Crunching the numbers on WDWguruDH's resort satisfaction poll

[Granny]

Originally posted by crisi
If you want to repost larger numbers (I don't want to dig in each thread today), I'll rerun the intervals. The margin should get smaller.

(And we could always transform all that non-normal data. Algebraic log functions are my favorite!)

crisi....well, call me dumb but here we go with the numbers as of 6:13 pm on August 10.

So, without further ado....(I'll repeat my caveats from the first post):

The first summary is a combination of the first two responses....Excellent & Good. I included Good because it was defined as meeting all expectations. So the resorts should not be "punished" by high expectations. To me, if a resort is rated Good or Excellent, that's pretty impressive.

Votes to date by resort (Votes for Excellent & Good/Total Votes)

VB = 66/74 = 89%
BCV = 79/116 = 68%
VWL = 103/112 = 92%
BWV = 119/146 = 82%
OKW = 146/160 = 91%
HH = 53/54 = 98% (Holy smokes!)
SSR = 22/32 = 69%



Or, to look at it the other way, we can rate dissatisfaction with a resort by combining the "Disappointed & Really Disappointed" votes.

Votes to date by resort (Votes for Disappointed & Really Disappointed /Total Votes)

VB = 3/74 = 4%
BCV = 14/116 = 12%
VWL = 5/112 = 4%
BWV = 8/146 = 5%
OKW = 5/160 = 3%
HH = 1/54 = 2%
SSR = 5/32 = 16%



I don't know what the numbers are telling us. I just enjoy the statisticians speaking Greek here on this thread!


Dr. Tomorrow...it's all coming clear to me now.

 

[Jimbo]

Originally posted by Granny
I don't know what the numbers are telling us. I just enjoy the statisticians speaking Greek here on this thread!

Spelling flames aren't cool, I know, but I think it's spelled G-E-E-K, Granny.

 

[DrTomorrow]

[Leaping to Granny's defense:]

If we're talking stats, we're talking sigma, beta & chi - sounds like GREEK to me!

 

[kathleena]

Math is easy because it just makes sense. The numbers just react predictably to the equation.

Now - Spelling - there's a different story. Who can figure out some of the twists in spelling?

Ah, but I was always one of the only girls in the advanced Math and Science classes..................

 

[Doctor P]

You know, stats can be funny. I will tell you from experience that the theory and what actually happens with respect to confidence intervals is very different. As part of my job, I run a major policy survey. I can tell you that every year as soon as we have roughly 200 of our responses in, the results for the survey are generally very stable and don't change much with the next 1000 responses we typically get. Also, the general rule of thumb is 400 responses to be representative of a group at a plus or minus 3 percent level with 95% confidence. You get very little additional power with the additional responses. However, I will agree that these results are pretty close for all the resorts and are probably within the margin of error. However, confidence intervals really narrow when the true proportions are as high as these satisfaction numbers are likely to be. I won't bore you with details, but will say that fewer responses are necessary to discern correct proportions when the true proportions are either very high or very low.

 

[crisi]

Range based on numbers above

Satisfied

VB 73-98%
BCV 51-82%
VWL 80-98%
BWV 69-91%
OKW 81-97%
HH 86-100%
SSR 38-92% - note the range with the small (32 responses) sample

Dissatisfied:

VB 0-15%
BCV 3-24%
VWL 0-12%
BWV 1-13%
OKW 0-9%
HH 0-13%
SSR 7-43% (once again - huge range, small sample)

This means that I am 95% confident that the true proportion of satified customers (given the sampling limitations- which are pretty extreme here) at SSR is between 38 and 92%. And I'm 95% certain that the true proportion of dissatisfied OKW guests is between 0 and 9%.

(But yeah, things trend towards the mean - which means that the numbers have more chance of tightening around where they are now than suddenly skewing off into OKW is horrible.)

 

[CarolMN]

Originally posted by crisi
....(snip)....... numbers have more chance of tightening around where they are now than suddenly skewing off into OKW is horrible.)

OOOOOOOOOOOO! Goody! Goody! Goody! Another opportunity for mischief!

OK. Let's see. OKW. Should I vote "horrible" to even things out or should I vote "wonderful" so more people will stay there instead of my beloved BWV? What to do? Oh what to do!


Sincere note of warning to Newbies:
Nothing in this poll is worth serious consideration. Pay it no mind.
Best wishes -