One of the exciting things about election season for data people like myself is the wealth of excellent political data analysis.
I have written before about the excellent work done by my friend Mark Blumenthal at HuffPost Pollster (here’s a link to my interview with him).
In my view the other must-read author on election polling is Nate Silver of the New York Times’ FiveThirtyEight blog. (By the way, I’m in the middle of Silver’s fabulous new book, The Signal and the Noise).
Both Silver and Blumenthal use multiple polls to come up with their composite forecast.
Silver recently wrote an excellent post of the importance of not relying on too few data points: “Falling Prey to the Dangerous Temptation to Cherry-Pick Polls.” Blumenthal also discussed poll averaging on a recent appearance on the Diane Rehm Show.
This is a lesson I think many marketers – and even market researchers – often forget. We often put too much stock in a single dataset.
Data is powerful, but it has its limitations. Whenever possible use multiple sources of data and triangulate to find the truth.
About Dana Stanley
