Trump

[shocka khan]

Sample sizes of 15 out of 300 million are obviously pretty useless.

Sample sizes of hundreds of thousands, or even millions, are a totally different animal.

Quit playing dumb.

I disagree. Take a look at this:

https://statswithcats.wordpress.com/2010/07/11/30-samples-standard-suggestion-or-superstition/

PwC and E&Y both consider a sample of 25 items as sufficient to predict the behavior of a population under study. 'Your mileage may vary'.

It will depend on the variability of the population (i.e. deviation from mean). A larger variance would indicate that the sample mean may vary more from the population, but the gist of this is that assuming a normal population distribution, a sample size of more than 10 will be fairly accurate (incorrect less than 10 percent of the time). Doubling sample sizes from that point does increase precision (as noted in the article), but the precision from the larger sample sizes diminishes as the size goes up.

Many large polling organizations poll only a few people (not thousands, more like hundreds or even 10's) in order to obtain an accurate result.

FYI, sample design is also important. One of the reasons Gallup got their numbers so wrong on the last election is because they were polling people who had land line phones. Most of the people who have land lines these days are older. As a result, this introduced sample bias.

So design of the sample and variance from the mean in the population are probably the two most important factors in determing how accurate the sample will be.

[Jamar Howard 4 President]

Geez. I specifically quoted you and that framed the discussion -- which had nothing to do with the general election. Your singular comment has only to do with performance in the primary and nothing else.

You tried to argue against a single sentence quote of mine outside of its larger context. I had no interest in your rabbit trail. I think we've already been down that one anyway.

[Jamar Howard 4 President]

Stop lying and quote where I argued this point with you. From recollection all I did was agree with you on this point. Stop lying.

I was under the impression that you originally disagreed before eventually agreeing. If you agreed all along, cool. Glad we actually agree on something.

So you if you agree that primary turnout is not well correlated with general turnout, why do you keep using primary turnout as proof that Trump will do well with R's in the general? It seems you claim there is little correlation, but then keep implying their is correlation via your arguments.

[Jamar Howard 4 President]

I disagree. Take a look at this:

https://statswithcats.wordpress.com/2010/07/11/30-samples-standard-suggestion-or-superstition/

PwC and E&Y both consider a sample of 25 items as sufficient to predict the behavior of a population under study. 'Your mileage may vary'.

It will depend on the variability of the population (i.e. deviation from mean). A larger variance would indicate that the sample mean may vary more from the population, but the gist of this is that assuming a normal population distribution, a sample size of more than 10 will be fairly accurate (incorrect less than 10 percent of the time). Doubling sample sizes from that point does increase precision (as noted in the article), but the precision from the larger sample sizes diminishes as the size goes up.

Many large polling organizations poll only a few people (not thousands, more like hundreds or even 10's) in order to obtain an accurate result.

FYI, sample design is also important. One of the reasons Gallup got their numbers so wrong on the last election is because they were polling people who had land line phones. Most of the people who have land lines these days are older. As a result, this introduced sample bias.

So design of the sample and variance from the mean in the population are probably the two most important factors in determing how accurate the sample will be.

So you are saying small samples have more usefulness than we give them credit for? And therefore large samples would also be very useful? And therefore @Kung Wu:'s crazy analogy to skinheads and Hitler is highly flawed?

[shocka khan]

So you are saying small samples have more usefulness than we give them credit for? And therefore large samples would also be very useful? And therefore @Kung Wu:'s crazy analogy to skinheads and Hitler is highly flawed?

What I'm saying is that smaller samples are more effective and efficient. Would you rather evaluate a sample of 25 or a sample of 300 if the difference was less than 2%? Strictly from a time standpoint, you would not.

As far as the analogy to skinheads and Hitler, not all skinheads are white supremacists, this is a logical fallacy ('that which is true of the parts is not true of the whole'), but that has nothing to do with statistics (unless you want to say if your sample included nothing but skinheads voting for Hitler and ignore all other choices they might have had, in which case, the sample would not be designed properly, as it did not encompass any other choices a skinhead might have had).

Other than that, my comments were only directed to sampling and efficiency, not the logical validation of someone's arguments.

[Jamar Howard 4 President]

Good grief.

[ShockdaWorld]

This looks like a whole bunch of people all trying to see who can troll the others the hardest and make them fall for it.

[Jamar Howard 4 President]

This looks like a whole bunch of people all trying to see who can troll the others the hardest and make them fall for it.

I'm really trying to have a real discussion, but clearly when @Kung Wu: said he was here to be a dick (his own words, not mind) he meant it. @shocka khan: must have decided he wanted to try the same routine.

Both will be voting Trump this fall. Sad!

[ShockTalk]

Clinton opens up double-digit lead over Trump nationwide - Reuters/Ipsos poll

"Some 46 percent of likely voters said they supported Clinton, while 35 percent said they supported Trump, and another 19 percent said they would not support either, according to the survey of 1,421 people conducted between May 30 and June 3."

[shocka khan]

I'm really trying to have a real discussion, but clearly when @Kung Wu: said he was here to be a dick (his own words, not mind) he meant it. @shocka khan: must have decided he wanted to try the same routine.

Both will be voting Trump this fall. Sad!

Here's what's sad.....Hillary is not in jail (or at least indicted) regarding her mishandling of classified information.

What's sadder is that we have to choose between these two clowns.

If I have to choose, I'm going to vote for my own self-interests and yes, I'll be voting for Trump.

It certainly beats voting for Hillary. Her bunch wants to castrate you, me, and any and all men. She will be at least as bad (if not worse) than Ann Richards was when she was gov of Texas.

If it weren't for the women Hillary has surrounded herself with, I'd probably vote for the other Donald.....Donald Duck

[Eric]

If either of my kids behaved and spoke like these two candidates, I would not have tolerated or rewarded it. I will not reward either of these equally reprehensible people either. Hillary's shortcomings are well documented, and Trump is for Trump, not America. Yes, a third party vote is a waste in a way, but here in Kansas, it won't make a difference. We keep electing this type of candidate despite their continued behavior.

[SB Shock]

Clinton opens up double-digit lead over Trump nationwide - Reuters/Ipsos poll

"Some 46 percent of likely voters said they supported Clinton, while 35 percent said they supported Trump, and another 19 percent said they would not support either, according to the survey of 1,421 people conducted between May 30 and June 3."

Just looking into the poll - this poll is an outlier (sits on extreme side for Hillary). I wouldn't be surprised after Hillary wins the nomination and consolidates the Bernie supporters that's she would pull to 4-6% lead.

Then again with the beatings that trump supporters have been taking this week - I wouldn't be admitting anything to anybody at the moment I was going to be voting to for Trump. Already seeing here on shockernet the Hillary supporter trying to shame people.

[MoValley John]

What's sadder is that we have to choose between these two clowns.

In the rare moment of a June day, I find a comment from you in which I agree.

In another rare moment, I disagree with Kung Wu on an issue and can't support Trump.

I have a brother who is very wealthy and a dyed in the wool Republican. He is not supporting anyone, but will be voting self interest for Clinton. He knows she's a crook, he knows her policies, he knows she's bought and paid for. But he also knows what she will do, how to invest while she is in office and knows that he will come out ahead. My brother doesn't fear an economic downturn, he fears uncertainty and the unknown.

I'm not wealthy, I'm voting for Montgomery Brewster. Seriously. If they published that data, I bet Brewster would surprise people with how many votes he receives.

[ShockTalk]

None-of-the-above[1].jpg

none-of-the-above[2].jpg

[SB Shock]