In the realm of probability and statistics, skewness or skew is a measure of the extent to which a data distribution is distorted from a symmetrical normal distribution. The distortion is in one direction, either toward higher values or lower values. In the former case, the distribution is positively skewed; in the latter, it is negatively skewed.
Positive skew is often shown in data based on cycle time, such as hold times for customer service representatives. The lower limit on any measure of time is zero, yet there is no comparable upper limit. For this type of data, the mean is higher than the median, which is higher than the mode. Housing prices often show positive skew, making it important to differentiate between mean prices and median prices.
Zero skew represents a symmetrical data distribution. This type of distribution is often seen for physical data such as height or weight in a random sample.
Negative skew may appear when a measure has a natural upper limit or when another reason exists for clustering of data points at the upper end of the distribution. For example, a chart of years of school completed might have a mode of 16, representing college graduation, with data points not greatly exceeding that (for graduate or professional school). At the lower end are data points representing individuals who did not go to school or who dropped out at various points throughout primary and high school.
Within a plane, the only lines that never intersect are parallel lines. In a three-dimensional space, lines can be in different planes and not intersect. These lines are referred to in geometry as "skew lines." For example, a line running along the sidewalk in front of a house and a line from the front to the back of the house's roof are skew lines, as they do not intersect and are not parallel.