they did what Gallup had done in 2012 - suddenly switch from reporting "registered voters" to now reporting "likely voters". I.e. (from your link), -
A major difference may be that the previous poll was conducted among registered voters, while the new survey is the first that Quinnipiac did this year of likely voters. (As Quinnipiac notes: “This is the first survey of likely voters in this race by the independent Quinnipiac University Poll, and can not be compared to earlier surveys of registered voters.”)
I remember during the 2012 Presidential election when Gallup did their switch from "registered" to "likely", which caused a similar significant shift, and eventually a later big fail because of who they considered "likely voters", not taking into account that their methodology of using the same old thinking of who was "likely" to vote, could not be carried over and accurately applied in that instance.
This is where pollsters are going to really have to make some decisions and not just go with "past practices" which introduce an automatic bias. The same thing happened (outside of the Russian meddling that had some impact) in 2016, where there were a group of voters who were not factored in as "likely" (those disaffected deplorable Drumpsters, many of whom had never voted before or had even previously voted for Democrats).
After the 2012 fiasco (where Gallup no longer does this poll), Huffington summarized what Gallup said happened -
Misidentification of Likely Voters. Like most other media pollsters, Gallup attempts to determine which respondents are likely to actually vote in November, using a procedure developed in the 1950s that involves a battery of questions, such as how often they’ve voted in the past and how much attention they’re paying to the election. Last year, this likely voter model moved Gallup’s estimate of the margin separating Obama and Romney 4 points in Romney’s direction.
“That, in and of itself,” Newport said, “was at least 1 point more towards Romney than the average of other polls that were using some time of likely voter model.” Thus, “on that sense alone,” he concluded, “they moved things too far in moving the sample towards Mitt Romney.” While Gallup is continuing to investigate its likely voter procedures with a series of experiments in the fall 2013 state elections, it pointed out one big issue in 2012: that many Obama voters said they hadn’t given much thought to the election, removing them from the likely voter pool even though they intended to cast a ballot. Gallup is researching whether it needs to majorly overhaul or even replace the way it identifies likely voters.
Under-Representation of Regions. Gallup also weights its data by a variety of factors, including broad geographical regions such as the Midwest and the South. But each region contains several time zones. Due to differing response rates, Gallup didn’t interview enough people in certain time zones within some regions, effectively undersampling states that vote more Democratic.
Faulty Representation of Race and Ethnicity. As HuffPost first reported in June 2012, Gallup in recent years has used an unusual method to ask about race that distorted the racial composition of its samples when the data were weighted. Unlike most other pollsters, who ask respondents to select from a list of racial and ethnic categories, Gallup asked respondents whether they identified with each racial and ethnic group one by one. This led to a disproportionate number of people who said they were multiracial, and that in turn distorted the weighting procedure, effectively giving too much weight to some white voters.
https://www.huffingtonpost.com/2013/06/04/gallup-poll-2012_n_3384882.html
From Gallup's report (PDF) -
Conclusion: The fundamental assumption justifying likely voter models is that a broad national adult sample, or even registered voter sample, does not reflect the demographic or political profile of the real electorate. As long as the underlying sample is representative of the U.S. public, applying a likely voter model makes sense. Also, historical turnout patterns show that likely voter models should have the effect of increasing support for the Republican candidate. This is how Gallup’s model worked in 2012, and apparently how other firms’ models worked but to a lesser degree.
Nevertheless, nearly all firms underestimated support for Obama. Gallup’s likely voter analysis largely speaks to the difference between Gallup’s likely voter model and other firms’ models, indicating the need for improvements that might be made based on further research. By making the vote choice less Democratic/more Republican, Gallup’s likely voter model performed in the same general way as the models that all other election polling firms used. Gallup’s four-point shift toward Romney, however, was greater than other polls’ shifts. And, various combinations of the use or weighting of likely voter questions would have changed Gallup’s final estimate of the presidential vote to be more in line with other firms’. Thus, Gallup will continue to investigate the model’s performance, both with upcoming validation studies of the 2013 gubernatorial elections, and in the context of additional research on the underlying sample. If changes in the research and election environments have shifted national samples closer to likely voters than has been the case historically, then the likely voter model needs to be adapted to correct for it.
https://news.gallup.com/file/poll/162887/Gallup%202012%20Presidential%20Election%20Polling%20Review.pdf
So bottom line is that we'll have to hope that aggregate polls give a better idea of what is going on. But even then all you have to do is look at what happened in 2016 (although most said it was "close" but the actual result produced literally polar opposite outcomes of what was supposedly "close" when comparing the actual electoral votes vs the popular vote).
And a real tell-tale sign of misjudging "likely voters" and "actual voters", was what happened in Alabama with Doug Jones. Blacks in general would have definitely not been considered "likely voters" -
particularly during a special election based on "past practice" (bias). Yet here we are with Doug Jones elected out of Alabama thanks to an enhanced turnout of black women.
= new reply since forum marked as read
