Monday, September 16, 2013

Skewing the Polls

We got a call on Sunday from the Gallup organization; they were conducting a poll and wanted our input.  I agreed to play along with them.  The only policy-related questions were about trust in government agencies, whether I had enough confidence in the privacy protections used by those agencies, and whether (for example) the Census Bureau should be allowed to use data from other agencies rather than direct surveys.

However, many of the questions were about what I consider atmospherics, of the order of "In the last week, have you read a local newspaper?" or "In the last week, have you visited a doctor?"  After a bunch of questions focusing on the last week, my interrogator then switched the context:  "OK, for the next questions, I want you to think about how you felt and what you did yesterday."

I explained that the previous day was Yom Kippur, and so my answers would by no means be representative.  The pollster said no matter, just answer.  So, we went through a bunch of questions.  On that day, was I well rested?  Would I describe my mood as happy, normal, or depressed?  Did I have any of the following:  Headache, backache, digestive woes, etc? (I always get a headache from being dehydrated).  Was I anxious?  If so, was it about money, family, etc?  Was I hungry, or did I have to skip any meals?  Again and again, I explained that the answers would be skewy, but we kept going.

Finally, the last few questions were demographic, age, race, etc--including religion: would I describe myself as Christian, Jewish, Muslim, Hindu, Buddhist, Atheist, or other?

It was amusing, and if you find yourself being told that you're hungrier and more introspective and prone to headaches than you think you are, then you can blame me.

No comments:

Post a Comment