Well in this case we know that the sample isn't representative, given all the other larger surveys showing dumbphones are still at 40%+ of the total subscriber population.
That doesn't have any bearing on this survey. Dumbphone users could have intentions of purchasing an iPhone too.
Since we're on the subject of surveys, Canalys released their latest results on the smartphone market today.
"Canalys today published its final worldwide country-level Q2 2011 smart phone market estimates, showing substantial market growth in all regions. . . Of the 56 countries Canalys tracks around the world, Android led in 35 of them and achieved a global market share of 48%. Asia Pacific (APAC) remained the largest regional market, with 39.8 million units shipping there, compared with 35.0 million in Europe, the Middle East and Africa (EMEA), and 32.9 million in the Americas.
Android, the number one platform by shipments since Q4 2010, was also the strongest growth driver this quarter, with Android-based smart phone shipments up 379% over a year ago to 51.9 million units. . .
The final country-level data delivered to clients today shows there were particularly strong performances from Android devices in APAC countries, such as South Korea, where Android holds an 85% platform share, and Taiwan, where it has 71%. . .
With shipments of 20.3 million iPhones and a market share of 19%, iOS overtook Nokia?s Symbian platform during the quarter to take second place worldwide. In doing so, Apple also became the world?s leading individual smart phone vendor, stripping Nokia of its long-held leadership position. . .
Microsoft is also eager to see Nokia?s first Windows Phone products, along with those from its other OEM partners, ship with its Mango update. . . Fewer than 1.5 million Microsoft-based smart phones shipped during the quarter, equating to a mere 1% share of the global market, down 52% against shipments a year ago.
Windows Phone OEM partner HTC saw Android driving the vast majority of its portfolio, but Canalys expects it to continue to be a leading provider of Windows Phone products. HTC achieved particular success in North America this quarter, climbing to a 21% share and consolidating its second place position, while Apple?s North American share dropped from 31% last quarter to 25% this quarter.
RIM had a challenging quarter in North America, with its market share slipping to 12%, down from 33% a year ago, leading to negative press coverage in the United States. But RIM?s global shipments grew 11% year on year, keeping it the number one vendor in Latin America with a 28% share. "
That doesn't have any bearing on this survey. Dumbphone users could have intentions of purchasing an iPhone too.
I'm not talking about the second chart (intent), I'm talking about the first - current handset. It's completely relevant, it shows us just how unrepresentative of the US smartphone consumer population the sample was. Android was lower than RIM in the sample which hasn't been true for a long old while.
Comscore and Nielsen independently produce numbers showing very similar levels of Android and iPhone adoption, and those numbers look nothing whatsoever like these.
I'm not talking about the second chart (intent), I'm talking about the first - current handset. It's completely relevant, it shows us just how unrepresentative of the US smartphone consumer population the sample was. Android was lower than RIM in the sample which hasn't been true for a long old while.
Comscore and Nielsen independently produce numbers showing very similar levels of Android and iPhone adoption, and those numbers look nothing whatsoever like these.
Homer Simpson: "Oh, people can come up with statistics to prove anything. 14% of people know that."
Since we're on the subject of surveys, Canalys released their latest results on the smartphone market today.
Interesting numbers, WP7 seems still stuck pining for the fjords. One really interesting bit you didn't include. Seems that a huge portion of Android's gain came from Nokia, but not as much as it could - because Samsung also saw huge Bada growth.
‘Samsung has failed to fully capitalize on Nokia’s weakened state around the world, as the Finnish company rides out a challenging transitional period,’ said Jones. ‘It’s the best placed vendor to grow at Nokia’s expense, taking advantage of its global scale and channel reach, but it hasn’t yet done enough to capitalize on this, particularly in emerging markets.’
Samsung was the largest Android device vendor and the number two vendor overall in the market with shipments of its own-branded devices at 17.0 million units. Its year-on-year growth of 421% was helped by significant growth of 355% in its bada smart phone shipments. Samsung also acts as an ODM for the Google Nexus S and T-Mobile Sidekick 4G, collectively estimated to have shipped 0.7 million units.'
I'm not talking about the second chart (intent), I'm talking about the first - current handset. It's completely relevant, it shows us just how unrepresentative of the US smartphone consumer population the sample was. Android was lower than RIM in the sample which hasn't been true for a long old while.
Comscore and Nielsen independently produce numbers showing very similar levels of Android and iPhone adoption, and those numbers look nothing whatsoever like these.
You've got a point about the charts. I looked at the intentions part and assumed the current ownership part was accurate. Indeed, this survey appears to be very skewed. I wonder how a survey with such an obvious bias would ever make it to the public. Are they unconcerned with credibility?
You have a valid point. Larger sample sizes generally lead to increased precision when estimating unknown parameters. This is a function of measures of central tendencey, specifically standard deviation (SD). However, data collection, reduction, and analysis times increase with the number of study participants. Cost then becomes a factor.
When reporting the results of studies such as this, it's best to stick to descriptive statistics in the reporting. Attempting to generalize the results to a larger population (inferential statistics) would be inappropriate (threats to external validity). This is especially true with small and/or non-representative samples, sampling techinques, and the associated Type VI error.
Thanks for catching that. Should have been a Type IV error. Sometimes I get my mix all talked up...
Type IV - Inconsistency between the statistical prodedure and research question. Ocurrs when the statistical model used has been unrelated or tangentally related to the question(s) of interest.
As an unsolicited FYI:
Type I - The likelyhood of rejecting the null hypothesis (i.e., there is no relationship or significant difference).
Type II - Related to the power of a test, the probability of detecting a difference when one exists.
Type III - Occurs when you get the right answer to the wrong question, sometimes called a Type 0 error. Also occurs when you correctly conclude that the two groups are statistically different, but you are wrong about the direction of the difference.
I think Apple may have under estimated how delaying the release of a next generation phone could hurt sales. Is anyone enthusiastic about buying a phone that is now well more than a year old? Probably not. And sometimes you need to replace your phone -- maybe its because the old phone breaks, or damaged, or is just too old to continue functioning -- and if a model is 15 months old, its probably not at the top of the list for many buyers.
It does make sense to move the iphone to a September cycle, making it available just before the Christmas rush. However, given that every previous iteration of the phone became available in the early summer period, moving the update cycle by 3 (or still possible more) months has left a number of buyers stuck with older phones. If Apple had been anticipating moving their release date, they probably should have considered either a minor update or a price adjustment midway through the year to more accurately reflect the age of the existing iphone 4. Releasing a long delayed white model is not an "update," either.
You've got a point about the charts. I looked at the intentions part and assumed the current ownership part was accurate. Indeed, this survey appears to be very skewed. I wonder how a survey with such an obvious bias would ever make it to the public. Are they unconcerned with credibility?
Well they did explain where and how the sample was acquired, in Minneapolis, so it's a relatively affluent urban demographic.
Just because the sample isn't representative of the full market doesn't mean it offers no insight, it just means we need to be careful how we view it.
I think Apple may have under estimated how delaying the release of a next generation phone could hurt sales. Is anyone enthusiastic about buying a phone that is now well more than a year old? Probably not. And sometimes you need to replace your phone -- maybe its because the old phone breaks, or damaged, or is just too old to continue functioning -- and if a model is 15 months old, its probably not at the top of the list for many buyers.
The problem is that the data is against you. Not only did Apple have their highest ever sales figures for iPhone last quarter, the YoY increase was up too, above levels seen in the previous recent quarters. The iPhone-4 may be old but it still holds up incredibly well compared to upper end Android phones.
The iPhone market seems to be permanently constrained by supply, as fast as markets such as the US move into a supply/demand balance Apple is adding new carriers and new markets.
I'm not saying that this will always be true, or that Apple can switch to 18month product cycles, or anything crazy like that - but the fact is that Apple hasn't paid any kind of price so far for the extended iP-4 lifespan.
Not stated is that if you don't get the results you want, just don't report the results. Reword it or change the population and try again.
Thanks for the informative links. Oftentimes a study yields inconclusive results or unexpected findings. As stated above, it's generally best to abandon the study as attempts to 'fix' things typically result in principle investigator bias. However, the study can serve as a basis for further research.
I'm not American, but to me that title reads that 74% of current iPhone buyers (read owners) are awaiting the new model. But the article states that it is a "significant number of Verizon customers"
When looking at the title I was pretty amazed that 74% of Verizon iPhone owners would renew their phone that quickly, but apparently I misread it. Anyone else as well?
That's definitely how it's phrased, American English or other. Badly phrased.
I'm not American, but to me that title reads that 74% of current iPhone buyers (read owners) are awaiting the new model. But the article states that it is a "significant number of Verizon customers"
When looking at the title I was pretty amazed that 74% of Verizon iPhone owners would renew their phone that quickly, but apparently I misread it. Anyone else as well?
You totally read it wrong...or misinterpreted it. "74% of current iPhone buyers" does NOT mean they are current owners. It means that 74% of potential iPhone buyers are waiting for the next iPhone...
Thanks for catching that. Should have been a Type IV error. Sometimes I get my mix all talked up...
Type IV - Inconsistency between the statistical prodedure and research question. Ocurrs when the statistical model used has been unrelated or tangentally related to the question(s) of interest.
As an unsolicited FYI:
Type I - The likelyhood of rejecting the null hypothesis (i.e., there is no relationship or significant difference).
Type II - Related to the power of a test, the probability of detecting a difference when one exists.
Type III - Occurs when you get the right answer to the wrong question, sometimes called a Type 0 error. Also occurs when you correctly conclude that the two groups are statistically different, but you are wrong about the direction of the difference.
To put the notion of Type 1 and Type 2 errors in remotely the same league as some silly, cute-sy, colloquial, Wikipedia-based, social science researcher catch-phrases ("Type 3" and "Type 4") betrays utter ignorance of the basics of the scientific method of enquiry.
Even given the poor sample I'd say we can still draw from the results that iPhone is proving to be 'stickier' than Android. Not that I think that half of android owners are really going to switch, likely most of them will see the price of iPhone and balk - but that half of the Android owners want to switch when almost none of the iPhone owners did is indicative of some significant differences in satisfaction.
I'd love to see this done with a much bigger sample and also with an indication of the android model owned. If dissatisfaction is mostly in the low-end android owners then it's not relevant because they can't afford to switch platforms anyway. Dissatisfaction at the premium end of android is much more serious. Unfortunately this survey offers nothing beyond speculation on that.
Comments
Well in this case we know that the sample isn't representative, given all the other larger surveys showing dumbphones are still at 40%+ of the total subscriber population.
That doesn't have any bearing on this survey. Dumbphone users could have intentions of purchasing an iPhone too.
"Canalys today published its final worldwide country-level Q2 2011 smart phone market estimates, showing substantial market growth in all regions. . . Of the 56 countries Canalys tracks around the world, Android led in 35 of them and achieved a global market share of 48%. Asia Pacific (APAC) remained the largest regional market, with 39.8 million units shipping there, compared with 35.0 million in Europe, the Middle East and Africa (EMEA), and 32.9 million in the Americas.
Android, the number one platform by shipments since Q4 2010, was also the strongest growth driver this quarter, with Android-based smart phone shipments up 379% over a year ago to 51.9 million units. . .
The final country-level data delivered to clients today shows there were particularly strong performances from Android devices in APAC countries, such as South Korea, where Android holds an 85% platform share, and Taiwan, where it has 71%. . .
With shipments of 20.3 million iPhones and a market share of 19%, iOS overtook Nokia?s Symbian platform during the quarter to take second place worldwide. In doing so, Apple also became the world?s leading individual smart phone vendor, stripping Nokia of its long-held leadership position. . .
Microsoft is also eager to see Nokia?s first Windows Phone products, along with those from its other OEM partners, ship with its Mango update. . . Fewer than 1.5 million Microsoft-based smart phones shipped during the quarter, equating to a mere 1% share of the global market, down 52% against shipments a year ago.
Windows Phone OEM partner HTC saw Android driving the vast majority of its portfolio, but Canalys expects it to continue to be a leading provider of Windows Phone products. HTC achieved particular success in North America this quarter, climbing to a 21% share and consolidating its second place position, while Apple?s North American share dropped from 31% last quarter to 25% this quarter.
RIM had a challenging quarter in North America, with its market share slipping to 12%, down from 33% a year ago, leading to negative press coverage in the United States. But RIM?s global shipments grew 11% year on year, keeping it the number one vendor in Latin America with a 28% share. "
That doesn't have any bearing on this survey. Dumbphone users could have intentions of purchasing an iPhone too.
I'm not talking about the second chart (intent), I'm talking about the first - current handset. It's completely relevant, it shows us just how unrepresentative of the US smartphone consumer population the sample was. Android was lower than RIM in the sample which hasn't been true for a long old while.
Comscore and Nielsen independently produce numbers showing very similar levels of Android and iPhone adoption, and those numbers look nothing whatsoever like these.
I'm not talking about the second chart (intent), I'm talking about the first - current handset. It's completely relevant, it shows us just how unrepresentative of the US smartphone consumer population the sample was. Android was lower than RIM in the sample which hasn't been true for a long old while.
Comscore and Nielsen independently produce numbers showing very similar levels of Android and iPhone adoption, and those numbers look nothing whatsoever like these.
Homer Simpson: "Oh, people can come up with statistics to prove anything. 14% of people know that."
Since we're on the subject of surveys, Canalys released their latest results on the smartphone market today.
Interesting numbers, WP7 seems still stuck pining for the fjords. One really interesting bit you didn't include. Seems that a huge portion of Android's gain came from Nokia, but not as much as it could - because Samsung also saw huge Bada growth.
‘Samsung has failed to fully capitalize on Nokia’s weakened state around the world, as the Finnish company rides out a challenging transitional period,’ said Jones. ‘It’s the best placed vendor to grow at Nokia’s expense, taking advantage of its global scale and channel reach, but it hasn’t yet done enough to capitalize on this, particularly in emerging markets.’
Samsung was the largest Android device vendor and the number two vendor overall in the market with shipments of its own-branded devices at 17.0 million units. Its year-on-year growth of 421% was helped by significant growth of 355% in its bada smart phone shipments. Samsung also acts as an ODM for the Google Nexus S and T-Mobile Sidekick 4G, collectively estimated to have shipped 0.7 million units.'
Source Canalys
I'm not talking about the second chart (intent), I'm talking about the first - current handset. It's completely relevant, it shows us just how unrepresentative of the US smartphone consumer population the sample was. Android was lower than RIM in the sample which hasn't been true for a long old while.
Comscore and Nielsen independently produce numbers showing very similar levels of Android and iPhone adoption, and those numbers look nothing whatsoever like these.
You've got a point about the charts. I looked at the intentions part and assumed the current ownership part was accurate. Indeed, this survey appears to be very skewed. I wonder how a survey with such an obvious bias would ever make it to the public. Are they unconcerned with credibility?
You have a valid point. Larger sample sizes generally lead to increased precision when estimating unknown parameters. This is a function of measures of central tendencey, specifically standard deviation (SD). However, data collection, reduction, and analysis times increase with the number of study participants. Cost then becomes a factor.
When reporting the results of studies such as this, it's best to stick to descriptive statistics in the reporting. Attempting to generalize the results to a larger population (inferential statistics) would be inappropriate (threats to external validity). This is especially true with small and/or non-representative samples, sampling techinques, and the associated Type VI error.
What's a 'Type VI' error?
I'm pretty certain I could survey another 216 people and come up with whatever results I'd like to see.
If you're 'pretty certain,' you should do it and report back.
What's a 'Type VI' error?
Thanks for catching that. Should have been a Type IV error. Sometimes I get my mix all talked up...
Type IV - Inconsistency between the statistical prodedure and research question. Ocurrs when the statistical model used has been unrelated or tangentally related to the question(s) of interest.
As an unsolicited FYI:
Type I - The likelyhood of rejecting the null hypothesis (i.e., there is no relationship or significant difference).
Type II - Related to the power of a test, the probability of detecting a difference when one exists.
Type III - Occurs when you get the right answer to the wrong question, sometimes called a Type 0 error. Also occurs when you correctly conclude that the two groups are statistically different, but you are wrong about the direction of the difference.
It does make sense to move the iphone to a September cycle, making it available just before the Christmas rush. However, given that every previous iteration of the phone became available in the early summer period, moving the update cycle by 3 (or still possible more) months has left a number of buyers stuck with older phones. If Apple had been anticipating moving their release date, they probably should have considered either a minor update or a price adjustment midway through the year to more accurately reflect the age of the existing iphone 4. Releasing a long delayed white model is not an "update," either.
You've got a point about the charts. I looked at the intentions part and assumed the current ownership part was accurate. Indeed, this survey appears to be very skewed. I wonder how a survey with such an obvious bias would ever make it to the public. Are they unconcerned with credibility?
Well they did explain where and how the sample was acquired, in Minneapolis, so it's a relatively affluent urban demographic.
Just because the sample isn't representative of the full market doesn't mean it offers no insight, it just means we need to be careful how we view it.
I think Apple may have under estimated how delaying the release of a next generation phone could hurt sales. Is anyone enthusiastic about buying a phone that is now well more than a year old? Probably not. And sometimes you need to replace your phone -- maybe its because the old phone breaks, or damaged, or is just too old to continue functioning -- and if a model is 15 months old, its probably not at the top of the list for many buyers.
The problem is that the data is against you. Not only did Apple have their highest ever sales figures for iPhone last quarter, the YoY increase was up too, above levels seen in the previous recent quarters. The iPhone-4 may be old but it still holds up incredibly well compared to upper end Android phones.
The iPhone market seems to be permanently constrained by supply, as fast as markets such as the US move into a supply/demand balance Apple is adding new carriers and new markets.
I'm not saying that this will always be true, or that Apple can switch to 18month product cycles, or anything crazy like that - but the fact is that Apple hasn't paid any kind of price so far for the extended iP-4 lifespan.
If you're 'pretty certain,' you should do it and report back.
I'm pretty certain you're intelligent enough to understand how to slant a survey to get the results you want. I don't need to prove that one too do I?
Some articles on surveys, and the ways they can be manipulated, sometimes inadvertantly.
http://stattrek.com/ap-statistics-2/...ling-bias.aspx
http://fitbiztools.com/2011/06/how-y...urvey-results/
http://en.wikipedia.org/wiki/Opinion_poll
Not stated is that if you don't get the results you want, just don't report the results.
I'm pretty certain you're intelligent enough to understand how to slant a survey to get the results you want. I don't need to prove that one too do I?
Some articles on surveys, and the ways they can be manipulated, sometimes inadvertantly.
http://stattrek.com/ap-statistics-2/...ling-bias.aspx
http://fitbiztools.com/2011/06/how-y...urvey-results/
http://en.wikipedia.org/wiki/Opinion_poll
Not stated is that if you don't get the results you want, just don't report the results.
Thanks for the informative links. Oftentimes a study yields inconclusive results or unexpected findings. As stated above, it's generally best to abandon the study as attempts to 'fix' things typically result in principle investigator bias. However, the study can serve as a basis for further research.
I'm not American, but to me that title reads that 74% of current iPhone buyers (read owners) are awaiting the new model. But the article states that it is a "significant number of Verizon customers"
When looking at the title I was pretty amazed that 74% of Verizon iPhone owners would renew their phone that quickly, but apparently I misread it. Anyone else as well?
That's definitely how it's phrased, American English or other. Badly phrased.
I'm not American, but to me that title reads that 74% of current iPhone buyers (read owners) are awaiting the new model. But the article states that it is a "significant number of Verizon customers"
When looking at the title I was pretty amazed that 74% of Verizon iPhone owners would renew their phone that quickly, but apparently I misread it. Anyone else as well?
You totally read it wrong...or misinterpreted it. "74% of current iPhone buyers" does NOT mean they are current owners. It means that 74% of potential iPhone buyers are waiting for the next iPhone...
I'm pretty certain you're intelligent enough to understand how to slant a survey to get the results you want. I don't need to prove that one too do I?
And, you're smart enough to know that consistent cynicism sucks.
And, you're smart enough to know that consistent cynicism sucks.
Absolutely agree
Thanks for catching that. Should have been a Type IV error. Sometimes I get my mix all talked up...
Type IV - Inconsistency between the statistical prodedure and research question. Ocurrs when the statistical model used has been unrelated or tangentally related to the question(s) of interest.
As an unsolicited FYI:
Type I - The likelyhood of rejecting the null hypothesis (i.e., there is no relationship or significant difference).
Type II - Related to the power of a test, the probability of detecting a difference when one exists.
Type III - Occurs when you get the right answer to the wrong question, sometimes called a Type 0 error. Also occurs when you correctly conclude that the two groups are statistically different, but you are wrong about the direction of the difference.
To put the notion of Type 1 and Type 2 errors in remotely the same league as some silly, cute-sy, colloquial, Wikipedia-based, social science researcher catch-phrases ("Type 3" and "Type 4") betrays utter ignorance of the basics of the scientific method of enquiry.
Enough said.
Absolutely agree
Even given the poor sample I'd say we can still draw from the results that iPhone is proving to be 'stickier' than Android. Not that I think that half of android owners are really going to switch, likely most of them will see the price of iPhone and balk - but that half of the Android owners want to switch when almost none of the iPhone owners did is indicative of some significant differences in satisfaction.
I'd love to see this done with a much bigger sample and also with an indication of the android model owned. If dissatisfaction is mostly in the low-end android owners then it's not relevant because they can't afford to switch platforms anyway. Dissatisfaction at the premium end of android is much more serious. Unfortunately this survey offers nothing beyond speculation on that.