CRA poll does not show “surging support” for the NDP

Written by Bruce Wark on May 9, 2017

To the editor,

A front-page story published in the May edition of The Brief reports incorrectly that a recent Corporate Research Associates poll shows “surging support” for the New Brunswick NDP while support for the provincial Green Party “declined sharply” between November 2016 and February of this year.

The story goes on to claim that “NDP support nearly doubled from 7% to 12%” and that the “poll also shows that the NDP has fully recovered from a slump in support.”

I wonder if your reporter was suffering from an acute attack of wishful thinking.

The NDP’s apparent five point rise and the Green Party’s four point decline fade away into nothing when the CRA poll’s margin of error is taken into account.

In fact, with a margin of error of plus or minus 4.4 percentage points, there wasn’t much movement in party support at all since any movement was pretty much within that margin of error.

Margins of error are determined by sample size. The smaller the sample, the wider the margin. I’d say that a poll based on a sample under 800 is generally not worth reporting. In this CRA poll, the sample size of decided voters was only 492.

To quote Shakespeare, your dramatic story of shifting political support was a tale “full of sound and fury, signifying nothing.”

Editor’s note: A statement about the margin error was added to the online version of this story.

Comments are closed.