FlyerTalk Forums

FlyerTalk Forums (https://www.flyertalk.com/forum/index.php)
-   MilesBuzz (https://www.flyertalk.com/forum/milesbuzz-370/)
-   -   Surprising credit card Airpoll (https://www.flyertalk.com/forum/milesbuzz/6624-surprising-credit-card-airpoll.html)

Family flyer Jun 29, 2002 2:59 pm

Surprising credit card Airpoll
 
From Randy's June 28 MilesLink:

How many mileage-earning credit cards do you own?
None 22.22 %
1 44.44 %
2 22.22 %
More than 2 11.11 %

Considering the responses posted in this thread, I'm shocked that almost one-quarter of the people have no mileage-earning card and only one in 10 have more than two cards.

VolleyballFerd Jun 29, 2002 5:01 pm


<font face="Verdana, Arial, Helvetica, sans-serif" size="2">Originally posted by Family flyer:
From Randy's June 28 MilesLink:

How many mileage-earning credit cards do you own?
None 22.22 %
1 44.44 %
2 22.22 %
More than 2 11.11 %

Considering the responses posted in this thread, I'm shocked that almost one-quarter of the people have no mileage-earning card and only one in 10 have more than two cards.
</font>
Don't be too surprised - the actually figures for the 4 categories are:

2 in 9, 2 in 9, 4 in 9 and 1 in 9 - exactly.

In order to get those numbers exactly, you would almost certainly have only a total of 9 respondents. Possibly it could occur with 18 - but even that is a pretty small sample size.

mauld Jun 29, 2002 5:06 pm

As someone who travels 90% on company business and who, if I want to get reimbursed for that travel have to use the company card--I don't have any mileage affinity cards! I'm also of that rare breed that uses cash to pay for items. Yes, I know I can use an affiliated AMEX card, but sorry I don't. While I like acquiring miles/hotel credits and have amassed a fairly large amount through general travel, I do not get quite as excited as some at purchasing just to get miles. Although, that said-- I have been using my regular green Amex to get MR points lately http://www.flyertalk.com/forum/wink.gif

[This message has been edited by mauld (edited 06-30-2002).]

Pickles Jun 30, 2002 9:00 pm


<font face="Verdana, Arial, Helvetica, sans-serif" size="2">Originally posted by VolleyballFerd:
Don't be too surprised - the actually figures for the 4 categories are:

2 in 9, 2 in 9, 4 in 9 and 1 in 9 - exactly.

In order to get those numbers exactly, you would almost certainly have only a total of 9 respondents. Possibly it could occur with 18 - but even that is a pretty small sample size.
</font>
Actually, these percentages would occur with ANY multiple of these fractions. And if your sample was 81, what is wrong with 18/81,18/81,36/81, and 9/81? Any more (or less) likely than 20/90,20/90,40/90 and 10/90? The basic laws of sampling, arithmetic distribution and association apply even in FlyerTalk.

Counsellor Jul 1, 2002 9:38 am


<font face="Verdana, Arial, Helvetica, sans-serif" size="2">Originally posted by Pickles:
Actually, these percentages would occur with ANY multiple of these fractions. &lt;snip&gt; The basic laws of sampling, arithmetic distribution and association apply even in FlyerTalk. </font>
But really, that's the point. While a sample of 900 might come out in one category with a round hundred responses (and thus look the same as "1 out of 9"), what are the odds that the remaining 800 would also all divide into exactly even hundreds? Pretty low, I would submit. Therefore, the fact that all categories are divisible by nine indicate that it is likely that only 9 (or maybe 18) folks have responded.


doc Jul 1, 2002 10:47 am

Perhaps "likely," but not neceassarily, by any means! http://www.flyertalk.com/forum/smile.gif

Nontheless, missing here for sure is the "N" number!

One should always include this "N" value, among other things, so as to ascertain how relatively meaningful, if at all, the data is! http://www.flyertalk.com/forum/smile.gif

Counsellor Jul 2, 2002 1:35 am

Ah, memories (nightmares?) of the Stat course in grad school.

UAL Traveler Jul 2, 2002 7:54 pm


<font face="Verdana, Arial, Helvetica, sans-serif" size="2">Originally posted by Counsellor:
But really, that's the point. While a sample of 900 might come out in one category with a round hundred responses (and thus look the same as "1 out of 9"), what are the odds that the remaining 800 would also all divide into exactly even hundreds? Pretty low, I would submit. Therefore, the fact that all categories are divisible by nine indicate that it is likely that only 9 (or maybe 18) folks have responded.
</font>
Absolutel correct. As the sample size, N, gets large, the probability for any unique distribution of results from an assumed random process goes to zero (but remains finite).


VolleyballFerd Jul 5, 2002 8:57 am


<font face="Verdana, Arial, Helvetica, sans-serif" size="2">Originally posted by Pickles:
The basic laws of sampling, arithmetic distribution and association apply even in FlyerTalk. </font>
Exactly true - which is why I posted in the first place (I won't go into the math, since others have already).

The important comment is from Doc - whenever you come across poll results you be aware of the sample size, and also how (and if) the respondents have been randomly selected from some population. If they haven't, then using those results to make any conclusions about anyone but the specific respondents would be invalid, statistically.

Also, there is often ambiguity in a question, and this always makes results suspect:

How many "mileage earning" credit cards do I own? Depends on if I count my hotel credit cards or only my airline card. To any who think the answer is obvious, then they haven't had much experience in seeing how people respond to survey questions. A lot would depend on the context of the other survey questions, or if the instructions expressly tell you to include or exclude hotel cards.

Main point - be careful when reading published survey results.

dc10forlife Jul 5, 2002 1:37 pm

Why not try another survey, and perhaps also account for those of us who would rather have cash back than miles? I would venture to say that many people have made this decision.

Right now I've got Fleet's 2% cashback rebate card (which has tiered cashback levels). For spending $25,000, I'd rather have the cashback (about $390 w/ Fleet) than having to pay an annual fee in order to get one or possibly two free airline tickets. I figure if I really need the miles, I can just choose to do a mileage run.

Of course, my expirience with redeeming award miles has also never been the best. The dates I want never seem to be available. In fact, right now I've got tons of miles sitting in my account b/c I can never use them!

[This message has been edited by dc10forlife (edited 07-05-2002).]


All times are GMT -6. The time now is 9:03 pm.


This site is owned, operated, and maintained by MH Sub I, LLC dba Internet Brands. Copyright © 2026 MH Sub I, LLC dba Internet Brands. All rights reserved. Designated trademarks are the property of their respective owners.