Polls Bias and Inaccuracies Explained
Nate Silver, the author of the article, analyzes the tendencies for firms to prefer a certain political party over another. As the chart demonstrates, polling figures tend to be highly skewed towards the favor of one party or another, depending on organization of where the poll was taken. The polls taken by the organizations are not intentionally favoring one or the other, but the polling firms get their data and information, unintentionally, from sources which have a greater favor one way or the other, the article explained. The "likely voter adjustment" is another method used by the author to estimate the accuracy. This method is most simply explained as evaluating polling results from firms' results for registered and likely voters. The article also explains the necessity for random sampling. Random Sampling is a method to ensure the data is not favoring a side in any way. To ensure this, all variables must be randomized; time of day a person is called, the wording of the question asked, mode of contact (not land-line exclusive), and location of the people being surveyed must all be random. Nate Silver uses polls from many different sources, despite the fact that they are inaccurate. He finds their historical bias and adjusts their polls using the aforementioned methods. Then, the adjusted polls are used to find the most accurate outcome from the calculated estimate.
Nate Silver has been successful at predicting election outcomes because he realizes the consistent mistakes/tendencies that polling firms make. His insight to the common error is supposedly scientific polls allows for more more improvement for the future. For example, if one were to point out an obvious flaw and fairly simple way to fix it, then it would be extremely beneficial to fix the glaring error. Because of his ability to find errors, the accuracy of future polls should be improved, assuming the firms attempted to reduce their inaccurate practices. The important of proper random sampling cannot be stressed enough. The process to get truly random and accurate data is meticulous, but the results' accuracy should be worth more than the extra effort. If the accuracy of the polling firm goes up, then their credibility should as well. As credibility goes up, one would imagine, the ability to make more money would go up as well.
Link: Calculating 'House Effects' of Polling Firms
Nate Silver has been successful at predicting election outcomes because he realizes the consistent mistakes/tendencies that polling firms make. His insight to the common error is supposedly scientific polls allows for more more improvement for the future. For example, if one were to point out an obvious flaw and fairly simple way to fix it, then it would be extremely beneficial to fix the glaring error. Because of his ability to find errors, the accuracy of future polls should be improved, assuming the firms attempted to reduce their inaccurate practices. The important of proper random sampling cannot be stressed enough. The process to get truly random and accurate data is meticulous, but the results' accuracy should be worth more than the extra effort. If the accuracy of the polling firm goes up, then their credibility should as well. As credibility goes up, one would imagine, the ability to make more money would go up as well.
Link: Calculating 'House Effects' of Polling Firms