Originally Posted by nht
The two "devices" are iPhones and Android phones. Both are in the same category. Why is this hard to understand? Ignore the column that says iPod Touch. It is irrelevant to this comparison.
So you are denying what you said yourself earlier, this is the sentence of yours that I responded to:
Originally Posted by nht
Incorrect. The data gathered is for each phone platform AND the iPod touch.
I guess it doesn't surprise me that you are denying what you said in plain words in your previous post.
Incorrect. What they said is that for all three phone platforms measured about 25% of respondents were under 24.
This has absolutely nothing to do with the percentage of iPhone, Android Phone and WebOS phone users under 24 that answered the survey. Zero, nada, zilch.
You apparently think so, not me. You are the first one to bring statistics of iPod touch purchase and usage into the discussion, I never did. The only thing I ever mentioned before is that I own an iPod (and always have at least one at any time), in an exchange about who is an "apple hater" and such. This is an iPhone
forum with discussions on iPhones
and other phones
The dimensions ARE fixed as much as they can be in a survey of this kind. There is no comparison of iOS vs Android except in your mind. The comparison is between iPhone users and Android users that responded to the survey.
I never compared iOS carte blanche with Android OS as a whole, the only think I ever compared was devices on cellular network
with iOS vs with android. I think I have made this abundantly clear in my previous posts, you might want to read those again before posting.
Yes, this is selection bias and I noted that. It's also done by AdMob which introduces another set of biases.
If you want to know the methodology simply read the article and follow the links. CNN did not hide that from you.
The link to admob piece by businesswire contains no reference to selection bias, or bias of any kind within the design of the study. This is the only portion of the article that referred to the study:
Originally Posted by businessweek
Highlights from the survey and the January 2010 AdMob Mobile Metrics Report include:
91 percent of iPhone users and 88 percent of iPod touch users would recommend their device, compared to 84 percent of Android users and 69 percent of webOS users.
iPod touch owners download an average of 12 applications a month, 37 percent more than iPhone and Android users who download approximately nine new apps. webOS users downloaded an average of six applications per month.
iPod touch users spent an average of 100 minutes per day using applications. webOS users spent an average of 87 minutes per day, followed by Android users at 80 minutes and iPhone users at 79 minutes per day.
73 percent of Android users are male, compared to only 56 percent of iPhone OS users. The average iPhone user is 14 years older than the average iPod touch user of which 78 percent are below the age of 24.
iPhone represented 47 percent of US smartphone usage in AdMob’s network in January 2010, followed by Android, RIM and webOS devices at 39, seven, and three percent, respectively.
AdMob first ran this survey in August 2009 and again six months later, in February 2010. The survey was conducted with 960 respondents over a two week period, spanning consumers on Android, iPhone, iPod touch and webOS devices on the more than 15,000 mobile Web sites and applications in AdMob’s network in February 2010. The survey did not include the RIM platform as AdMob does not currently serve ads into Blackberry applications.
Methodology, no it's not in the article, except for some vague references such as "opt-in" and "in Februrary", which do not stand up in the standards of any academic work.
"All data in the feature section is based on an opt-in survey taken by users on their mobile device. Respondents were sourced by responding to mobile ads throughout AdMob's iPhone OS, Android and webOS networks. There was no incentive offered to participate in the survey.
There were 963 total respondents: 318 Android, 244 iPhone, 356 iPod touch and 45 webOS. The survey was run from February 5th - February 16th.
Of course it was conducted by mobile ads, what do you think admob as a company does? Of course, it also doesn't answer the questions that I posed at all, whether it was conducted when the device was purchases, registered, during installation of an app, visiting certain site, etc, etc.
The geographic representation of the respondents was designed to approximate the distribution of users in the AdMob network. The respondents were sourced from English-speaking countries in the AdMob network. "
Not relevant to our discussion, almost everyone in the industry knows what markets admob serves anyways, so mostly redundant information.
You know, after a certain point you're simply trying to bluster your way past the point that you have no data whatsoever to support your assertion that iPhone users are significantly older than Android users.
My quarrel is not with your iphone vs android phone stats, it's always been with your dragging in another entire category of devices into an irrational comparison, let me refresh your mind on another place where you mentioned (among numerous instances):
Originally Posted by nht
The iPod stuff was highlighted to show that Apple is positioning well for future iPhone users.
I have never said that their iPhone stats were wrong, and in fact, this is what I said about it:
Originally Posted by HardBall
the correct comparison would be to fix all dimensions that are intended to be independent, and then compare devices between iOS and android, which make only the iPhone vs. android Phone comparison valid.
You can't run away from what you said earlier so easily.
They are reporting the demographics of the respondents. You can use this data to extrapolate to the wider population as long as you understand the inherent biases of an opt in survey of users of ad driven apps. Unless you have data to show that there is an inherent age bias this information is probably representative of the age demographics of the entire population for that time period.
You apparently have never done any first hand study involving rigorous statistical methods, or statistical inference, and don't know how terms like "extrapolate", "representative" and "inherent" are normally used in such domains. I rest my case.
Given that AdMob and Nielsen generate these reports for Advertising purchasers do you really believe that
a) There is significant statistical analysis in determining the ages of respondents? (ans: no)
b) They are likely to completely hose up the categorization of age demographics? (ans: unlikely)
Whether the reported demographics matches that of the target population is debatable but not for their uses: the demographics of the eyeballs that view ads.
You are right, it is "debatable", so are the conclusions that you think admob and nielsens have drawn, so what's the quarrel here?
For someone who asserts expertise in statistics you seem not to grasp this fairly basic nature of opt-in surveys...there is always selection bias in opt in surveys.
Why, because the links you have provided does not define the methodology, period. Calling something opt-in survey does nothing to define the methodology. A large portion of all statistical data in the IT industry come from opt-in surveys, it means nearly nothing to state that something is an "opt-in survey".
Given that Nielsen did not provide methodology in their blog I'll go along with their analysis...that the age difference is slight. Not having any desire to pay $$$ for their actual report which lists their collection methodology I'll live with that assessment.
How odd that you require empirical data when you steadfastly provide none of your own while rejecting data that has already been provided.
If you are willing to draw that kind of that kind of conclusion, then you better be willing to pay for that. That and a statistical package such as SAS or SPSS.
Find me a source that shows that iPhone users are predominantly in the 44+ demographic.
Also, you don't need data to make an assertion. That's what makes it an assertion vs a statement of fact.
You are right, my saying that iPhone users are typically older is a conjecture (and is a much simpler conclusion to draw than yours, mine simply involves analysis of two bimodal distributions in the same DF), and based on some small sample of data I see around my own life. Of course it is not a statistically valid conclusion to draw given the small sample that I have, and I never claimed it to be. If I ever have the resources to conduct a large scale data collection, I would, but I'm just an average student. You on the other hand, try to draw some conclusions from ambiguous and imprecise data that does not in any way warrant the type of conclusions that you want, so you simply push aside all guidelines of sound scientific research to forge toward the conclusion that you want.