We are a commune of inquiring, skeptical, politically centrist, capitalist, anglophile, traditionalist New England Yankee humans, humanoids, and animals with many interests beyond and above politics. Each of us has had a high-school education (or GED), but all had ADD so didn't pay attention very well, especially the dogs. Each one of us does "try my best to be just like I am," and none of us enjoys working for others, including for Maggie, from whom we receive neither a nickel nor a dime. Freedom from nags, cranks, government, do-gooders, control-freaks and idiots is all that we ask for.
Our Recent Essays Behind the Front Page
Monday, October 24. 2016
If you've been under a rock, you might have missed out on the news that the internet sorta crapped out on Friday. You couldn't twit, or twat, or twunt, or twisp, or whatever you call that internet grunting you guys do. You couldn't go to Reddit to visit subreddits that consist of one founder, one reader, and one moderator, adding up to one person. You couldn't binge-watch Game of Cards on Netflix. Or was it Harry of the Rings you like? I forget. Anyway, you couldn't do it.
I didn't notice. You know, if you ask me, the internet is now old enough to be judged on its merits. It kinda sucks. When it doesn't work because your baby monitor and a Cuisinart in the next country don't want it to work, it demonstrates that it was set up and is being run by fools.
Still, here we are. On to the links!
Tech oligarchs buy real businesses with borrowed money from fake businesses. You're going to see a lot of people without swim trunks when this tide goes out. I'm not sure this guy is smart enough to buy a real business, so he really better keep a weather eye on the tide chart. At least Bezos has a paper route to fall back on.
I never had one. Now no one has one. I feel some Latin coming on.
Hard to tell. He's not playing In-A-Gadda-Da-Vida
The purpose of rigging polls is not discouraging opponents and attracting money from people who want to back the winner. That's gravy. The real purpose is cover for stuffing ballot boxes. James Michael Curley could have told you that.
Do they hand out Nikes and purple blankets? Oh well. No children to pull the plug on you in the hospital when the nurse isn't looking. The nurse will do it when she notices no children come to visit you.
Well, it's Monday, and last time I checked, the internet still worked. You're going to have to think of another excuse for not getting anything done this week. Tell them your Samsung Galaxy phone caught on fire, ignited your car's air conditioner, and you got oversampled at Sunday's wine tasting. It's worth a shot.
Display comments as (Linear | Threaded)
Oversampling is a legitimate technique called stratification. You oversample demographics with larger variance and undersample demographics with smaller variance. Obviously if all Republicans vote the same way, you only need to sample one Republican.
You are supposed to weight the results according to their sample sizes though in putting it back together into a final poll result, which will be better than one constructed with random sampling.
For instance, if your overall sample of American opinion is 1000, your margin of error will be about ±3% (given a couple of reasonable assumptions). About 130 of these persons will be black, so your margin of error on black opinion will be about ±8½%. If you want to have more resolution of black opinion, you would oversample blacks. If you were to poll 260 blacks, twice the rate in the overall population, this would give you a margin of error of about ±6%. You would then adjust their numbers by half when including them in the overall survey.
Per the late Ray Kinsella (Shoeless Joe, Field of Dreams), the Cubs win the last pennant before Armageddon. Does this mean Hillary wins? or the Donald?
However if you sample 1000 people and only get to those who have land lines and are home during the day you miss a major segment of the population. Or if you want one candidate to win and believe you can demoralize many who oppose her than you sample selectively to get the results that you want. Or you can select an honest representative sample but ask questions that will skew the answer. Or you can sample 1200 people and select 1000 that give you a result that you like. Or... There are a thousand ways to phony up a statistical poll.
GoneWithTheWind: However if you sample 1000 people and only get to those who have land lines and are home during the day you miss a major segment of the population.
Sure. There are a number of possible biases. This is continually evaluated by polling agencies, who make their money by providing accurate results.
GoneWithTheWind: Or if you want one candidate to win and believe you can demoralize many who oppose her than you sample selectively to get the results that you want.
The email cited in the original post conflated oversampling with intentionally loading the results.
GoneWithTheWind: Or you can select an honest representative sample but ask questions that will skew the answer.
Sure. However, most polls publish the questions, which are usually along the lines of "If the election were held today, for whom would you vote?"
GoneWithTheWind: Or you can sample 1200 people and select 1000 that give you a result that you like.
Or you could just make up stuff without any evidence.
Best 10-year old drummer? Well, until he turned 11. The kid is GOOD.
re Thousands of California soldiers forced to repay enlistment bonuses a decade after going to war
Well. This will certainly undercut the credibility of the California National Guard, and the Pentagon as well.
Good luck offering bonuses to recruits in the future.
Of course if your intent is to weaken The Guard it explains the claw back.
One wonders how often they try to get money back after it has been incompetently spent?
It's always easiest to push people around who are least able to push back.
So explain for me what the purpose was of oversampling Democrats by 34 points in an Arizona poll? Which was used on Real Clear Politics to create the average for AZ?
The polls are a crock this year. All over the map. Swinging wildly back and forth. One national poll shows Hillary up 12 points, another shows Trump ahead by 2. And you think we should be paying attention to POLLS???
It's crap this year. Just plain crap.
MissT: So explain for me what the purpose was of oversampling Democrats by 34 points in an Arizona poll?
In Arizona, registration is about 35% Republican, 30% Democrat. The poll was adjusted to weight the proportions found in the registered voter database by county, party, age, and gender.
MissT: All over the map. Swinging wildly back and forth. One national poll shows Hillary up 12 points, another shows Trump ahead by 2.
Polls are a statistical sampling. If the margin of error is ±5%, then we would expect a range of 10% about the mean 95% of the time. The given range is well within what might be expected when there are that many polls being taken. Averaging multiple polls tends to increase precision, and helps improve accuracy by balancing different methodologies.
The drummer stayed in the pocket. No fills, no solo, no trading fours. So, we believe he was adequate.
MissT: Swinging wildly back and forth.
Actually, polls have been fairly consistent. Perhaps this will help. Compare the blue line to the red line. There is virtually no chance that statistical fluctuations alone can explain the persistent effect of the blue line always being above the red line over so many polls over so much time.
Of course, there could be some inherent bias, but these are many different polling companies, using many different methodologies. In the last presidential election, aggregate polling was within a couple of percent of the final result. You can also weight polls based on their past performance. For instance, Rasmussen tends to trend Republican, and missed the call on the last presidential election.
Just laughable. In AZ with an R+5, you should not need to 'adjust' the numbers so that D's come out plus +35. Your logic makes ZERO sense.
Please show me your master's in Statistics. Then maybe I'll listen.
Look back at the AZ polling in 2012, you see NOTHING like this in any of the polls. Just nothing. Your explanation is ridiculous.
MissT: In AZ with an R+5, you should not need to 'adjust' the numbers so that D's come out plus +35. Your logic makes ZERO sense.
No. The poll returned an excess of Democrats. The weighting reduces the effect of those excess Democrats.
Let's say we know, from register voter lists, that men and women each make 50% of the registered voter population. We take a poll, and the results return 70% women, and only 30% men. (Perhaps women are more likely to answer the survey.) We then weight the results to match the actual population. Let's say 60% of the women polled say Clinton, while 30% of the men say Clinton.
Unweighted, we have 51% Clinton. Weighted, because women are over-represented, we have 45% Clinton.
700 women, 60% Clinton = 420 women
300 men, 30% Clinton = 90 men
510/1000 = 51%
500 women, 60% Clinton = 300 women
500 men, 30% Clinton = 150 men
450/1000 = 45%
MissT: Look back at the AZ polling in 2012, you see NOTHING like this in any of the polls.
The RCP average for Arizona showed Romney by 7.5%. The final result was Romney by 9.1%. Shave away a few points due to changing demographics, and due to missteps by Trump, and you have a close race.
The Wikileaks releases confirm what we've always known, that most polls are rigged by the Democrats and their media to create the illusion there is more support for them than there is. This sets the stage for people not questioning voter fraud programs that are instituted to steal elections.
Look at a Hillary rally (where she is lucky to get 100 people) versus a Trump rally (where he gets five to ten thousand) and you will see where the real support is.
MissT--Typical leftist babble from Zachriel. They over-sample Democrats to mirror the large turnout by illegals voting that they are hoping for. During the Obama administration, the Democrats have done everything to encourage the invasion of Arizona by illegals and to wrest control from Americans in favor of the illegals. This includes invalidating laws and arresting the sheriff on trumped up charges who is trying to enforce the law.
Seems to me these military personnel and vets need a good class action attorney. These claims by California and the feds are totally bogus; defenses should include:
1. The screwup was California's, not the serviceperson's.
2. The serviceperson relied on the bonus offer in deciding to enlist or extend service (detrimental reliance, which creates a contract claim and defense against California and the feds).
3. Given how long ago this happened, the statute of limitations has probably expired for trying to get the money back (here, the longest statute of limitation for this type of claim is 6 years).
The fact it seems like from the story they settle with those who take them on shows they have a weak claim for reimbursement.
Another good example of the anti-military / anti-vet mindset of the Obama administration and the leftist California government.
Unfortunately, yet another Democrat-created problem that President Trump will have to address once he takes office.
Speaking of rigged polls, here's an arrticle about how many people in the mainstream media are actually Hillary Trolls. So we know the media is also rigged:
Jim: They over-sample Democrats to mirror the large turnout by illegals voting that they are hoping for.
It's standard statistics. If the sample over-represents Democrats, then the results are adjusted down to represent their actual proportion in the population.
Jim: The Wikileaks releases confirm what we've always known, that most polls are rigged by the Democrats and their media to create the illusion there is more support for them than there is. This sets the stage for people not questioning voter fraud programs that are instituted to steal elections.
Instead of accepting the obvious conclusion that Trump is losing, you conjecture a conspiracy that involves multiple independent polling companies, as well as fifty state election commissions.
Please also note that the original claim concerning Wikileaks and oversampling is a misunderstanding of statistics combined with a refusal to be skeptical of preconceptions.
"Or you could just make up stuff without any evidence."
Exactly. ZeroHedge takes a technical word, "oversampling", which he doesn't understand, then uses it to make up stuff.
Jim: Look at a Hillary rally [/i]
Hillary Clinton in Philadelphia