PollQuant Poll Shows Mrs. Clinton Leads by 2.5 Points in Presidential Election - Nov. 7, 2016

PollQuant "Modeled" Poll of Polls Result (4-Way)

November 7, 2016

Mrs. Clinton 44.65 (Modeled)

Mr. Trump 42.10 (Modeled)

Mrs. Clinton leads by 2.55 Points -- a.k.a. Statistical Tie

(Click on the image to enlarge)

Worth Knowing

1. How does PollQuant "Modeled" Poll of Polls differ from the other Poll of Polls? 
PollQuant models the national polls while the competition uses an average. 

2. Why is PollQuant Superior? 
Modeling reduces the impact of the divergent issues**, making the comparison apples-to-apples. PollQuant does not apply any judgment, instead lets the model decide.

3. Is PollQuant forward looking? 
No, it's backward-bending, with polls thru 11-06-2016, but could be used as a predictive tool or a validation tool to authenticate an independent poll. 

4. Does PollQuant collect its own survey data? 
No, PollQuant models the outcomes and attributes of the national polls.

5. Does PollQuant model all the national polls? 
No, PollQuant models those within the most recent 4-to-7 day survey period. As we head into the poll, the time period tightens.

6. Do you disclose your Model values?
No, PollQuant model values are strictly internal and are used to recompute the cumulative adjusted poll values as shown above.

7. What are the different issues that PollQuant attempts to handle? 
Here are some of the Issues** that PollQuant model tends to address:

--- Sample Size (750 vs. 3200) 

--- Sample Bias (over/under-sampled category)

--- Sample Make-up (likely vs registered / source of population stratification)

--- Sample Period (one day can make a difference / major news skews polls)

--- Sample Collection Method (telephone interviews vs. online data collection)

--- Sample Margin of Error (statistical tie 46 vs. 44 w/MoE 2; 46 vs. 41 w/MoE 5) 

--- Sample Outlier Results (Candidate 1: Poll 1 shows +4 while Poll 6 shows -7)

--- Plus other modeling issues... 

By accessing this Blog you agree to our Disclaimer (click here to read)

Press Release

http://pollquant.blogspot.com/2016/11/press-release-launching-pollquant.html

PollQuant Poll Shows Mrs. Clinton Leads by 2 Points in Presidential Election - Nov. 6, 2016

PollQuant "Modeled" Poll of Polls Result (4-Way)

November 6, 2016

Mrs. Clinton 45.06 (Modeled)

Mr. Trump 42.73 (Modeled)

Mrs. Clinton leads by 2 (2.33) Points -- a.k.a. Statistical Tie

(Click on the image to enlarge) 

Worth Knowing

1. How does PollQuant "Modeled" Poll of Polls differ from the other Poll of Polls? 
PollQuant models the national polls while the competition uses an average. 

2. Why is PollQuant Superior? 
Modeling reduces the impact of the divergent issues**, making the comparison apples-to-apples. PollQuant does not apply any judgment, instead lets the model decide.

3. Is PollQuant forward looking? 
No, it's backward-bending, with polls thru 11-05-2016, but could be used as a predictive tool or a validation tool to authenticate an independent poll. 

4. Does PollQuant collect its own survey data? 
No, PollQuant models the outcomes and attributes of the national polls.

5. Does PollQuant model all the national polls? 
No, PollQuant models those within the most recent 4-to-7 day survey period. As we head into the poll, the time period tightens.

6. Do you disclose your Model values?
No, PollQuant model values are strictly internal and are used to recompute the cumulative adjusted poll values as shown above.

7. What are the different issues that PollQuant attempts to handle? 
Here are some of the Issues** that PollQuant model tends to address:

--- Sample Size (750 vs. 3200) 

--- Sample Bias (over/under-sampled category)

--- Sample Make-up (likely vs registered / source of population stratification)

--- Sample Period (one day can make a difference / major news skews polls)

--- Sample Collection Method (telephone interviews vs. online data collection)

--- Sample Margin of Error (statistical tie 46 vs. 44 w/Moe 2; 46 vs. 41 w/Moe 5) 

--- Sample Outlier Results (Candidate 1: Poll 1 shows +4 while Poll 6 shows -7)

--- Plus other modeling issues... 

By accessing this Blog you agree to our Disclaimer (click here to read)

PollQuant Poll Shows Mrs. Clinton Leads by 2 Points in Presidential Election - Nov. 5, 2016

PollQuant "Modeled" Poll of Polls Result (4-Way)

November 5, 2016

Mrs. Clinton 45.14 (Modeled)

Mr. Trump 42.84 (Modeled)

Mrs. Clinton leads by 2 (2.30) Points -- a.k.a. Statistical Tie

(Click on the image to enlarge)

Worth Knowing

1. How does PollQuant "Modeled" Poll of Polls differ from the other Poll of Polls? 
PollQuant models the national polls while the competition uses an average. 

2. Why is PollQuant Superior? 
Modeling reduces the impact of the divergent issues**, making the comparison apples-to-apples. PollQuant does not apply any judgment, instead lets the model decide.

3. Is PollQuant forward looking? 
No, it's backward-bending, with polls thru 11-04-2016, but could be used as a predictive tool or a validation tool to authenticate an independent poll. 

4. Does PollQuant collect its own survey data? 
No, PollQuant models the outcomes and attributes of the national polls.

5. Does PollQuant model all the national polls? 
No, PollQuant models those within the most recent 4-to-7 day survey period. As we head into the poll, the time period tightens.

6. Do you disclose your Model values?
No, PollQuant model values are strictly internal and are used to recompute the cumulative adjusted poll values as shown above.

7. What are the different issues that PollQuant attempts to handle? 
Here are some of the Issues** that PollQuant model tends to address:

--- Sample Size (750 vs. 3200) 

--- Sample Bias (over/under-sampled category)

--- Sample Make-up (likely vs registered / source of population stratification)

--- Sample Period (one day can make a difference / major news skews polls)

--- Sample Collection Method (telephone interviews vs. online data collection)

--- Sample Margin of Error (statistical tie 46 vs. 44 w/Moe 2; 46 vs. 41 w/Moe 5) 

--- Sample Outlier Results (Candidate 1: Poll 1 shows +4 while Poll 6 shows -7)

--- Plus other modeling issues... 

By accessing this Blog you agree to our Disclaimer (click here to read)

PollQuant Poll Shows Mrs. Clinton Leads by 1 Point in Presidential Election - Nov. 4, 2016

PollQuant "Modeled" Poll of Polls Result (4-Way)

November 4, 2016

Mrs. Clinton 44.15 (Modeled)

Mr. Trump 43.17 (Modeled)

Mrs. Clinton leads by 1 (0.98) Point -- a.k.a. Statistical Tie

Alternate Scenario

Without the Reuters Poll

Mrs. Clinton 44.17 (Modeled)

Mr. Trump 43.71 (Modeled)

Mrs. Clinton's lead drops to 1/2 (0.46) Point -- a.k.a. Statistical Tie

(Click on the image to enlarge)

 Worth Knowing

1. How does PollQuant "Modeled" Poll of Polls differ from the other Poll of Polls? 
PollQuant models the national polls while the competition uses an average. 

2. Why is PollQuant Superior? 
Modeling reduces the impact of the divergent issues**, making the comparison apples-to-apples. PollQuant does not apply any judgment, instead lets the model decide.

3. Is PollQuant forward looking? 
No, it's backward-bending, with polls thru 11-03-2016, but could be used as a predictive tool or a validation tool to authenticate an independent poll. 

4. Does PollQuant collect its own survey data? 
No, PollQuant models the outcomes and attributes of the national polls.

5. Does PollQuant model all the national polls? 

6. Do you disclose your Model values?No, PollQuant models those within the most recent 4-to-7 day survey period. As we head into the poll, the time period tightens.

No, PollQuant model values are strictly internal and are used to recompute the cumulative adjusted poll values as shown above.

7. What are the different issues that PollQuant attempts to handle? 
Here are some of the Issues** that PollQuant model tends to address:

--- Sample Size (750 vs. 3200) 

--- Sample Bias (over/under-sampled category)

--- Sample Make-up (likely vs registered / source of population stratification)

--- Sample Period (one day can make a difference / major news skews polls)

--- Sample Collection Method (telephone interviews vs. online data collection)

--- Sample Margin of Error (statistical tie 46 vs. 44 w/Moe 2; 46 vs. 41 w/Moe 5) 

--- Sample Outlier Results (Candidate 1: Poll 1 shows +4 while Poll 6 shows -7)

--- Plus other modeling issues... 

By accessing this Blog you agree to our Disclaimer (click here to read)

Press Release: Launching PollQuant – A "Modeled" Poll of Polls

Press Release
https://www.prlog.org/12599105

ORLANDO, Fla. - Nov. 3, 2016 - PRLog -- Sid Som, owner and publisher of prominent data analytics and info sites like ChefQuant, Homequant, TownAnalyst and LocValu is pleased to announce the launch of PollQuant, a scientifically "modeled" poll of polls.

While a couple of poll of polls exists, they are average-based, without a scientific noise-reduction mechanism. Since PollQuant is modeled, it significantly reduces the incidences of noise and influences of outliers. When at least 15 major polling organizations are competing for visibility and recognition, polarizations would be expected -- for example, with one tracking poll registering +4 while another showing -10 for the same candidate during the same survey period -- thus negatively impacting the consistency and reliability of the average-based poll of poll.

Similarly, divergent sample sizes like 750 vs. 3200, unsound sampling bias leading to over/under sampling of likely voting groups, conflicting survey methods, e.g., direct telephonic contact vs. online polling, differing methodologies like moving averages vs. simple averages, etc. often make the component polls apples-to-oranges. Moreover, arithmetic mean-based calculations help perpetuate the influence of outlier outcomes, causing more volatility and instability in day-to-day movements.

On the other hand, when those competing polls' attributes are modeled out in a single equation in a statistically significant manner, the component polls largely converge to the same results, laying the foundation for a more commensurable comparative platform and paving the way for more reliable and explainable outcomes. Furthermore, since the model realigns outcomes of outlier polls, no special judgment or treatment would be needed in handling or removing them before the final line-up is decided upon.
Though the traditional polls of polls are inherently backward-bending, PollQuant is not. Based on the concept and practice of predictive modeling, PollQuant produces results that could easily be used to forecast some short-term outcomes. In fact, it may not be an exaggeration to assume that when the poll pundits understand the underlying concept and usefulness of PollQuant, they will use it as a predictive tool to validate their own internal results.

In a nutshell, PollQuant serves multiple purposes. First and foremost, it's the first statistically significant poll of polls. Secondly, it's also an independent predictive poll, without the subjective nuances of an atomic poll, making it a perfect addition to a short line-up to strengthen its reliability. Finally, it could be used as a validation poll to authenticate one's own internal polling results.

In a recent conversation, Sid Som, the inventor of PollQuant, stated, "With the introduction of our modeled poll of polls invention, we are adding quantitative science to an age-old market mechanism which was in desperate need of modernization. Now, pundits can comfortably rely on our poll of polls without having to worry if certain outlier polls skewing the overall metallurgy of the overly simplistic average-based outcomes. Similarly, our invention makes issues like sample illiquidity, bias, methodology, margin of error, etc. irrelevant. Our modeling process equalizes all of that, creating a truly commensurable comparative platform. I have no doubt that in the not-too-distant future our invention will force the poll of polls landscape to reinvent itself with a dose of science."

http://pollquant.blogspot.com/

PollQuant Poll Shows a Statistical Tie in Presidential Election - Nov. 3, 2016

PollQuant "Modeled" Poll of Polls Result (4-Way)

November 3, 2016

Mrs. Clinton 43.60 
Mr. Trump 43.57 
Statistical Tie 

(Click on the image to enlarge)

Worth Knowing

1. How does PollQuant "Modeled" Poll of Polls differ from the other Poll of Polls? 
PollQuant models the national polls while the competition uses an average. 

2. Why is PollQuant Superior? 
Modeling reduces the impact of the divergent issues**, making the comparison apples-to-apples. PollQuant does not apply any judgment, instead lets the model decide.

3. Is PollQuant forward looking? 
No, it's backward-bending, with polls thru 11-02-2016, but could be used as a predictive tool or a validation tool to authenticate an independent poll. 

4. Does PollQuant collect its own survey data? 
No, PollQuant models the outcomes and attributes of the national polls.

5. Does PollQuant model all the national polls? 
No, PollQuant models those within the most recent 4-to-7 day survey period. As we head into the poll, the time period will tighten.

6. Do you disclose your Model values?
No, PollQuant model values are strictly internal and are used to recompute the cumulative adjusted poll values as shown above.

7. What are the different issues that PollQuant attempts to handle? 
Here are some of the Issues** that PollQuant model tends to address:

--- Sample Size (750 vs. 3200) 

--- Sample Bias (over/under-sampled category)

--- Sample Make-up (likely vs registered / source of population stratification)

--- Sample Period (one day can make a difference / major news skews polls)

--- Sample Collection Method (telephone interviews vs. online data collection)

--- Sample Margin of Error (statistical tie 46 vs. 44 w/Moe 2; 46 vs. 41 w/Moe 5) 

--- Sample Outlier Results (Candidate 1: Poll 1 shows +4 while Poll 6 shows -7)

--- Plus other modeling issues... 

By accessing this Blog you agree to our Disclaimer (click here to read)

PollQuant Poll Shows Mrs. Clinton Leads by 1.82 Points in Presidential Election - Nov. 2, 2016

PollQuant "Modeled" Poll of Polls Result (4-Way)

November 2, 2016

Mrs. Clinton 44.60 
Mr. Trump 42.78
Mrs. Clinton Leads by 1.82 points

(Click on the image to enhance)

Worth Knowing

1. How does PollQuant "Modeled" Poll of Polls differ from the other Poll of Polls? 
PollQuant models the national polls while the competition uses an average. 

2. Why is PollQuant Superior? 
Modeling reduces the impact of the divergent issues**, making the comparison apples-to-apples. PollQuant does not apply any judgment, instead lets the model decide.

3. Is PollQuant forward looking? 
No, it's backward-bending, with polls thru 11-01-2016, but could be used as a predictive tool or a validation tool to authenticate an independent poll.

4. Does PollQuant collect its own survey data? 
No, PollQuant models the outcomes and attributes of the national polls.

5. Does PollQuant model all the national polls? 

6. Do you disclose your Model values?
No, PollQuant models those within the most recent 4-to-7 day survey period. As we head into the poll, the time period will tighten.

No, PollQuant model values are strictly internal and are used to recompute the cumulative adjusted poll values as shown above.

7. What are the different issues that PollQuant attempts to handle? 
Here are some of the Issues** that PollQuant model tends to address:

--- Sample Size (750 vs. 3200) 

--- Sample Bias (over/under-sampled category)

--- Sample Make-up (likely vs registered / source of population stratification)

--- Sample Period (one day can make a difference / major news skews polls)

--- Sample Collection Method (telephone interviews vs. online data collection)

--- Sample Margin of Error (statistical tie 46 vs. 44 w/Moe 2; 46 vs. 41 w/Moe 5) 

--- Sample Outlier Results (Candidate 1: Poll 1 shows +4 while Poll 6 shows -7)

--- Plus other modeling issues... 

By accessing this Blog you agree to our Disclaimer (click here to read)