|
|
Subject:
Is this statistically significant?
Category: Science > Math Asked by: pizzamik-ga List Price: $20.00 |
Posted:
15 Oct 2005 15:44 PDT
Expires: 14 Nov 2005 14:44 PST Question ID: 580736 |
A professional real estate site analysis company was hired to survey all the locations of a 75 unit restaurant chain, in an effort to optimize future location selection criteria. Over 50 attributes (freestanding, end-cap, square footage, traffic counts, visibility, etc.) were evaluated, and based on that information, a sales prediction matrix was created. The 75 unit chain has 5 drive through locations. Of the top ten overperforming stores (actual sales vs. predicted matrix sales), 4 were drive through. The 5th drive through location indexed very close to predicted sales. Question: Has drive through been proven mathematically significant to the site selection criteria, and if so, how significant is it? Please show your math and explain how you arrived at your decision. Thanks so much! |
|
There is no answer at this time. |
|
Subject:
Re: Is this statistically significant?
From: myoarin-ga on 15 Oct 2005 19:58 PDT |
Have you read the FAQs about Google Answers' policy on homework assignments? |
Subject:
Re: Is this statistically significant?
From: pizzamik-ga on 16 Oct 2005 12:18 PDT |
I work for this company. I regret if my wording of the problem makes it look as if it is a homework assignment. I assure you this is quite real and is something we are exploring right now. My gut feel is this is very significant but this is based on experience, not math. The sample size is small and that is my concern. |
Subject:
Re: Is this statistically significant?
From: khephri-ga on 16 Oct 2005 15:18 PDT |
Not homework? Maybe you should be writing textbooks. OK, I'll fall for it. Mathematically, the probability of seeing such an unexpected arrangement of the data is statistically significant, assuming that you are using the usual 0.05 cut-off for signficance. The easiest way to approach this problem is with a two-by-two table, or: Drive through Not drive through Top 10 4 6 the rest 1 64 There are many ways to evaluate this table, including both the Chi2 and Fisher's exact. Chi2 is the standard method taught for 2x2 tables, but would not be ideal here, because one expects to find, on the average, less than one drive-through in the top 10 performers, and Chi2 should have at least 5 as an expected value. So the Fisher's exact would be preferred. The p value is 0.0008, or very much less than 0.05. Show the math? No way. The calculation requires multiple factorials, with 75!the largest. Just go to one of the many websites that offer on-line Fisher's exact (such as http://www.unc.edu/~preacher/fisher/fisher.htm) and put in your numbers. Just be sure not to put in 6,10,1,65. The site also gives you the formula, in case your are interested (you need to calculate for this data set and add anything more extreme, i.e. all 5 drive-throughs showing up in the top 10). |
If you feel that you have found inappropriate content, please let us know by emailing us at answers-support@google.com with the question ID listed above. Thank you. |
Search Google Answers for |
Google Home - Answers FAQ - Terms of Service - Privacy Policy |