The Imperialist Dog is the rare specimen that enjoyed his experience in research methods:

Of course, the vast majority of undergraduates hated the class and panned it (disclosure: I liked it and got an A-). They are apparently happy being spoonfed and unable to analyze data for themselves. Given the tendency to take the easiest possible path, the department will probably make the class an elective at some point, then abolish it entirely.

Everyone doesn’t need to know how to do multiple linear regression, but a knowledge of what terms mean (sampling, confidence interval, etc.) and how data may be manipulated would prevent some of the more egregious deceptions perpetrated by misusers of statistics.

One of the things I bumped into teaching the equivalent course here at the University of Mississippi is that a lot of the students didn’t seem to get the point of the statistical portion of the course (which may have been partially my fault, since it was my first time teaching the course). Understanding why we’d want to test hypotheses and talk about variances is perhaps more important than the actual algebra involved, but I’m not sure you can have a solid understanding of the former without comprehending at least some of the latter. (That’s not to say I know the formulas for most of these things off the top of my head; that’s why we have R and Stata, not to mention Greene and Kennedy.)

I firmly believe nobody should draw any conclusions from survey data unless they fully comprehend what the terms “margin of error” and “confidence interval” mean. Furthermore, anyone who ever uses the results of a “web poll” to decide or justify anything more important than the SportsCenter showcase highlight ought to be publicly executed.