Monday, October 1, 2012
Political Polls--How Do They Justify Their Confidence Interval
There has been a lot of debate over how accurate political polls are since pollsters presume that Democrats have higher turnouts than Republicans, and, consequently, give Democratic candidates (i.e., Obama) a 5 to 10% bump. Now, this article indicates that only 9% of households respond to political polls. If the response rate is so low, how can the pollsters justify the confidence interval for their polls? How do they account for the non-response error, or, since many non-responses may be due to lack of "land line" telephones, non-coverage errors? Perhaps someone with a background in statistics can help explain this.