I am currently taking a statistics course for my B.A. in sociology and I am having trouble understanding how to choose the alpha level. I understand the reasoning of why it is chosen, but I was wondering do you just choose whichever level you want (normally, we choose .05) ? My professor hasn't elucidated on just how you know which level to choose, so I was just wondering if anyone here may be able to help.
Its based on the level of uncertainity you are willing to take..in other words risk...in terms of estimating the true value. higer alpha value will result in Type II error....accepting the good lot. lower alpha error will result in Type I error... rejecting the good lot. It also depends on what you are testing and the sample size. 5% of alpha corresponds to 95% of the area under the normal curve which is around 2 standard deviations. This is safe always to go by 5 % risk in the estimation than changing it from .05