What is Statistical Noise?
Many businesses depend a great deal on statistics. Statistical information is used to identify customer preferences and purchasing habits, production costs, and the efficiency of operating structures. While the generation of statistics is a great way to gain a better understanding of how a business should run and the direction it should take, the process can also create some worthless data. This is where the concept of statistical noise comes into being. Strictly defined, statistical noise is a slang term for the acknowledged variation that is found within a given data sample or formula. For example, within the context of developing production statistics related to the number of yards of cloth that can be produced within one hour, there are a number of factors that can impact the average number of yards produced. Quality of the base product, machinery malfunctions, operator error, and even the temperature and humidity level on the plant floor can impact the direction that the data will
But there’s some science that this may be tightening up and that John McCain may be making gains. But that could still be statistical noise. I want to see another day of those signs before I say it really means something. VAN SUSTEREN: In order for me to give a lot of credence to polls, I need to know this–looking back at the primaries, the polls done with the primaries, did either candidate poll a particular direction, poll high from what he got or poll low from what he got? BARONE: We saw a pretty clear pattern with Barack Obama, with just a couple exceptions, one of which was New Hampshire, where Barack Obama tended to get from the voters the same percentage that he got in the pre- election polls. So if you were going into the Ohio contest on March 4, Barack Obama was running about had 44 percent, 45 percent. And he got about 45 percent of the votes from the voters in those polls. That was a pretty consistent pattern. And there are some Democratic consultants, like (INAUDIBLE) in M
