PS Overlapping sets

This topic has expert replies
Master | Next Rank: 500 Posts
Posts: 344
Joined: Sat Nov 12, 2011 3:21 am
Thanked: 1 times
Followed by:2 members

PS Overlapping sets

by sud21 » Sat Sep 19, 2015 5:10 am
200 people responded to a survey that asked them to rank three different brands of soap. The percentage of respondents that ranked each brand 1st, 2nd, and 3rd are listed above.
If no respondents rated the soaps in the order Y, Z, X, how many respondents rated the soap in the following order: X, Y, Z?

100

60

50

30

25
Attachments
overlapping sets.png

Master | Next Rank: 500 Posts
Posts: 363
Joined: Sun Oct 17, 2010 3:24 pm
Thanked: 115 times
Followed by:3 members

by theCEO » Sat Sep 19, 2015 12:50 pm
sud21 wrote:200 people responded to a survey that asked them to rank three different brands of soap. The percentage of respondents that ranked each brand 1st, 2nd, and 3rd are listed above.
If no respondents rated the soaps in the order Y, Z, X, how many respondents rated the soap in the following order: X, Y, Z?

100

60

50

30

25
There a 6 ways in which we can order x, y, z where the first letter is the first position, etc
x y z
x z y
y x z
y z x
z x y
z y x

We are told that y,z,x has no respondent so if we label each available option, we have
a = x y z
b = x z y
c = y x z
d = z x y
e = z y x

we are told that 15% of the respondent choose x as the last choice,
therefore e = 15% x 200 = 30

we are told that 40% of the respondent choose y as the second choice
therefore a + e = 40% x 200 = 80
therefore a = 80 - e = 80 - 30 = 50

therefore 50 people choose x,y,z

ans = c

GMAT Instructor
Posts: 2630
Joined: Wed Sep 12, 2012 3:32 pm
Location: East Bay all the way
Thanked: 625 times
Followed by:119 members
GMAT Score:780

by Matt@VeritasPrep » Thu Sep 24, 2015 11:04 am
Since nobody ranked the soaps Y > Z > X, we know that EVERYONE who thought Y was best must have given the order Y > X > Z. (If you rank Y #1, these are the only two arrangements.)

So Y > X > Z = 60 people.

Using similar logic, since nobody ranked the soaps Y > Z > X, we know that EVERYONE who thought Z was second best must have given the order X > Z > Y.

So X > Z > Y = 30 people.

Since there are only two ways to rate X best (either X > Z > Y or X > Y > Z), these two arrangements must sum to 80, the number of people who rated X best. We have 30 already from X > Z > Y, so the number who gave X > Y > Z must be 50.

Master | Next Rank: 500 Posts
Posts: 344
Joined: Sat Nov 12, 2011 3:21 am
Thanked: 1 times
Followed by:2 members

by sud21 » Sun Sep 27, 2015 3:52 pm
Hi, theceo,

how are you getting 15%? Should not it be 20% (1/5)?

Master | Next Rank: 500 Posts
Posts: 363
Joined: Sun Oct 17, 2010 3:24 pm
Thanked: 115 times
Followed by:3 members

by theCEO » Sun Sep 27, 2015 5:30 pm
sud21 wrote:Hi, theceo,

how are you getting 15%? Should not it be 20% (1/5)?
Hi sud21,

In order to find the percentage of respondent that rank z > y > x, we need to find the graph that addresses x when x is the 3rd choice. Looking at graph #3, we see that 15% of respondent choose x as the 3rd choice.

15% x 200 respondent =30 respondent choose x as the last choice.

Let me know if this help.

How did you get 20%?

Master | Next Rank: 500 Posts
Posts: 344
Joined: Sat Nov 12, 2011 3:21 am
Thanked: 1 times
Followed by:2 members

by sud21 » Mon Sep 28, 2015 4:41 pm
I got that. I got 20% by picking one scenario where x is last out of the 5 scenarios.

Master | Next Rank: 500 Posts
Posts: 363
Joined: Sun Oct 17, 2010 3:24 pm
Thanked: 115 times
Followed by:3 members

by theCEO » Mon Sep 28, 2015 4:46 pm
sud21 wrote:I got that. I got 20% by picking one scenario where x is last out of the 5 scenarios.
Understood. If the graphs were not presented then there will be a 20% chance that X is the last choice.