Stats are super complex!! After taking introductory and advanced statistics class, and perhaps the more I take classes, the more I feel I'm lost and need help.
It's because:
1. There are countless possibilities of your statistical model, whether the variance of the residuals is equal or not (usually not!), whether the residuals are normally distributed or not, and the types of analyses you want to do!! It's next to impossible to be an expert of all unless you are a stats major.
2. Need to depend on R packages to do the analyses, but not all packages are perfect: some cannot deal with certain types of models, and some cannot handle heteroscedasticity and so on. So.. you would either spend a lot of time searching for 'the perfect package', or use the incomplete method that's available in your package of choice. And it's very hard to judge whether the alternative method is good 'enough'.
Stats.. seems to confuse many researchers that are not a statistician. How can we improve this situation?
1. Mandate ~4 courses to cover all possible models and analyses methods to arm them?
2. A stats consultant on every floor? -- But what if when we graduate and go into an industry?
No comments:
Post a Comment