Also if this helps, there was 50 questions, each question being worth 2 points equally.
We had a class of 15 students who took an exam: test scores were: 56, 65, 65, 54, 28, 65, 55, 47, 68, 41, 71, 86, 65, 67, 65 respectively on a non multiple choice exam of 50 questions. The passing mark was exactly a 65, so there has been an accusation thrown at our medical school, he decided to bring up the scores. For the people who scored a 65 or above, he wouldnt let them see it, only the people who failed. I know correlation to a statistical test does not determine causation, but I need to find a probability of this happening. Before I accuse someone of something, I want to throw out a statistical significance of an event happening. Anyone know what I can do? Thank you
Also if this helps, there was 50 questions, each question being worth 2 points equally.
What are you thinking of accusing them of?Before I accuse someone of something, I want to throw out a statistical significance of an event happening.
I don't have emotions and sometimes that makes me very sad.
I want to accuse them of boosting up grades so more people wouldn't fail, the department set a stringent passing mark of 65 so people who were close were given that mark, otherwise they would have to fail more students. However, by university guidelines, this is illegal to do. I need a probability of how 5 people could have attained a score of a 65 on a non mcq test of 50 questions with a class of 15 people. thank you
You better hope that there isn't a (Capital "F") Faculty Union that would get involved in this matter.
for all your psychometric needs! https://psychometroscar.wordpress.com/about/
How is the scoring done? If I assume 0 points for a wrong answer and 2 points for a correct answer, and no partial points, then no one can get 65 points.
I'm surprised that some students were not allowed to see their work. At our University of Applied Science (in the Netherlands) students are always entitled to feedback on their own work.
I'm no expert in probability but would think of something like what is the probability that 5 or more students out of 15 get the same score.
Moderation of results etc? Happens everywhere. Hilarious that you're trying to catch him out though - good luck!
I would go for checking the probability on 5 or more students scoring the same grade. You would also be surprised if 6, 7, 8 etc. or more would have the same grade.
This can be seen as a variation on the famous 'birthday problem'. Instead of having 367 possible days, instead of having 2 or more people having the same birthday, we have 5 or more students having the same grade, and we know that the group has 15 people.
The formula for then determining this probability can be found at the bottom of http://math.stackexchange.com/questi...-same-birthday.
swamifez (10-05-2014)
Would you be so vehement or determined if there were 5 100%?
Stop cowardice, ban guns!
No I wouldn't, but in the history of the English program at our university, there has not been ever one 100%. by him allowing students to go through, he screwed the small minority of students who failed and he knew absolutely what he was doing. i am leaving my university if i can't move onto next year, i spent a lot of money and its an arguement i want to make before i leave.
I am trying to figure this out, let X be the number of marks achieved in a single question; then X follows B(2;1/3). Assuming independence, S=X1+⋯X50, the sum of all marks, follows B(100,1/3). Now compute p:=P(S=65). Define Z to be the number of students with S=65. Clearly Z follows B(18,p). Now compute P(Z=5). Would this work?
You also mentioned somewhere that for each of the 50 questions,
0, 1 or 2 points were awarded. What was the format of questions
and answers? Was it multiple choice? If yes, how did the presentation
of question and answers actually look like, and how was the scoring
performed? Or were these open ended questions, and the rater had to
decide how many points to give to written answers?
With kind regards
K.
swamifez (10-05-2014)
Tweet |