, a survey question might be considered a dependent variable. For another problem, that same question might be considered an independent variable. Since survey questions are just variables, they can be analyzed using any of the methods we've learned in our leveling course and in the Hubbard text - one sample tests, two sample tests, ANOVA, regression, Chi-square, simulation, etc. Having said all of that, discuss things that should be reflected upon and/or considered in constructing a survey.