Friday

April 18, 2014

April 18, 2014

Posted by **holly** on Thursday, July 5, 2007 at 5:17pm.

Let N be the number of questions. You are asking for the probability of getting N/5 or fewer wrong by randomly guessing.

Assume N>> 1 and use an error-function approximation for the number of questions answered correctly. The mean will be N/2 and the standard deviation will be sqrt[N*p*(1-p)] = (1/2)sqrtN, since p = 1/2 is the probability of getting each one right.

Inegrate the probability of correct answers from 0.8 N to N, which you can treat as infinity when N is very large.

The larger the value of N, the less likely you will be to get 80% right.

For N=25, I get a standard deviation of 2.5 about a mean of 12.5, and the probability of getting 20 or more right will be 0.0013. It becomes a much lower probability of N = 100. You need to use an error function table for this sort of problem.

**Related Questions**

stats again - ok I get it, so once I get the prob figred for one person, how do ...

stats - A study by Hewitt Associates showed that 79% of companies offer ...

stats - Test scores on a university admissions test are normally distributed, ...

Statistics - 5. Scores on a recent national statistics exam were normally ...

Stats - A group of 56 randomly selected students have a mean score of 30.8 and a...

Stats - Consider the following Hypotheses test: Ho:m>=80 Ha:m<80 A sample ...

Stats - National Packing Materials Company claims its X20 box can hold loads up ...

Probablity - Samir had prepared the problem tests for Stages 1 to 5 of Geometry ...

Writing inequalities - you must have an average score of at least 80 to get a B...

math - A random sample of 2000 adults showed that 1120 of them have stopped at ...