prob

[ritu]

Icici. Bank collects data on 10,000 respondents
.........dse 2006 question 28....hw to do it

[aditi5000]

P(M)=.68
P(W)=.32
P(C\M)=42/68
P(C\W)=25/32
P(U|C*M)=1200/4200
P(U|C*W)=1400/2500

P(man with unpaid bal) = P(M*u|c)= P(u|M*c)P(c|m)P(m)/ P(u|M*c)P(c|m)P(m)+ P(u|w*c)P(c|w)P(w) = .12/.12+.14 (everything cancels out!)
Fill in the blanks, AND SUBTRACT FROM 1.. you'll get 0.54... not 0.56 but close enough i guess :P
did anyone do this in any other way?

[ritu]

hi aditi....can u pls explain in detail...m nt getting this....

[Chinni18]

Ritu
First you have to see how many men without unpaid balance. That would be 6800-1200=5600
Total respondents are 10,000/
So prob is 5600/10000 = 0.56
The thing here is that men without credit cards also do not have unpaid balances, thats why you do 6800-1200, not 4200-1200

[Chinni18]

Can someone help with this question:
A poll conducted in the class of ECO revealed that 60 percents of students believe that Hong Kong should have minimum wage law. Consider a random sample of 10 students from the ECO class and count x; the number of students who support the minimum wage law.
How would we do this kind of question?

[aditi5000]

Ritu, this is Bayes' Theorem - direct application.
I have simply used the formula.

P(person having unpaid bal given that he's a man) = 12/12+14

Required prob = P(person not having unpaid bal given that he's a man) = 1- P(person having unpaid bal given that he's a man)