In article <[EMAIL PROTECTED]>, John Kane <[EMAIL PROTECTED]> wrote: >Herman Rubin wrote:
>> In article <[EMAIL PROTECTED]>, >> John Kane <[EMAIL PROTECTED]> wrote: >> >Stan Brown wrote: >> >> Herman Rubin <[EMAIL PROTECTED]> wrote in sci.stat.edu: >> >> >Test for understanding, not for imitation of robots. Give >> >> >a few multi-part problems, and be sure to give partial credit. >> >> Excellent advice. I do (try to) test for understanding, by posing >> >> problems in real-world terms and seeing if the students know which >> >> test or calculation to do in which circumstances. Calculators are >> >> allowed for all work. And yes, I give partial credit where it's >> >> warranted. On the other hand, I don't give credit where work is not >> >> shown. In my view, the "right" answer is worthless if you can't >> >> justify it. >> >If there is a 'right' answer how do you justify not giving full credit >> >for a 'right;' answer? >> >For partial answers of course you need to see the work however if the >> >answer is corret then unless you suspect cheating I cannot see how you >> >cannot give full marks. This assumes there is an objective numerical or >> >verbal result. >> I recall one test in an elementary decision theory (college >> algebra level, following probability) in which in one problem, >> students were asked to compute the Bayes risk as a function >> of the discrete action, and in the next problem to state which >> action was best. The grading of this answer was based on the >> answer to the previous problem. >> One can ask which event has a higher probability, and give no >> credit if the answer is based on incorrect ideas. >So you are marking on non-independent questions if I understand you. Ergo >assuming a mistake on A then a mistake on B is given if the candidate >correctly understands what he/she is asked? I hope I am misreading this. In practical situations, the user will compute little. Knowing what the problem means, and what the answers mean, is far more important. This is especially true for the one who will not be an expert. -- This address is for information only. I do not claim that these views are those of the Statistics Department or of Purdue University. Herman Rubin, Dept. of Statistics, Purdue Univ., West Lafayette IN47907-1399 [EMAIL PROTECTED] Phone: (765)494-6054 FAX: (765)494-0558 ================================================================= Instructions for joining and leaving this list and remarks about the problem of INAPPROPRIATE MESSAGES are available at http://jse.stat.ncsu.edu/ =================================================================
