Thursday, October 09, 2008

Why i usually ignore the polls

I don't waste much time with political polls. Like all survey tools, they are at best a snapshot of a point in time and election cycles are all about movement. After all, it wasn't that long ago that the media was polling the Hilliary-Rudy general election race.

Plus, it is hard to tell if a survey was done competantly or in a slipshod fashion.

Also, i know for a fact that respondents lie. I work in marketing and i've used market research and analysis for many years. Time and again i've run into results where customers answer questions contrary to the facts: Credit card customers who claim they never carry a balance even though the survey sample was drawn only from people who had carried a balance for six straight months. Farmers who claimed to buy only John Deere equipment even though we were surveying people who had purchased Case implements in the past years. Etc. etc.

Sometimes the results are a matter of simple confusion and low involvement. For example, the last time i was phone surveyed i had a hard time remembering if i bought my last printer at Office Max or Office Depot. I know which store i went to, but i can never remember the name. Being the scrupulous, honest sort, i answered "I don't remember". At least i think i did. Maybe i just picked one so i didn't sound stupid.

At other times, the reponses are shaped by perceived social stigma. Many people do not want to admit that they carry a credit card balance. Hence, they tell the surveyer (pollster) they do not. In the right environment, Walmart shoppers will claim they never visit the big store. There is perceived social pressure even in an ephemeral survey encounter.

Knowing all that, i don't get bent out of shape about something that could be incorrect today or irrelevant two weeks from now.

No comments: