Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
What would the margin of error be for a random survey of 200 people from a population of 10-15 million? 5-7%?

Say around 6% with an confidence level of around 90%. In survey terms that is bad. That would be with just 2 answers.
Fact that there are more here increases it it and the standard deviation would be a lot higher. The spread is to high, it bad survay.

When doing a random sample of people it requires a lot more requirements to make sure your sample is good.

People in Texas will give much different results on average than people in California. Or Big city vs small city.

200 people is not enough to even get a handle on getting those variables under control.
 
Did this survey really need to happen? I've yet to meet anyone dissatisfied with something brand new they were waiting over a year for and only had one option to select from. Only the truly brave are willing to try out something not iPhone when they had iPhone; if they were coming from something else (Android, feature) who is going to admit failure?

The problem here is the contracting required for these phones. You BETTER be completely satisfied, because late 2013 is a long time to be with something you hate.

----------

Even if you disagree with the sample size, is it really that hard to imagine? It's an iPhone FFS.

https://www.macrumors.com/2011/09/0...rankings-of-smartphone-consumer-satisfaction/

Not a surprise. You'll get the same results with a sample size of 2000 people. How can you not? 1) It's an iPhone, 2) It's an improvement over the iPhone 4, 3) it has Siri.

Connect the dots.

Please start ending your posts with /sarcasm because no rational human being could actually believe what you just wrote.
 
MacRumors is also a skewed sampling. In fact EVERY forum is a skewed sampling. So 100 posts about bad battery life may seem like the whole world, but it's not.

Fandroids look like they are taking over the world, but surprise! they're not. They're just all on the forums blabbing away, all day long, making it much more annoying and useless to even read these comments anymore. Thanks for reading mine though! :)

Don't confuse actual statistics with informal polls.

You're trying to argue that hard data that shows Android as the world's most used smartphone operating system is flawed? Stop being in data denial.
 
200 of say 10-15 million users is accurate enough. Consider when Gallup does a survey they typically use a sample size of 1000 to represent a country of 300 million with a ±4% margin of error. According to their data increasing the sample size beyond 1000 only has a marginal improvement in accuracy.

The size of the total population isn't relevant to how big your sample size needs to be. It's defined by the variability in the population how big your sample needs to be to accurately reflect the population.
 
Did this survey really need to happen? I've yet to meet anyone dissatisfied with something brand new they were waiting over a year for and only had one option to select from. Only the truly brave are willing to try out something not iPhone when they had iPhone; if they were coming from something else (Android, feature) who is going to admit failure?

The problem here is the contracting required for these phones. You BETTER be completely satisfied, because late 2013 is a long time to be with something you hate.

----------



Please start ending your posts with /sarcasm because no rational human being could actually believe what you just wrote.

http://www.gottabemobile.com/2011/09/13/iphone-4-sales-still-strong/

Now add a faster processor + Siri.

Winner.

You don't need to be a statistician to figure this stuff out.
 
Last edited:
Wirelessly posted (Mozilla/5.0 (iPhone; CPU iPhone OS 5_0_1 like Mac OS X) AppleWebKit/534.46 (KHTML, like Gecko) Version/5.1 Mobile/9A405 Safari/7534.48.3)

I'm 100% satisfied with my 4S even though this is my first iPhone :)
 
Say around 6% with an confidence level of around 90%. In survey terms that is bad. That would be with just 2 answers.
Fact that there are more here increases it it and the standard deviation would be a lot higher. The spread is to high, it bad survay.

When doing a random sample of people it requires a lot more requirements to make sure your sample is good.

People in Texas will give much different results on average than people in California. Or Big city vs small city.

200 people is not enough to even get a handle on getting those variables under control.

Hmm. You know all of the terms, but you (obviously) can't be bothered to actually do the calculations.

Given the numbers we have, you can determine the confidence interval for any given confidence level for the result given the population size and sample size.

With a population of 15,000,000, and a sample size of 200, a 96% response means:
  • 95% confidence that between 93.28% and 98.72% are satisfied with their iPhone 4s
  • 99% confidence that between 92.43% and 99.57% are satisfied with their iPhone 4s
Looks like a pretty solid survey to me.
 
With a 95% confidence interval, the margin of error is about 6.9% for a sample size of 200 on this large of a population.

The standard in most industries to to target a 5% margin of error at a 95% confidence interval.

The sample size would need to be 385 for industry standard statistical significance. However, this analysis is fairly close and I'd be confident that it is directionally accurate so long as the sample population was not biased in any way.
 
Last edited:
Hmm. You know all of the terms, but you (obviously) can't be bothered to actually do the calculations.

Given the numbers we have, you can determine the confidence interval for any given confidence level for the result given the population size and sample size.

With a population of 15,000,000, and a sample size of 200, a 96% response means:
  • 95% confidence that between 93.28% and 98.72% are satisfied with their iPhone 4s
  • 99% confidence that between 92.43% and 99.57% are satisfied with their iPhone 4s
Looks like a pretty solid survey to me.

Do the math.

It is a 90% that the result is with in 90%-100%. 6% error is a pretty wide margin of error.

Also that would be on a much more control sample group of say manufacturing.
Human surveying is harder to do. You have a large geographic area to cover with a range of populations and 200. This is a bad survey in how check.
It is nothing more than a spot check. Not really valid than a quick check.
used this to sit for the quick cals.
http://www.custominsight.com/articles/random-sample-calculator.asp

It has been a long time since I seen the formulas. It not really much good than say you wanted something about the same of NYC. It is worthless for the entire nation.
 
Hmm. You know all of the terms, but you (obviously) can't be bothered to actually do the calculations.

Given the numbers we have, you can determine the confidence interval for any given confidence level for the result given the population size and sample size.

With a population of 15,000,000, and a sample size of 200, a 96% response means:
  • 95% confidence that between 93.28% and 98.72% are satisfied with their iPhone 4s
  • 99% confidence that between 92.43% and 99.57% are satisfied with their iPhone 4s
Looks like a pretty solid survey to me.

You can't say that without knowing how they implemented the survey. 200 is an incredibly small number for a company that does this as their business.
 
Dissatisfied customer

I've been a mac user since the Power PC days, as well as had all versions of the iPhone, and have worked as a previous Apple employee. I disappointingly have to say that I utterly regret purchasing the iPhone 4s. I loved everything about the previous iPhone 4, however 32GB was insufficient for my needs, and upgraded to the iPhone 4s. From day one, I had issues. Much poorer battery life, and a phone that was supposed to be twice as fast, kept getting stuck in applications, and became unresponsive for 20-30 seconds at a time. When going to take a photo, that usually means missing the event that I wanted to capture. Apple has replaced my phone 3 times already, with no remedy to these issues. I'm hoping that it is not hardware related, and can be fixed in a software update. But when you want to throw your phone against a brick wall almost every time I go to use the iPhone, that's horrible. Only if I hadn't sold my iPhone 4; I would return this device in a heartbeat.
 
I have seen people who switch from Android to iPhone but never someone who switched from iPhone to Android. I think it says a lot about iPhone itself :)
 
LOL at America.

These people are locked into two-year contracts. Psychologically, how can someone justify being unsatisfied when they have potentially 24 months at up to $100/mo left to pay?

Ask the prepaid users in Europe (who paid 749€ upfront) how satisfied they are and you won't get the same answer, you'll get a more accurate one.
 
There should be a "super-dee-duper ultra mega satisfied" option for complete accuracy!
 
LOL at America.

These people are locked into two-year contracts. Psychologically, how can someone justify being unsatisfied when they have potentially 24 months at up to $100/mo left to pay?

Ask the prepaid users in Europe (who paid 749€ upfront) how satisfied they are and you won't get the same answer, you'll get a more accurate one.

unlocked iPhones are sold everywhere, just as well as 2 years contracts. I don't get your point for Europe v. US.
There is an easy way to check your last point, it's the return rates.
 
I'm 96% satisfied with mine.

I want to be able to buy iTether to connect it to my MacBook when I'm on the road.
 
Do the math.

It is a 90% that the result is with in 90%-100%. 6% error is a pretty wide margin of error.

Also that would be on a much more control sample group of say manufacturing.
Human surveying is harder to do. You have a large geographic area to cover with a range of populations and 200. This is a bad survey in how check.
It is nothing more than a spot check. Not really valid than a quick check.
used this to sit for the quick cals.
http://www.custominsight.com/articles/random-sample-calculator.asp

It has been a long time since I seen the formulas. It not really much good than say you wanted something about the same of NYC. It is worthless for the entire nation.

His numbers were correct.

This site has a decent calc: http://www.surveysystem.com/sscalc.htm

Your calc doesn't take into account the percentage of people answering 4 or 5 vs those answering 1, 2, or 3. This changes the accuracy of the calculation.

At 95% confidence, the interval is 2.72%, so it's +-2.72% on 96%. That means at 95% confidence, the spread is 5.44%. At 99%, it's 7.14%, or +-3.57%. That's a reasonably large spread in market research, but still puts us above 90% rated a 4 or 5 in either case, which makes over 90% of people at least somewhat satisfied with the iPhone regardless. This is accurate so long as the sample is entirely random, but I don't know how they selected their sample.

Source: I worked for a market research firm for five years.
 
unlocked iPhones are sold everywhere, just as well as 2 years contracts. I don't get your point for Europe v. US.
There is an easy way to check your last point, it's the return rates.

Let me know how well an iPhone4S without a contract works on AT&T, T-Mobile, Verizon or Sprint, i.e. the four large carriers in the US.

The survey is ridiculous, at best, because it can't separate the physical iP4S from the service in the US, whereas that disconnect can be made in the rest of the world.

----------

it's the return rates.

that's an even worse metric, because people who are unlikely to be satisfied won't purchase in the first place.
 
Let me know how well an iPhone4S without a contract works on AT&T, T-Mobile, Verizon or Sprint, i.e. the four large carriers in the US.

The survey is ridiculous, at best, because it can't separate the physical iP4S from the service in the US, whereas that disconnect can be made in the rest of the world.

This is correct. For the most part, it tells us how satisfied 4S owners with AT&T, Verizon, and Sprint are, as I'm sure there aren't a whole ton of unlocked 4S phones out there yet.
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.