John Navas <spamfilter0@navasgroup.com> wrote in
news:9bus039p4obrkdqqgap15icv90jnp35vui@4ax.com:
> On Fri, 30 Mar 2007 23:17:11 -0700, SMS <scharf.steven@geemail.com>
> wrote in <460dd2b7$0$27237$742ec2ed@news.sonic.net>:
>
lf the area of the lower 48, with
>>many variations within the region.
>>
>>Even though the Consumer's Union survey only targeted Consumer's
>>Reports subscribers, any differences between subscribers and the
>>population at large cancelled out, since subscribers were only rating
>>their own carrier's quality.
>
> Nope. Statistics don't work that way. CU surveys (as I've explained
> several times) suffer from serious flaws:
Oh, sweet Jesus, here we go again.
> * Non-representative population
They are not aliens, therefore they are a very representative population.
> * Self-selected sample
Just as any survey is. I'm not aware of a survey where a gun is held to
the head of the participantys to force them to take the survewy against
their will.
> * Small sample size when broken down by region
Untrue.
> * Improper methods (e.g., lumping together D-AMPS and GSM,
> ENS and non-ENS, iDEN and CDMA2000)
Impropoer by whose definition? Yours doesn't count.
> * Poor methodology
You don't have a clue about what you're talking about- what a surprise.
You have no practical experience on the subject, so why don't you just shut
your hole and wrap yourself up in your CIngular blanket.
Scott wrote:
> John Navas <spamfilter0@navasgroup.com> wrote in
> news:9bus039p4obrkdqqgap15icv90jnp35vui@4ax.com:
>
>> On Fri, 30 Mar 2007 23:17:11 -0700, SMS <scharf.steven@geemail.com>
>> wrote in <460dd2b7$0$27237$742ec2ed@news.sonic.net>:
>>
> lf the area of the lower 48, with
>>> many variations within the region.
>>>
>>> Even though the Consumer's Union survey only targeted Consumer's
>>> Reports subscribers, any differences between subscribers and the
>>> population at large cancelled out, since subscribers were only rating
>>> their own carrier's quality.
>> Nope. Statistics don't work that way. CU surveys (as I've explained
>> several times) suffer from serious flaws:
>
> Oh, sweet Jesus, here we go again.
>
>
>> * Non-representative population
>
> They are not aliens, therefore they are a very representative population.
>
>> * Self-selected sample
>
> Just as any survey is. I'm not aware of a survey where a gun is held to
> the head of the participantys to force them to take the survewy against
> their will.
The key is that any bias automatically cancels out because the
respondents are only rating their own service. No one believes that a
Cingular respondent is more likely to be more critical of Cingular, than
a Verizon respondent is likely to be critical of Verizon.
>> * Small sample size when broken down by region
>
> Untrue.
This is where Navas really loses all credibility. Even broken down by
region, the sample size is still large enough for an extremely small
margin of error. The problem is that he doesn't understand statistics.
>> * Improper methods (e.g., lumping together D-AMPS and GSM,
>> ENS and non-ENS, iDEN and CDMA2000)
>
> Impropoer by whose definition? Yours doesn't count.
It's a representative sample of users of each carrier. The fact that it
mixes in users of various technologies is fallout from the fact that
these carriers often have multiple technologies in use.
>> * Poor methodology
>
> You don't have a clue about what you're talking about- what a surprise.
> You have no practical experience on the subject, so why don't you just shut
> your hole and wrap yourself up in your CIngular blanket.
The one good thing about the CU surveys is that the statistical
methodology they use is very sound.
Of course who needs independent, unbiased surveys like the CU survey,
when we have the single sample Navas survey, LOL.
On Sat, 31 Mar 2007 13:46:48 -0700, SMS <scharf.steven@geemail.com>
wrote in <460ec8c4$0$27190$742ec2ed@news.sonic.net>:
>Scott wrote:
>> John Navas <spamfilter0@navasgroup.com> wrote in
>> news:9bus039p4obrkdqqgap15icv90jnp35vui@4ax.com:
>>> * Self-selected sample
>The key is that any bias automatically cancels out because the
>respondents are only rating their own service. No one believes that a
>Cingular respondent is more likely to be more critical of Cingular, than
>a Verizon respondent is likely to be critical of Verizon.
Sorry, but statistics doesn't work that way. Worse, you're making an
assumption without anything to back up or support it.
>>> * Small sample size when broken down by region
>This is where Navas really loses all credibility. Even broken down by
>region, the sample size is still large enough for an extremely small
>margin of error. The problem is that he doesn't understand statistics.
Sorry again, but you have that backwards. The sampling error is
actually large relative to the reported differences.
>>> * Improper methods (e.g., lumping together D-AMPS and GSM,
>>> ENS and non-ENS, iDEN and CDMA2000)
>It's a representative sample of users of each carrier. The fact that it
>mixes in users of various technologies is fallout from the fact that
>these carriers often have multiple technologies in use.
That fallout invalidates the results, much like saying the average
person in the USA has one breast.
>>> * Poor methodology
>The one good thing about the CU surveys is that the statistical
>methodology they use is very sound.
Just the opposite.
>Of course who needs independent, unbiased surveys like the CU survey,
>when we have the single sample Navas survey, LOL.
JDPowers is quite a bit more credible.
You have a personal agenda (anti-GSM in general, anti-Cingular in
particular, as your repeated trolling here makes clear), and seize upon
anything that seems to support it, no matter how unsound.
--
Best regards, SEE THE FAQ FOR CINGULAR WIRELESS AT
John Navas <http://en.wikibooks.org/wiki/Cingular_Wireless_FAQ>
John Navas <spamfilter0@navasgroup.com> wrote in
news:rog21395retu2bauoo22rm4i45tg90evrq@4ax.com:
> On Sat, 31 Mar 2007 13:46:48 -0700, SMS <scharf.steven@geemail.com>
> wrote in <460ec8c4$0$27190$742ec2ed@news.sonic.net>:
>
>>Scott wrote:
>>> John Navas <spamfilter0@navasgroup.com> wrote in
>>> news:9bus039p4obrkdqqgap15icv90jnp35vui@4ax.com:
>
>>>> * Self-selected sample
>
>>The key is that any bias automatically cancels out because the
>>respondents are only rating their own service. No one believes that a
>>Cingular respondent is more likely to be more critical of Cingular,
than
>>a Verizon respondent is likely to be critical of Verizon.
>
> Sorry, but statistics doesn't work that way. Worse, you're making an
> assumption without anything to back up or support it.
Actually, they do work that way. Anybody with experience would know
that, newbie.
>
>>>> * Small sample size when broken down by region
>
>>This is where Navas really loses all credibility. Even broken down by
>>region, the sample size is still large enough for an extremely small
>>margin of error. The problem is that he doesn't understand statistics.
>
> Sorry again, but you have that backwards. The sampling error is
> actually large relative to the reported differences.
Not noticeably larger than any other survey available on the subject.
Try again, neophyte.
>
>>>> * Improper methods (e.g., lumping together D-AMPS and GSM,
>>>> ENS and non-ENS, iDEN and CDMA2000)
>
>>It's a representative sample of users of each carrier. The fact that
it
>>mixes in users of various technologies is fallout from the fact that
>>these carriers often have multiple technologies in use.
>
> That fallout invalidates the results, much like saying the average
> person in the USA has one breast.
You are the only one claiming this- no statistician would exclude any
portion of the customer base when measuring satisfaction with a brand or
product.
>
>>>> * Poor methodology
>
>>The one good thing about the CU surveys is that the statistical
>>methodology they use is very sound.
>
> Just the opposite.
Only in your eyes, child.
>
>>Of course who needs independent, unbiased surveys like the CU survey,
>>when we have the single sample Navas survey, LOL.
>
> JDPowers is quite a bit more credible.
Yeah- they take money from the very people they survey. A good way to
build credibility.
>
> You have a personal agenda (anti-GSM in general, anti-Cingular in
> particular, as your repeated trolling here makes clear), and seize
upon
> anything that seems to support it, no matter how unsound.
>
Re: Replacement Phone Quality HAH! 
Linear Mode
