Wednesday, December 7, 2011

Dec. 7: Lies, damn lies, and....

Statistics can, to say the least, be misleading. They often are - and are often deliberately misleading.  (Even when they tell the truth, they can be - and often are - used to give a false impression.) They can also be slanted by the precise wording of the question.

To add to the problem, polling companies are hired to do polls. Unlike Santa Claus, somebody pays them for what the they - and that somebody often makes it clear what  sort of results he or she would like to see.  A good reporter or a good news editor should know that.

Evidently, The Moncton Times and Tribune is short on good reporters and good editors. Today's lead story is that the Alward government's popularity is up, with the support of 57% of NBers. The Liberals got 28%, the NDP 23%.

Oops! That doesn't add up to 100%. It adds up to 108%. But that's not all. Not by a long shot.

A later part of the report says that 20% either refused to reply or didn't know; ( it isn't clear whether this was referring to a particular question or to to the whole thing.) In any case, the conservatives did not get got support of 57% of all New Brunswickers - or even of 108% of all New Brunswickers. They got the support of 53% of those who answered. But those who answered represented only 71% of New Brunswickers.

That means real Conservative support was more like 42% - and there's over a three percent margin of error. Nor is even that the whole story.

The company hired to do the research is Corporate Research Associates, a purely commercial outfit frequently employed by large corporations. Is it possible CRA might have a bias? Is it possible it might, for example, skew the results with clever wording of the questions? (We aren't told the wording of the questions.) Is it possible they might be selective in their choice of districts in which to conduct the survery?  There's a hint of some, considerable bias in paragraphs two, three, and four of the front page story.

The CEO of Corporate Research associates told the rerorter that the party's handling of the shale gas issue might be the reason for it's good results. "The way they've handled shale gas has been pretty even-handed, " he said.  Even-handed? What - they've provided full information?  They've given both sides of the issue? They've prosecuted SWN for deliberately breaking the law in its most recent exploration in Sussex?  Get real.

He also said proof of the government's even-handedness is shown in its promise to bring down the toughest regulations of any jurisdiction to protect the environment.

We've had no regulations to speak of for a dozen years - from either Conservatives or Liberals. This Conservative government has had power for over a year. It's been promising tough regulations since the start. We still haven't seen any.

It has taken no significant action against SWN, even though SWN publicly admitted it broke the law, and did so to save money.


This, says CEO of Corporate Research Associates, is proof of even-handedness.  Gee. I wonder which party he supports.

Like many statistical reports, this one is not to be trusted.

Even the CEO of Corporate Research Associates must know that 57%  support in a poll in which 29% of those questioned refused to answer does not represent 57% support from all New Brunswickers. And even if all the polling methods were honest and competent (- the CEO's statement does not make this reader confident of that -), it still would not support the claim of 57% of all New Brunswickers.

Trained statisticians would know that. Trained journalists would know that. There are only two explanations for such a poll being reported as this one was. Some people are unethical or some people are incompetent. Possibly, some people are both.

This sort of reporting (combined with the triviality of the reporting in general, the manipulation and sloppiness of editors, the bubblegun silliness of its City Views and Hump Day on the op ed page, the obvious bias and rant of much of its editorializing and commentary) is what makes The Moncton Times and Tribune the worst paper (in a lump with the rest of Brunswick Media) in Canada.

It has also and consistently been deplored by parliamentary studies of the quality of news media in this country. The only print media that the majority of New Brunswickers get are designed to keep them in ignorance of what is going on in the world, in Canada, in New Brunswick, and ieven in Moncton. It is designed to distribute propaganda to keep people obedient to their real masters, but happily absorbed in reading horoscopes, advice from Miss Manners, and an op ed page that is largely children's stories for adults. Try the two gems on today's op for examples.

Oh - Norbert let his enthusiasm for buzz words and bias slip in today's column. As a major leader in setting up social programmes, Britain is often criticized, especially by the rich, as a "nanny" state, one which spoils the poor with things they don't deserve - like health care.

Nannay state. Good choice of words. Sets up a clear image. People like Norbert pick up words of that sort - and use them instead of thinking.

In today's column, Norbert refers to Britain as a nanny state, That buzz term, in the view of Norbert's of this world, say all there is to say.

Evidently, Norbert doesn't know that Britain has the largest wage gap between rich and poor of any developed country. Yes, substantially larger than either the US or Canada. And, like the US and Canada, the rich of Britain are getting even richer, and the poor poorer.

That, of course, is because it's a nanny state. Even worse than New Brunswick, Britain has massive poverty and indebtedness because it's a nanny state - for corporations.

1 comment:

  1. > . Is it possible CRA might have a bias?

    Well they might, or they might not. But if you're going to suggest that they do, you need to have some evidence, not just a leading question.

    I would recommend following http://www.threehundredeight.blogspot.com/ which is the most authoritative coverage of polling in Canada I know of. Regular reading will give you a very good idea of which polling outfits can be trusted and which ones are fly-by-night.

    ReplyDelete