My Math Forum Comparing multiple data sets with different n values

 April 7th, 2008, 11:53 AM #1 Newbie   Joined: Apr 2008 Posts: 1 Thanks: 0 Comparing multiple data sets with different n values I'm not sure if this is the correct place to toss this topic out there, so if this is the wrong place, I apologize. I'm having a wine tasting party in a few weeks. I have little cards that I hand out to everyone who tastes the wine, and they score it based on a 1-5 rank, 1 being the worst and 5 being the best. If everyone voted on every wine, it'd be easy to select a winner at the end of the night, but my problem is that based on past parties, not everyone ends up voting on each wine, and so the value of n varies from one set to another, and my ancient stat classes in college aren't of any help, though I'm pretty sure that's my fault and not the stat classes. How can I compare these sets in a statistically accurate way and determine a winner? As an example, let's say that the Merlot (sorry Sideways) gets the scores {1,3,3,4,5,5,5}, the Sauvignon Blanc gets {2,2,3,4,4,4,4,4,5,5,5,5}, and the Pinot Grigio gets {1,1,2,3,5,5,5,5,5}. Is there a way to weight them according to the number of responses to make the comparisons accurate, or is the data not directly comparable?
April 7th, 2008, 01:43 PM   #2
Global Moderator

Joined: May 2007

Posts: 6,804
Thanks: 715

Re: Comparing multiple data sets with different n values

Quote:
 Originally Posted by cboldon I'm not sure if this is the correct place to toss this topic out there, so if this is the wrong place, I apologize. I'm having a wine tasting party in a few weeks. I have little cards that I hand out to everyone who tastes the wine, and they score it based on a 1-5 rank, 1 being the worst and 5 being the best. If everyone voted on every wine, it'd be easy to select a winner at the end of the night, but my problem is that based on past parties, not everyone ends up voting on each wine, and so the value of n varies from one set to another, and my ancient stat classes in college aren't of any help, though I'm pretty sure that's my fault and not the stat classes. How can I compare these sets in a statistically accurate way and determine a winner? As an example, let's say that the Merlot (sorry Sideways) gets the scores {1,3,3,4,5,5,5}, the Sauvignon Blanc gets {2,2,3,4,4,4,4,4,5,5,5,5}, and the Pinot Grigio gets {1,1,2,3,5,5,5,5,5}. Is there a way to weight them according to the number of responses to make the comparisons accurate, or is the data not directly comparable?
The simplest way is to compute the average for each wine. This would be reasonable as long as each wine has some entries.

 April 7th, 2008, 03:01 PM #3 Global Moderator     Joined: Nov 2006 From: UTC -5 Posts: 16,046 Thanks: 938 Math Focus: Number theory, computational mathematics, combinatorics, FOM, symbolic logic, TCS, algorithms It's a somewhat hard problem since people have different standards -- one person might give out mostly 4s and 5s, and another rate wines 1, 2, 3, 4, and 5. But ignoring that point for a moment, in the interest of simplicity: I would recommend taking the median, then the average as a tiebreaker. That way there's less incentive, as it were, to rate a wine you didn't like artificially low.

 Tags comparing, data, multiple, sets, values

 Thread Tools Display Modes Linear Mode

 Similar Threads Thread Thread Starter Forum Replies Last Post ka0ttic Computer Science 3 March 17th, 2014 04:08 AM cd_gary Algebra 4 March 16th, 2012 09:49 AM willf Advanced Statistics 0 January 24th, 2012 04:22 AM carter22 Algebra 2 March 1st, 2010 05:49 PM cd_gary Abstract Algebra 0 December 31st, 1969 04:00 PM

 Contact - Home - Forums - Cryptocurrency Forum - Top