Bloomberg's Poorly Written Poll by @DavidOAtkins

Bloomberg's Poorly Written Poll

by David Atkins

During my lunch break on work project today I received a phone call from the Bloomberg News Service. It was a standard poll asking my opinion on a range of issues including the economy, the President's job performance, approval opinions for the Republican candidates, whom I would vote for in the Congressional and Presidential elections, and binary choices on my opinions of the hot-button issues of the day ranging from birth control to fracking.

As a research professional, I was taken aback at the simplistic nature of many of the questions, and at the gross omissions in the questionnaire. For instance, when asked about the most important issues of the day, the list did not include climate change or environmental issues of any sort, which would have been my answer. In this day and age when climate change is by far the most pressing issue facing mankind, that sort of oversight is inexcusable (I answered jobs and unemployment instead.)

But more interesting were the questions asking simply whether I approved or disapproved of 1) the President's negotiations with the Republicans on the budget, and 2) the President's handling of the deficit.

Of course, my first instinct was to answer a resounding no to both questions. But I hesitated and reconsidered. These were not open-ended questions, and there was no option for a third "other" response.

Now, any researcher worth their salt would tell you that this is terrible polling design. The questions should have been phrased and scaled something like this: "Considering the President's negotiations on the budget, would you say that 1) the President has not done enough to compromise with Republicans on the budget; 2) the President has done enough to compromise with Republicans on the budget; or 3) the President has given up too much ground to compromise with Republicans on the budget?" A similar approach to the deficit question should have been taken.

Forcing respondents to choose a straight-up "yes" or "no" for approval/disapproval does not allow any indication even in a quantitative manner for why one might or might not disapprove.

I knew that any disapproval answers for the President's approach to the deficit and negotiating with Republicans would be interpreted by lazy researchers and a lazy press to mean that I had felt that the President was too partisan in his approach, and not concerned enough with closing the deficit. So I swallowed hard and said that I approved.

So watch for the Bloomberg poll results to come out in a few days, and know that whatever results are reported will be skewed significantly to the right, through the peril of honest answers to poorly designed questions.

In the future, Bloomberg News might want to consider using a competent firm to design its polling questionnaire--unless, of course, it's looking to drive a right-wing narrative. In which case, I suppose they should carry on as usual.


.