Contributors   |   Messages   |   Polls   |   Resources   |  
Comments
Newest First | Oldest First | Threaded View
Page 1 / 2   >   >>
Ariella
Ariella
12/14/2016 8:16:38 PM
User Rank
Author
Re: at some time in the future, everyone is virtual!
<But for industry polls, I'd like to maybe see more meta-analysis that looks back on how poll predictions stand up to reality. There have been all kinds of predictions for tech, and only Moore's Law seems to be true generally. The "trough of disillusionment" always seems to hit when marketing says it won't.....> @mhhf1ve I agree that there would be value in that. It seems to me that many can make all kinds of predictions with impunity because no one seems to come back and say, "You know, the numbers you said we'd see are very far from what happened."  They only do that with political predictions about winners because there is much less time elapsed between the call and the result. 

50%
50%
mhhf1ve
mhhf1ve
12/14/2016 6:57:41 PM
User Rank
Platinum
Re: at some time in the future, everyone is virtual!
> "Polls are not reliable, Ghani explained, for a number of reasons."

Thanks for the extended reply, Ariella. Polling is indeed very tricky to do "right" -- and even when it's done with the best of methodologies, it's not necessarily predictive in the way you might want it to be -- especially for US presidential elections, due to the electoral college and other complicating factors.

But for industry polls, I'd like to maybe see more meta-analysis that looks back on how poll predictions stand up to reality. There have been all kinds of predictions for tech, and only Moore's Law seems to be true generally. The "trough of disillusionment" always seems to hit when marketing says it won't.....

50%
50%
batye
batye
12/10/2016 6:25:05 PM
User Rank
Platinum
Re: at some time in the future, everyone is virtual!
@mhhf1ve I could not agree more... as it could be easy manupalited to give need outcome on paper... but reality is other story... I keep reading - https://www.wired.com/2016/08/man-behind-trumps-bid-finally-take-digital-seriously/ 

but always ask what is really true story behind... we would never know... I'm not say good or bad... I'm only say real life is more interesting than expected...

50%
50%
batye
batye
12/10/2016 5:54:25 PM
User Rank
Platinum
Re: at some time in the future, everyone is virtual!
@Ariella interesting observation :) in my books all the polls numbers could look good but on site it could be other story... In my world before invest it better to see physical location and do due diligence on site :) - How I see it... but I'm looking from the point of security...

50%
50%
JohnBarnes
JohnBarnes
12/9/2016 2:46:12 PM
User Rank
Platinum
Re: at some time in the future, everyone is virtual!
Ariella, With a stratified sample the goal is to produce roughly equivalent confidence levels across all the strata being combined in the final result. This will often require drastically different sized samples that are not at all proportional to the populations of the strata. Your example is extreme, but yes, it can be correct and necessary to barely sample New York and sample like crazy in Delaware. Another way polling and voting are different is that one person one vote is not really a principle of polling. Polls are not supposed to be fair; they're supposed to be accurate. That requires a different rule: maximize information per added participant at the margin.

50%
50%
Ariella
Ariella
12/9/2016 8:57:34 AM
User Rank
Author
Re: at some time in the future, everyone is virtual!
<
Polls within industries or markets are a different matter, of course, because large parts of the general population aren't represented, but it's a pretty good overall indicator.> @JohnBarnes and, of course, there's no electoral college issue at play.  Industry polls do vary in intnet sometimes -- not just to show overall trends but to show differences in different regions/countries. So I think in those cases, you'd have to check that each region actually is adequately represented and not skewed in some way that puts in just 10 for one area and 900 for another.

50%
50%
JohnBarnes
JohnBarnes
12/8/2016 10:45:34 PM
User Rank
Platinum
Re: at some time in the future, everyone is virtual!
ariella,

In the "good old days" for pollsters in the US, 1200 people was enough because there were only about 66 kinds of people with a 90% conformity -- i.e. 66 demographic groups, 90% of the members of which would have preferences aligned with the group's.  So all you needed was around 15-20 members of the demographic to get pretty close for each group, and then just multiplied them by appropriate population weights.

That was true about 1956-2004, which is to say from Gallup's really crude but brilliant early work up tlll almost today.

Last I saw, it was looking like you needed more like 300 groups and almost none of them had better than 70% conformity, which means you need more like 35-40 group members to get the same accuracy.  300 x 40 = 12,000, so you need to poll about ten times as many subjects.

Polls within industries or markets are a different matter, of course, because large parts of the general population aren't represented, but it's a pretty good overall indicator.

100%
0%
JohnBarnes
JohnBarnes
12/8/2016 10:37:19 PM
User Rank
Platinum
Re: at some time in the future, everyone is virtual!
mhhf1ve,

The poll analyst community is in fact going more than a little nuts about the reasons for its failures, because so many of the proposed causes are turning out not to work when the data is reanalyzed. You might say that at the moment they are really out of ideas.

I really think it's going to turn out to be pure bien pensant -- a very large part of the population says what they think the polltaker wants to hear. The equivalent of that for products? Well, in the imaginations of the socially aspiring, do the "right people" like butter, superhero movies, exercise machines, or hip hop more or less than they are "supposed" to?

50%
50%
Ariella
Ariella
12/8/2016 5:09:10 PM
User Rank
Author
Re: at some time in the future, everyone is virtual!
@mhhf1ve Unfortunately the choices one gets on these pull down menus do not include something like "reading purely for research and not for possible purchases." That's another problem with how things are framed. I try to answer the questions as honestly as I can, but they don't really let me clarify that they would be wasting their time when making a sales call. With Brexit and Trump there were other factors at play, I believe. With the former, of course, the vote was very close, which is actually something I predicted, though I didn't know which way it would go. With the latter, people lost sight of why polls are inaccurate.

A few months before the elections, I heard a lecture by Rayid Ghani, currently the Director of the Center for Data Science & Public Policy at the University of Chicago. He served as the Chief Scientist for Obama for America in 2012 re-election campaign.

Polls are not reliable, Ghani explained, for a number of reasons. One, any poll with a sample size that is under several thousand people doesn't cover enough of a range to be taken as an accurate represent. Two, even larger polls don't necessarily signify an outcome if they cover the whole country rather than offer a breakdown by state.  Finally, even a poll that is generally accurate would have some margin of error.

The people behind presidential elections campaigns aren't supposed to look at polls; they look at the data that will help them achieve their goal. That's because -- and these were his very words --the campaign is not about winning the most votes overall but winning at least 217 electoral votes. Ultimately, the swing states' electoral votes determine which candidate wins. 

100%
0%
mhhf1ve
mhhf1ve
12/8/2016 4:59:50 PM
User Rank
Platinum
Re: at some time in the future, everyone is virtual!
That just adds more data to suggest that these surveys are highly inaccurate-- since they use self-reported responses and may rely on the respondents to verify their validity as decision makers in the field. Polling is tough. I guess that's why polls didn't predict Trump or Brexit.

100%
0%
Page 1 / 2   >   >>


Latest Articles
Italy's 5G auction could exceed a government target of raising €2.5 billion ($2.9 billion) after attracting interest from companies outside the mobile market.
The emerging-markets operator is focusing on the humdrum business of connectivity and keeping quiet about some of its ill-fated 'digitalization' efforts.
Three UK has picked Huawei over existing radio access network suppliers Nokia and Samsung to build its 5G network.
Vendor says that it's its biggest 5G deal to date.
Verizon skates where the puck is going by waiting for standards-based 5G devices to launch its mobile service in 2019.
On-the-Air Thursdays Digital Audio
Orange has been one of the leading proponents of SDN and NFV. In this Telco Transformation radio show, Orange's John Isch provides some perspective on his company's NFV/SDN journey.
Special Huawei Video
10/16/2017
Huawei Network Transformation Seminar
The adoption of virtualization technology and cloud architectures by telecom network operators is now well underway but there is still a long way to go before the transition to an era of Network Functions Cloudification (NFC) is complete.
Video
The Small Cell Forum's CEO Sue Monahan says that small cells will be crucial for indoor 5G coverage, but challenges around business models, siting ...
People, strategy, a strong technology roadmap and new business processes are the key underpinnings of Telstra's digital transformation, COO Robyn ...
Eric Bozich, vice president of products and marketing at CenturyLink, talks about the challenges and opportunities of integrating Level 3 into ...
Epsilon's Mark Daley, director of digital strategy and business development, talks about digital transformation from a wholesale service provider ...
Bill Walker, CenturyLink's director of network architecture, shares his insights on why training isn't enough for IT employees and traditional ...
All Videos
Telco Transformation
About Us     Contact Us     Help     Register     Twitter     Facebook     RSS
Copyright © 2024 Light Reading, part of Informa Tech,
a division of Informa PLC. All rights reserved. Privacy Policy | Cookie Policy | Terms of Use
in partnership with