09 April 2018

Spin, bias, or just wrong? #nlpoli

If four media outlets all reported a story in precisely the same way despite some fairly obvious factual problems with their interpretation,  is it spin, bias, or just a mistake?

That's the logical question out of last week's post on the way local newsrooms had reported a recent political poll about premiers and popularity.

The answer is that it is more than a mistake.  It is less than spin.  There doesn't appear to be a deliberate misinterpretation.

Yet what happened is a form of bias, in the same sense that a research firm would look at bias as a source of error. 

The causes are not partisan.

They are systemic,  identifiable, and correctable.

But the story presented is incomplete and  therefore inaccurately describes what the poll results show.

At the simplest level,  folks in the newsrooms would argue they got it right.  On a spectrum of popularity,  Ball *was* the third most popular even if the number of people who approved of his performance was less not the largest number of responses.

True, but there are three things that work against that interpretation. 

First is the results themselves.  Two premiers above 50% approval.  Seven below, of which Ball had the highest rating.

Second, the newsrooms faced exactly the same result in the last part of 2017 with the same polling company and didn't report it the same way.  Same facts  - third from the top,  comparable jump in approval from the previous quarter, but still less than 50%  - and yet two different approaches.

Last time,  newrooms played on the size of the jump: Ball's popularity increases.  They used the perspective that his popularity had plummeted in early 2016 but he had been on a steady climb, at least with this polling firm.  Some newsrooms went back to the same context this time or brought in some comparison to another polling firm's results.

The third argument against the news interpretation being right is that most people would likely not consider someone with approval below 50% to a popular politician by any measure.  Take Donald Trump as an example.  His popularity - or approval rating - is currently running at about 42%.  Trump's approval in January, after a year in office, was the lowest of any President since there have been polls.

But notice that even at 40%,  Trump's approval rating was within 10 points of Ronald Reagan and Barack Obama at the same point.  He's just a couple of points shy of Gerald Ford and not really that far away from Harry Truman.  That lends some support to the newsroom interpretation but notice that people have not been looking at the approval ratings at 50% or below as being a good thing,  that is,  reflecting general approval.  They are generally held to be a mark of poor performance or poor standing even for popular presidents such as Obama and Reagan.

Another potential angle to take on the story is the context of the Angus Reid poll in itself. That tells another story far more accurate and compelling than the one the newsroom chose.

By way of a simple request last week,  your humble e-scribbler got the complete set of results of approval ratings for premiers in Newfoundland and Labrador for as long as Angus Reid has been doing these surveys.  They even helpfully colour-coded the thing blue and red to show party affiliation of the premiers in addition to labeling the individuals by name.

Even where it is at 42% approval - with 49% disapproval - Ball is only now in the neighbourhood of where Paul Davis was (43%) at the end of his time in office.  It's worth noting how much time Ball has spent in the approvals basement compared to Davis.  You can also compare that to the amount of time that Ball has either trailed or tied Opposition Leader Davis in other polling as "best choice for Premier."

Ball remains well below Tom Marshall's approval over three quarters of the year. Tom started at 49%,  rose to 59% and finished at 52%.

In comparison to Kathy Dunderdale,  Ball's approval is only now at the level Kathy Dunderdale enjoyed in the middle of 2012,  in the immediate aftermath of the Bill 29 fiasco.  After that it was all down hill for Dunderdale.

Ball is heading to a leadership review vote in June, one which his team are clearly very worried about.  The other parties are picking leaders this month and, as far as the Tories are concerned,  their current outgoing leader has been polling as well or better than the Premier.  That's not exactly a strong spot to be in, even if,  in the Reid results,  Ball has been climbing back up lately.  if you add how Ball's numbers appear to have climbed - turning away from the financial crisis and making enormous concessions to the public sector unions - there's even more context to help understand what has been going on and what might come in the next few months.

More, accurate information gives you all sorts of other ways of tackling the story than the one chosen by the province's conventional media.  What you read in the conventional media is not always accurate or complete by any means.  This doesn't mean there's any sort of  conspiracy or plot involved but it does mean that when they go looking for information,  people had best look at more than one or two sources to ensure they have an accurate picture.  In this case even four sources wouldn't give you accuracy:  the full results, available from Angus Reid Institute,  told a better version.  Not only sources matter but the types of sources matter as well.



SRBP acknowledges the assistance of Angus Reid Institute,  particularly Shachi Kurl and Ian Holliday, in supplying the historical data from ARI used in this post.  

The interpretations, comments, and analysis in this post are entirely those of SRBP and do not reflect the views of ARI.