Posted on 10/29/2018 11:23:35 AM PDT by jmaroneps37
In this final week before Election Day we will see an endless stream of “Hillary Clinton has a 95% chance of winning,” type polls.
Nevertheless, the numbers say otherwise.
The recently released Public Policy Institute (PPCI) of California poll of California voters makes this case.
These results were predictable.
They provided headline readers with the false impression that was intended; and did a very clever job of burying what they didn’t want readers to know.
It predictably told us the Democrat gubernatorial candidate was on his way to an easy win.
What PPCI worked hard to hide was far more important.
They said in a generic survey by 65/23 California’s Hispanics favored Democrats to run Congress.
Based on what their report hid, this is highly implausible.
On page 9 of their 30 page report, all the way at the end of a series of data points, PPCI whispered that by 49/44 Hispanics living in 11 identified battleground Congressional districts, generically prefer a Republican Congress.
This is earth shaking news that, when properly applied to several House races in California, threatens more than one dozen Democrat House members’ reelection.
Only on page 22 in some decidedly small print does the report name the 11 CDs they referred to on page 9.
Those CDs were 4, 7, 10, 16, 21, 25, 39, 45, 48, 49, and 50.
Two of these are Democrat seats.
There are two numbers used to describe each district that indicate how the electorate is self- reporting they will vote in this cycle; but as will become apparent one number will overcome the other.
The first number is the district’s D rating which indicates how Democrat the district has voted in the last two presidential elections relative to the popular vote percentage averaged by Democrat presidential candidates in 2012 and 2016.
So, if a district has a D+3 rating, on average, it has voted 3% higher than the average total vote of the Democrat candidates in the last two general elections.
Based on PPCI’s survey the second important number for each Democrat is the Hispanic population in his/her district.
PPCI might think they can soften the blow for Democrats by downplaying the 49/44 generic ballot among the Hispanics in the 11 battleground CDs favoring Republicans.
Nevertheless, the truth is that with such a large sample (Hispanics in these 11 CDs make up 18% of all of California’s Hispanics) it is statically impossible to ignore the magnitude of this 49/44 result favoring Republicans.
There is no statistically sound reason for PPCI to assume they are correct when they assert the survey numbers they “found” showing by 65/23 all of the rest of California’s Hispanics favor a Democrat Congress is correct.
There are 15 million Hispanics in California.
To conduct an accurate survey of this group a pollster would need responses from about 6,000 voters for each one million in the group.
Since the entire cohort is 15 million, a sample size of 90,000 respondents is required.
The Hispanics living in the 11 battleground Congressional districts number about 2.8 million which far exceeds the number needed to assert even a 99% confidence level as a sample group for 15 million people.
So if CCPI had a 99% confidence level result supporting the 49/44 split among Hispanics in the 11 CDs ,favoring Republicans, it must extend this to all Hispanics in California.
If not the burden is on CCPI to explain why the 11 CD Hispanics are so different from the rest of California’s Hispanics; and why we should care about the statewide 65/23 split favoring Democrats when it runs so counter to their own large sample subset data.
Moreover it makes no sense given data on Hispanic voting which is already publicly available.
PPCI’s report of California’s Hispanic’s 49/44 favorable generic split toward Republicans adds to a growing body of evidence that in this cycle Hispanic voters are shifting their allegiance toward Republicans.
Hispanic approval of President Donald Trump has moved from 28% support in 2016 to 45% today.
In Florida a leftist poll found the Republican gubernatorial candidate up 11 points among Hispanics who make up 27% of the population.
In a special election in Texas last month a Republican who lost an election for the same State Senate seat by 15 points in 2016, won by 6 points against a Democrat Latino former Congressman from the same area.
The seat had been in Democrat hands for 139 years.
The district is 68% Hispanic.
A coterminous US House district held by a Republican was once called a “toss up”; but now the Republican is leading by 18 points.
In September a Probolsky poll found the Democrat gubernatorial candidate in California was up just 44/39 and leading only 35/34 among Hispanics with 30% “undecided.”
“Undecided” is a dodge to avoid saying out loud how a person will vote.
By all indications, these California Hispanics are not “undecided.”
For the sake of fairness we will divide the 7 points of undecided generic numbers equally and add 3.5% to each side; and use a 52.5%/ 47.5% generic split among California’s Hispanics favoring Republicans.
In each of the following cases the D rating of the Democrat candidate is less than the assumed 47.5% remaining Hispanic support.
Combined with 52.5% Republican support for his/her opponent provides an obstacle too great to overcome for several Democrats’ whose past victories have relied too heavily/or completely upon Hispanic votes.
This assumption is correct because in these districts the Hispanic vote is so irreplaceable it cannot be replaced by any other group.
Now we can see the effect of this Hispanic Shift in voter allegiance at work in California.
In CD 3, Democrat John Garamindi who has a D+5 rating and is in a district that is 30% Hispanic could lose.
Subtracting 52.5% from 30% leaves just 14.25% of Hispanic support.
Losing 52.5% Hispanic support and being left with just 14.25% Hispanic support for this Democrat is more than the strength of a D+5 can make up for because it indicates the D rating was too dependent upon the now lost Hispanic support.
In CD 7, Democrat Ami Bera who has a D+3 rating and is in a district with an 18% Hispanic population could lose.
Subtracting 52.5% from 18% leaves just 8.55% of Hispanic support for this Democrat.
Losing 52.5% Hispanic support and being left with just 8.55% Hispanic support for this Democrat is more than the strength of a D+3 can make up for because it indicates the D rating was too dependent upon the now lost Hispanic support.
In CD 9, Democrat Jim Costa who has a D+9 rating is in a district with a 61% Hispanic population could lose.
Losing 52.5% Hispanic support and being left with just 28.975% Hispanic support for this Democrat is more than the strength of a D+9 can make up for because it indicates the D rating was too dependent upon the now lost Hispanic support.
In CD 16, Democrat Jerry McNerney who has a D+8 rating is in a district with a 37% Hispanic population could lose.
Losing 52.5% Hispanic support and being left with just 17.575% Hispanic support for this Democrat is more than the strength of a D+8 can make up for because it indicates the D rating was too dependent upon the now lost Hispanic support.
In CD 24, Democrat Salud Carabajal who has a D+7 rating is in a district with a 36% Hispanic population could lose.
Losing 52.5% Hispanic support and being left with just 17.1% Hispanic support for this Democrat is more than the strength of a D+7 can make up for because it indicates the D rating was too dependent upon the now lost Hispanic support.
In CD 26, Democrat Julia Brownley who has a D+7 rating is in a district with a 45% Hispanic population could lose.
Losing 52.5% Hispanic support and being left with just 21.375% Hispanic support for this Democrat is more than the strength of a D+7 can make up for because it indicates the D rating was too dependent upon the now lost Hispanic support.
In CD 31, Democrat Pete Aguilar who has a D+8 rating is in a district with a 53% Hispanic population could lose.
Losing 52.5% Hispanic support and being left with just 25.175% Hispanic support for this Democrat is more than the strength of a D+8 can make up for because it indicates the D rating was too dependent upon the now lost Hispanic support.
In CD 32, Grace Napolitano who has a D+17 rating is in a district with a 61% Hispanic population could lose.
Losing 52.5% Hispanic support and being left with just 28.975% Hispanic support for this Democrat is more than the strength of a D+17 can make up for because it indicates the D rating was too dependent upon the now lost Hispanic support.
In CD 35, Democrat Norma Torres who has a D+19 rating is in a district with a 69% Hispanic population could lose.
Losing 52.5% Hispanic support and being left with just 32.775% Hispanic support for this Democrat is more than the strength of a D+19 can make up for because it indicates the D rating was too dependent upon the now lost Hispanic support.
In CD 36, Democrat Raul Ruiz who has a D+17 rating is in a district with a 60% Hispanic population could lose.
Losing 52.5% Hispanic support and being left with just 28.5% Hispanic support for this Democrat is more than the strength of a D+17 can make up for because it indicates the D rating was too dependent upon the now lost Hispanic support.
In CD 38, Linda Sanchez who has a D+17 rating is in a district with a 36% Hispanic population could lose.
Losing 52.5% Hispanic support and being left with just 17.1% Hispanic support for this Democrat is more than the strength of a D+17 can make up for because it indicates the D rating was too dependent upon the now lost Hispanic support.
In CD 40, Democrat Lucille Roybal who has a D+33 rating is in a district with a 88% Hispanic population could lose.
Losing 52.5% Hispanic support and being left with just 41.8% Hispanic support for this Democrat is more than the strength of a D+33 can make up for because it indicates the D rating was too dependent upon the now lost Hispanic support.
In CD 41, Democrat Mark Takano who has a D+12 rating is in a district with a 62% Hispanic population could lose.
Losing 52.5% Hispanic support and being left with just 29.45% Hispanic support for this Democrat is more than the strength of a D+12 can make up for because it indicates the D rating was too dependent upon the now lost Hispanic support.
In CD 46, Democrat Lou Correa who has a D+15 rating is in a district with a 66% Hispanic population could lose.
Losing 52.5% Hispanic support and being left with just 31.35% Hispanic support for this Democrat is more than the strength of a D+15 can make up for because it indicates the D rating was too dependent upon the now lost Hispanic support.
In CD 47, Democrat Alan Lowenthal who has a D+13 rating is in a district with a 34% Hispanic population could lose.
Losing 52.5% Hispanic support and being left with just 16.15% Hispanic support for this Democrat is more than the strength of a D+13 can make up for because it indicates the D rating was too dependent upon the now lost Hispanic support.
This is all good news for Republicans.
It would really be news if only half the predictions would come true.
I would imagine that Hispanic’s would loose their Hollyweird support.
A Unicorn will will win the Kentucky Derby before a California Democrats like Roybal loses her seat. Her father had that seat in the early 1960’s.
I’m kinda shocked that Coach Collins is so statistically illiterate. Yes, this is good news, however. GOP might save 2-5 seats that RCP/NYT thik will go dem... if this poll is to be believed more than the NYT poll. (And by the way, extrapolating from subpopulations is VERY dangerous.)
Not mentioned in the OP but there’s a conservative GOP Hispanic, Omar Navarro, running against Maxine Waters who represents CA-43, S. Central L.A., Inglewood etc.
I would love to see Navarro win.
“I’m kinda shocked that Coach Collins is so statistically illiterate. Yes, this is good news, however. GOP might save 2-5 seats that RCP/NYT thik will go dem... if this poll is to be believed more than the NYT poll. (And by the way, extrapolating from subpopulations is VERY dangerous.)”
So an 18% sample size is a “dangerously” low subsection?
There is no way I can compete with your “feelings” and NYT and RCP ( both fake and can be) so I yield. BTW you do know that the NYT sometimes makes 60,000 calls to get 250 “likely voters” right? How do you think that works? Here’s a hint: THEY KEEP CALLING UNTIL THEY GET THE ANSWERS THEY ARE LOOKING FOR.
Brilliantly succinct tweet I read today, perhaps it applies here:
Remember kids: Trust the Bookies, not the Pollsters.
-PJ
>> So an 18% sample size is a “dangerously” low subsection? <<
THAT is exactly what I mean by the statistical illiteracy in this article. No, there was no “18% sample.” That would mean successfully conducting six million interviews. You could probably call everyone in the entire state of California, and reach about 400,000 people, if you’re lucky.
>> Here’s a hint: THEY KEEP CALLING UNTIL THEY GET THE ANSWERS THEY ARE LOOKING FOR. <<
No, they keep calling until they reach 250 “likely voters” who are willing to be interviewed by them. (And of course, they don’t publicize their results until they reach 500.) The incorrectness of a NY Times poll will come from two sources: (1) When there is a 4% margin of error, that means that a 50-42 result could actually be a 46-46 result. In other words, an 8-point landslide could be mere statistical noise. (2) Turnout modeling. If they believe 43% of voters will be Democrats, and they get only 35% Democrats, they regress their polls accordingly.
The Bookies were almost as bad. The betters are just reading the polls by election day.
This is interesting information, we will see what, if anything, it amounts to next Tuesday.
I would love to see some large surprises out of California Congressional districts, but I am not willing to bet the farm on it happening.
However, on the economy, the lower parts of the economy are FINALLY seeing some improvement after 2 administrations of neglect, and this may indeed cause minorities in particular to rethink their alliances.
Depending on your ethics or lack thereof you can prove anything with statistics.
Well you can smokescreen a lie using statistics. But what’s your point?
Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.