Student folklore holds that answer b is very often the correct one in conventional multiple-choice exams, where each question has five possible answers. To test whether this belief holds for certain instructors, the students plan to examine the next 20 multiple-choice questions asked by the instructor.
Isthebinomialapproachtheappropriateinthisinstance?In other words, would the binomial assumptions be reasonable in this situation?
Assumingthatonlyoneanswertoanyquestionisregardedascorrect, what is the natural null hypothesis to be tested?
Should the research hypothesis one-sided or two-sided?