When a student's answer on this exam doesn't match the answer key the computer assumes (by default) that the student was just rounding wrong and deserves the point. It then uses an algorithm to determine whether their response is too far from the correct answer and is actually wrong. If it determines the student response is "close enough" to the correct answer then it will give the student the points, but if it decides it's "too far" then it won't. The alpha level in this case determines how close is "close enough" to the correct answer. What are valid reasons to change the alpha from the typical 0.05? ---Select-- Use a 0.02 so that you don't give points to students who don't deserve them |---Select-- Use a=0.015 so that students who only had rounding issues don't lose points ---Select-- Use a 0.20 so students who are doing it wrong won't think they're getting it right ---Select-- Use a=0.15 so that you can give students as many points as possible ---Select-- Use a 0 so that there won't be any error in the decision |---Select--- ✓ Use a=0.001 so that we can have the most power in our statistical test