Tuesday, March 5, 2013

Do multiple choice tests really work?

Have you ever thought.. How many gatecrashers can enter a merit list through a multiple choice tests? How much would you gain or loose by random guessing? How tests with multiple correct choices reduces the scourge due to random guessing?

Have a look at an interesting conversation between Srinivas Bhogle and Rajeeva Karandikar about validity of Multiple Choice Tests:



  1. The post is really interesting & thought provoking, Enjoyed reading :)!

    I have a few questions,

    1. "In general, if there are n choices, the negative mark to eliminate the advantage of random guessing is 1/(n-1)."

    Is this also some estimate.. How do we get it?

    2. In the first figure, when 10% of candidates are guessing, the maximum no. of imposters should not exceed 10% of 1000 i.e. 100..


  2. Thanks a lot for sharing such an interesting stuff. Like Dr. Karandikar has said I too feel that traditional question-answer examination is better than multiple choice tests.