Bell-Shaped Curve

A BELL-SHAPED CURVE is associated with the well known normal distribution where most people achieve ratings or scores in the middle range, while a few appear at either end of the range. Politicians have adopted the bell-shaped curve in support of their election bids through the science of polling. Faith in poll results has allowed elected officials and rival candidates to obtain rapid results on the mood of the electorate on social issues.

European mathematicians found that large collections of numbers and their averages exhibit a statistical distribution, which is approximated by a normal probability distribution, often called the Gaussian Distribution. The mean and standard deviation are key parameters that, if known, contain all the information defined by the data set. This means that the original data sets need no longer be kept as these two parameters completely describe the relationship of the measurements or samples. Today, this curve is frequently referred to as the Bell Curve and its predictive power is derived from the technique of random sampling.

George Gallup, for example, predicted the surprising Roosevelt victory of 1936 more accurately than other leading polls, such as the massive survey conducted by the Literary Digest. Based on the predictive power of random sampling. Gallup knew his Bell Curve-based statistical approach would produce much more accurate polling results. Gallup's 3,000 randomly selected interviews turned out to be more powerful than the 10 million people polled by the Literary Digest using a survey technique called quota sampling.

Since then, social scientists have used the statistics of normal distributions to determine the relationships between voting and variables such as education, socioeconomic class, age, and gender. Numerous studies have shown that the highly educated tend to vote more. The college graduate is more likely to participate in politics than a high school graduate, just as a white-collar employee will vote more than a blue-collar worker.

Elections are won or lost based on the votes of the middle class and the elites in society. It is in these classes where the most highly educated people are found. A strong correlation is found between education and level of political knowledge, attitude towards [Page 62]government, and interest in voting. The notion of correlation assesses the strength of relationship between variables; social scientists look for sources of correlation among groups of people within society.

A bell-shaped curve, or standard deviation diagram, showing the majority in the middle and fringe elements on the edges.

Because the types of polling questions asked tend to vary, the raw scores are not a meaningful measurement. Instead, the standard deviation is used. It provides a level of standardization similar to the measurement of time in seconds. One standard deviation translates into a result at either the 16th or 84th percentile, while two standard deviations places the result way outside the normal distribution of responses at the 2nd and 98th percentiles.

A standard deviation greater than two indicates a response on the fringe, with respect to a given issue. The standard deviation thus can be thought of as the average of the differences from the mean. Most polls authorized by politicians are done with a random sample size of about 500–1,000 people, which assures one standard deviation of accuracy, assuring confidence that the bulk of the constituent's views are reflected.

...