Letter: Bonkers polling? - Businesslive

14 March 2019 - All of what Everatt says about the IRR’s poll is wrong, and frankly suggests he has no idea what constitutes best polling practice these days.

In her column “Opinion Polls and Their Vastly Diverging Results” (March 12), Carol Paton gives ample space to ANC pollster and Wits University academic David Everatt to criticise the Institute of Race Relations’ (IRR) most recent market research, but none to the IRR to return the favour.

All of what Everatt says about the IRR’s poll is wrong, and frankly suggests he has no idea what constitutes best polling practice these days. Allocating undecided voters is not “guesswork”, it is scientific and rational, and standard practice the world over among serious polling companies. Our sample is random, which, if he bothered to properly read the full methodology (which would require first asking for it), he would know. And so on and so forth.

But look, whatever. Here is the thing: Everatt’s poll has the DA on 14% in Gauteng. That is a mad number. Bonkers, in fact. The IRR has the DA on 32.4% on the provincial ballot, as of the end of February, and into the mid-30s on lower turnout.

But don’t take my word for it. We will see, won’t we? In a little over eight weeks we will know. I hope Business Day returns to Mr Everatt the day after the election and puts that 14% to him. Or just give us call, we will be more than happy to return that favour.

Gareth van Onselen
Head of Politics, Institute of Race Relations

https://www.businesslive.co.za/bd/opinion/letters/2019-03-14-letter-bonkers-polling/