Replies

If you do a thorough study of polls, you will find they are off more often than not. You will find an occasional poll that is accurate but it is part of what is called in statistics ‘the luck of the draw’. It is similar to dealing 3 back-to-back full house hands and then proclaiming that the recipient is a master player.

You can check my posting history. You will find I never believe in polls, that as a PhD in Statistics I know that all political polls are scientifically invalid. They are invalid because the sampling plans which are scientifically designed to survey a representative sample are never executed as required because most of the sample is not home, will not pickup, will pickup and hangup.

There is also the fact that landlines are becoming more scarce as mobile phone coverage grows. Those sampled are users of landlines. Mobile phone users are savvy enough to put their phone numbers on a do-not-call list and businesses are the same.

So in actuality, the poll sample is only about 10% of that which was planned and that 10% is not representative of the electorate.

So the polls are essentially garbage scientifically. Further, those conducting polls are the lowest rung of persons who perform statistical science. They are usually the rejects.

A professor or scientist with a great reputation can dirty their name in about one day by engaging in the activity known as political polling. Political polling is the rock bottom of the scientific community and is inhabited by losers.

And there is no way to elevate the field of polling statistics so that it is respectable. Census takers do better because most people recognize it is the law to respond to a census taker. But political poll statisticians are the dregs.

What it all means is that political polls are garbage put out by the low achievers in the field of statistics.

Now your question is not really a question. It is a snide remark. Such remarks are taught to you by the low-class media persons who day in and day out are making snide remarks. Such persons have to fill ‘dead air’ with their commentary and so they run their mouths nonstop and in so doing teach the viewing population that it’s Ok to run the mouth and make snide remarks.

In the 1950s TV taught people it was Ok to smoke and kids were taught it was Ok too. No different now with politics. Run the mouth without thinking, without trying to discern any intelligence.

Here is some intelligence regarding polls. Consider the question of why Donald Trump is the consistent leader in polls. Why is it that given polls are scientific garbage conducted by loser type people, why is it that he shows up as the leader and is the frontrunner?

Now here’s the same question with a number thrown in. Why is it that an invalid sample of 10% respondents to a political survey is showing a true signal pickup of a frontrunner?

In statistics, part of the art is to try and uncover an element of causality. Correlation is not causation but certain conditioning facts can point to a causal factor leading to a set of correlations.

The question that needs to be asked to get at causality is, just who is it that is sitting at home waiting for a phone call and just who is that will pickup the phone and actually respond to questions?

Here’s a key fact: there are almost 100 million Americans sitting at home out of work.

Political polls usually target about 0.1% of a voter sample and if only 10% of those bother to pickup, that’s about 0.01% of the targeted sample or 0.0001. If a region comprising 1,000,000 voters is to be sampled, 1,000 persons will be targeted, and about 100 will actually respond (polling backrooms will keep going through lists and backup lists to get to 1,000 but the representational validity will be ruined). So of 100 people, say 100 republicans, what is the chance that 50 of them are out of work? What is the chance that 65 of them are fed up with the government policy over illegal immigration? What is the chance that 70 of them see American jobs leaving for offshore destinations?

The unscientific political polls are reaching a segment of America that wants a strong leader who is believed to be capable of making needed changes. That would be Donald Trump.

Now, what about matchups with Hillary Clinton?

Again ask who is sitting at home in a sample of democrats?

One answer would be ‘welfare queens’.

When the press is reporting that Donald is losing to Hillary Clinton 46 to 41, what is likely happening is the unscientific sample of Donald’s out of work republican respondents are roughly equivalent to a sample of Hillary’s democrat welfare queens.

And then there are the questions of sampling the independents which have grown enormously from 24% of the electorate in 2010 to 42% in 2014.

And then there are questions related to poll sampling abuse, such as push polls and so on.

As to your question of 39% to 30% Trump to Cruz? It is meaningless. It shows nothing. Nothing has changed except that Cruz supporters are temporarily ebullient. There is no underlying causal change agent that is changing the reality of who is sitting at home. All the same conditions are still there.

And now Trump heads to states where he has great strength and widespread support.

Everything above discusses factors that are not statistical. They are hard facts given meaning and supplied with discussion on how they might appear in an invalid poll and how such polls might pickup a true ‘signal’.