#1  
Old July 13th, 2002, 07:30 PM
quetico
Guest
 
Posts: n/a
Default Some interesting stats


Found some interesting stats on some of the cruise lines. Results were derived from
the 'rate a cruise' section of cruisemates.com. Of the five possible categories, the middle one
has been ignored and the sum of the top two are divided by the same of the bottom two...
sort of gives a satisfaction quoteint, I guess.
A score of three was suggested as the lowest acceptable, but I would think four would
be more like it.
Anyway, they're informative....

FOOD SERVICE CABINS VALUE
NCL

NORWAY 1 1 1 1
DREA M 0 0 0 0
MAJESTY 1 1 0 1
SEA 1 1 0 1
SKY 1 1 0 1
STAR 1 1 1 1
SUN 1 1 1 1
WIND 1 7 1 1
AVERAGE 1 1 1 1

CARNIVAL

CONQUEST 3 3 7 5
PRIDE 10 10 95 9
SPIRIT 4 4 13 7
TRIUMPH 6 4 11 6
VICTORY 19 7 26 9
AVERAGE 9 6 30 7


CELEBRITY

CENTURY 31 15 93 22
GALAXY 6 6 21 6
HORIZON 22 14 14 8
INFINITY 14 45 17 13
MERCURY 14 13 20 9
MILLENIUM 5 4 7 5
SIMMIT 9 43 88 21
ZENITH 13 28 76 7
AVERAGE 14 21 42 11

PRINCESS

DAWN 15 15 13 15
GOLDEN 19 29 27 3
GRAND 4 4 5 5
OCEAN 16 44 26 28
REGAL 3 3 11 2
ROYAL 85 1 1 1
SEA 82 92 19 22
STAR 6 15 18 6
SUN 19 21 6 19
AVERAGE 28 17 14 11
Reply With Quote
  #2  
Old July 13th, 2002, 09:10 PM
ErnieMCC ErnieMCC is offline
Senior Member
Admiral
 
Join Date: Aug 2000
Posts: 1,274
Default Re: Some interesting stats

Well... as you seem to be "statistical minded", I assume you know that any survey that depends on voluntarily submitted input is basically an invalid survey. Interesting, but representative of nothing more than the feelings of the population that participated in the survey as compared to the population sailing on the ships.
Reply With Quote
  #3  
Old July 14th, 2002, 07:30 AM
quetico
Guest
 
Posts: n/a
Default Re: Re: Some interesting stats

Ahhh.. so that makes all of the poles that are taken worthless. You'd best contact CNN and the rest with your insight....
Reply With Quote
  #4  
Old July 14th, 2002, 11:41 AM
John John is offline
Senior Member
First Mate
 
Join Date: Aug 2000
Posts: 480
Default Re: Re: Re: Some interesting stats

What he is saying that is for statistical anaylisis to be some what reliable you need to have either a complete survey of the population or a random sampling. These polls that you gathered were just from people who found the website and answered the poll. We have no idea how many people were surveyed and what percentage of those that were on the boat so it really could not be considered a random sample there fore it can't be considered a valid survey for statistical purposes.

Can you tell that my college room mate for 2 years got his degree in statistics. Extremely boring stuff.
Reply With Quote
  #5  
Old July 15th, 2002, 06:42 AM
quetico
Guest
 
Posts: n/a
Default Re: Re: Re: Re: Some interesting stats

Point taken, John... but using the stats fromt this board are no more invalid than any poll taken by CNN, since they also poll only a section of people taken at random.
I f the stats came from boards labelled ' problems with cruises' or something like that, then I think they'd be in doubt, but this isn't the case.
Actually, our elective system works the same way... and to a degree, our judicial system as well.
But there does seem to be a trend i the data... sailing NCL, for instance, looks like problems consistently, right across the board. And my trip booked on the millenium may not be as good as my trip on the summit or the ocean.
Q
Reply With Quote
  #6  
Old July 15th, 2002, 11:24 AM
Marnie
Guest
 
Posts: n/a
Default Re: Some interesting stats

You're all joking, right? Who cares how many people were surveyed or if the survey was statistically valid? Nobody's going to make life-changing decisions based on it. Quetico, thanks for taking the time to share the results of cruisemates survey. If the information provided on Cruisemates.com is so useless why are you here?
Reply With Quote
  #7  
Old July 15th, 2002, 07:26 PM
quetico
Guest
 
Posts: n/a
Default Re: Re: Some interesting stats

Me???? I figure the info provided on cruisemates IS valid... that's why I took the time to post the stuff....
And you're right about life-changing decisions ..people come here for info, not major markers in life. ... but who ever implied otherwise?
Q
Reply With Quote
  #8  
Old July 16th, 2002, 08:10 AM
Marnie
Guest
 
Posts: n/a
Default Re: Some interesting stats

Sorry Quetico, I didn't address that correctly, I was asking the several other people why they are here if the information Cruisemates survey results was so useless.
Reply With Quote
  #9  
Old July 16th, 2002, 07:48 PM
quetico
Guest
 
Posts: n/a
Default Re: Re: Some interesting stats

Thanks... you did have me puzzled.... now if I can just get these shackels down...
Q
Reply With Quote
  #10  
Old July 17th, 2002, 11:07 AM
AR
Guest
 
Posts: n/a
Default Re: Re: Re: Some interesting stats

Ernie's initial comment is still exactly right: it is what survey researchers call a "self-selected sample" and is therefore of minimal value, some say no value.

CNN and other news organizations take two kinds of polls: one kind is the silly sort of "log onto our website and tell us what you think," which is nothing more than a parlor game. It has no validity whatsoever. The other kind is the poll taken using legitimate researchers with legitimate sampling techniques. This kind of polling is expensive, which is why news organizations often team up to pay the bill and share the results. You can spot these polls because they always quote a "margin of error," which means the sample was selected by the polling company based on demographics. You can't compute a meaningful margin of error for a self-selected sample.

From Quetico's intial explanation of how he "massaged" the raw data, I honestly can't determine what he actually did, but I have a feeling he may have added insult to injury with regard to data analysis technique.

AR
Reply With Quote
  #11  
Old July 17th, 2002, 10:52 PM
pamda pamda is offline
Senior Member
Admiral
 
Join Date: Oct 2000
Posts: 5,238
Default Re: Some interesting stats

I actually own a kid who is a statistician/pollster. The way results can be skewed is HUGE. And fascinating.

__________________
Reply With Quote
  #12  
Old July 18th, 2002, 06:46 AM
quetico
Guest
 
Posts: n/a
Default Re: Re: Re: Re: Some interesting stats

AR, you could be right. I didn't manage the data, I found it and checked out the numbers.
But I guess what you're saying is that if I were to ask 100 people about a cruise, and if 95 of them said it was great, that this is not significant. Personally I think it would have certain implications. My guess is I would probably enjoy the cruise, unless I happened to be one of the other 5. Conversely, if most of the responses inidcate dissatisfaction, I would expect problems.
And to be fair, it's not a self-selected sample... the cruisemates web site is for the good and the bad, so all in all it should balance out.
And one other thing. A margin of error in sampling theory is based on the fact that in a normal distribution, 95% of the population wil be within two standard deviations of the mean. The larger the number sampled, the smaller the standard deviation, and the more accurate the poll is. In a binomial distribution, which is a yes/no type poll, to achieve a margin of error of about 3% for 95% of the time the sample size must be about - roughly - 1150 people. That;s why these three numbers will always be cited in a poll.
The margin of error has nothing to do with demographics, and all polls, however poorly done, will have a margin of error.
And as far as 'massaging' the data goes, it appears it was done as follows. Take the poll for a room where the results are
excellent 50, good 20 average 10, poor 5 terrible 2 .
Ignore the 'average' category, since it can go either way. The number of people answering above the average is 70. The number of people answering below the average is 7. Divide the good by the bad and get 10.
The happier people are, the higher the number.
Mickey mouse approach? Maybe. Why was it done in the first place? Can't say.
But I've sailed on three of the four lines in the poll, and I can say that all the implications seem appropriate, especially with respect to NCL.
Q
Reply With Quote
  #13  
Old July 18th, 2002, 10:15 AM
AR
Guest
 
Posts: n/a
Default Re: Re: Re: Re: Re: Some interesting stats

Most non-scientific polls are done for a very simple reason: to boost advertising revenue for the organization conducting them. The random web polls are there to increase hits on websites, theoretically making the advertising on those sites more valuable.

Demos do have a lot to do with error rates, along with sample size of course, not to mention how the questions are framed and a bunch of other factors too.

There is great debate about the validity of "dropping" the middle point of a five point scale. Many say that it creates significant skewing, and some advocate four-point scales for that very reason.

You also make a huge assumption when you accept that those who respond in a self-selected manner are actually "qualified" to answer (e.g., have taken the cruise). Why would people evaluate things they haven't experienced? I have no clue, but they do.

All that said, there's certainly nothing wrong with looking at this kind of info to see whether there are any grave variations from what you already know (or think you know). If there are some big differences it might be interesting to try to find out why.

AR
Reply With Quote
  #14  
Old July 24th, 2002, 06:53 AM
quetico
Guest
 
Posts: n/a
Default Re: Re: Some interesting stats

I think it was Mark Twain who said there are three kinds of untruths...
lies, damn lies and statistics....
Q
Reply With Quote
Reply

Bookmarks

Similar Threads
Thread Thread Starter Forum Replies Last Post
Body Stats beentravelin Cruise Jokes & Trivia - Just For Laughs 3 January 10th, 2009 12:14 PM
some interesting stats quetico Norwegian Cruise Lines 18 August 11th, 2002 03:35 PM
some interesting stats quetico Princess Cruise Lines 0 July 15th, 2002 07:01 AM
some interesting stats quetico Carnival Cruise Lines 0 July 13th, 2002 07:32 PM


Thread Tools Search this Thread
Search this Thread:

Advanced Search
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump

 

All times are GMT -4. The time now is 11:24 AM.
design by: Themes by Design