Comments on: Survey: 6 percent of U.S. teens own iPhones

By: jawandapuck, Vancouver, BC

jawandapuck, Vancouver, BC — Fri, 11 Apr 2008 21:14:32 +0000

Even the author misinterpreted the numbers when he facetiously said, "…or at least these 389 teens — still have money to spend at the Apple Store,". He forgot that it is *6%* of 389 teens (or so says the "survey").

ex ped: Good catch! Fixed now. Thanks.

By: Bill, Willilamsburg, VA

Bill, Willilamsburg, VA — Fri, 11 Apr 2008 19:09:05 +0000

A sample of 384 that has 6% as the sample statistic produces a 95% confidence interval of from 3.6% to 8.4%, not 1% to 11%. if the sample had been of 1500 teens, that would have reduced the interval to 4.8% to 7.2%. Quadrupling the sample size halves the interval. The comments that question the selection method for those in the survey are much more relevant than questions about the sample size. It takes a pretty big sample to get the margin of error down below =/- 1% for a 95% confidence interval. This from a statistics instructor.

By: Alan, Dallas, TX

Alan, Dallas, TX — Thu, 10 Apr 2008 16:49:13 +0000

Steve from CT appears to be correct that the margin of error here appears to be +/-5%. As stated here, you don't need a huge or large sample base to create a statistic but if you want reliability and accuracy, the larger the sample the better. With a survey saying 6% own iPhones and there's such a large margin of error like there is here, it's nearly meaningless and certainly doesn't really tell us anything we didn't already know. We know teens have iPhones…we still have no idea how many own them. Another point is that as close to the original report that I've seen on this used the term "high school students" and also "teens." If there are 26M teenagers, then fine. But the U.S. DoE says there are 16.4M high school students. To me it's a moot point. Choosing a population of 16.4M vs an input of "infinity" comes out identical when plugging in the sample sizes and the 95% confidence level, etc. The sample is too small to give reliable results. I haven't seen the original report but what's been reported doesn't include even the margin of error. That's a "no-no" for reporting survey statistics in the news. Without that info being transparently reported by those releasing the data, the survey should be trashed. (it is possible that the full report gives that info…if so…why is it absent in the public form?…seems possible that the large margin of error would have not have attracted much attention from those seeking useful information)

By: Dave, Akron, Ohio

Dave, Akron, Ohio — Thu, 10 Apr 2008 12:51:58 +0000

Sample size accepted, but, is the sample a representative sample of the population. Does it cross economic, location, etc. lines to cast a broad net in obtaining the sample? Or is it concentrated in an economic niche that would generate the result give?

By: Steve, Rocky Hill CT

Steve, Rocky Hill CT — Wed, 09 Apr 2008 21:40:57 +0000

If you calculate the sample size based on a population of 20million with a confidence level of 95% and a confidence interval of 5…the sample size needed is 384. Based on this calculation the authors should have stated that were are 95% confident that 6% of US teens own an Iphone +/-5% (1% to 11%). I work as a market research analyst and I am very comfortable with calculating sample sizes.

By: Vince, California

Vince, California — Wed, 09 Apr 2008 17:12:50 +0000

I hope I read it wrong, but it seemed to favor Apple in this article as an advertising promotion. Anyone got paid from Apple?

A survey is only a survey and it seems to leave out the logistics of the survey. I am only surprised that CNN would post something like this.

By: Guy Lavallee, Edmonton, Canada

Guy Lavallee, Edmonton, Canada — Wed, 09 Apr 2008 16:13:18 +0000

"Teens flock to iPhones"

What do you expect? Teens are stupid.

By: Dave, Cleveland, Ohio

Dave, Cleveland, Ohio — Wed, 09 Apr 2008 15:19:25 +0000

No surprised that the younger crowd goesfor the Iphone since they've already got the Ipod, etc. It would be a natural customer base for Apple products. But since most teenagers depend upon their parents to obtain phones (like my boys), I would anticipate the percentage of teenagers that have cell phones that are not Iphones would dwarf the 6%.

By: Shane, Denver, Colorado

Shane, Denver, Colorado — Wed, 09 Apr 2008 14:54:52 +0000

While 389 is a sample survey, and it can reasonably justify a larger trend across a population (statistically speaking), it should be noted that a small sample group correlates to higher sampling errors. When the sample group size approches the population group size (i.e. teens in the U.S.), in theory this should bring the sampling error to near zero or to zero if testing the entire population. So, in effect the numbers the researcher is finding are not inaccurate, they are just not telling the whole story. I would say that the reason 6% doesn't add up to the actual U.S. teen population is due to a sampling error. In this case due to a variety of variables that are more than likely not factored (rural, urban, city, suburb, age, etc.). This in effect creates a ridiculous sampling error. For this reason I would have a hard time buying into what this says.

By: Huh!

Huh! — Wed, 09 Apr 2008 14:13:01 +0000

It is not surprising to me that many teens bought iPhones – considering how many have jobs after school/on weekends. My wife is an 11th grade teacher in a middle-income district, and the kids eerily reflect this statistic almost to the percentage point. More want to buy the Touch, but paying for the monthly $60 iPhone bill is no sweat…