How Do You Test For Normality (Bell Curve Distribution)?
Many measurements in real life have a distribution pattern that resembles a bell curve, formally known as the “normal distribution” or “gaussian” curve. For example, IQ is distributed normally. Or, if you flip a coin a hundred times, the expected number of heads follows a normal distribution. The logarithm of human height and foot length is normal as well (often called log-normal). In many statistical applications (such as quality control and error analysis) distributions are assumed to be normal. However, this is something that always needs to be verified, or else the analysis will be incorrect. The steps below describe an easy method for non-statisticians to check if a distribution is more or less normal. First, use a large enough random sample size for the normality test. To accurately verify whether or not a distribution is normal, you should have at least 50 data points. Next, compute the average (mean), median, range, and standard deviation of the sample. Call these numbers A, M,
