Question: The decision whether to publish papers submitted to scholarly journals is typically based on the opinions solicited by the editor of the journal from knowledgeable referees. An article in the American Economic Review reported the results of an experiment in which a randomly as-signed fraction of the referees of papers submitted for publication were not told of the identity of the author (the "Blind" sample). The goal was to determine whether there was any bias caused by the referee's knowledge of who the author of the paper was. The following table contains some of the results of the experiment:
a. Why is it important to assign papers to blind (as opposed to nonblind) referees randomly? What test can you propose for the success of randomization?
b. Test the null hypothesis that it makes no difference in acceptance rates whether papers are assigned to blind or nonblind referees.
c. Calculate the p-value associated with part (b).
d. Test the null hypothesis that it makes no difference in acceptance rates whether papers are assigned to blind or nonblind referees for male and female authors separately.
e. Do these data imply that there is discrimination against female authors?