Descriptive Statistics

Median

In statistics, the median is a measure of "central tendency" (which is, loosly speaking, an indication of the "middle" of the variable). Unlike the mean, which is calculated with a formula, the median is calculated based on the rank ordering of all the values. If you were to sort all the observed values from the smallest to the largest, the median would be halfway "down the list." That is, 50% of the observations are smaller than the median. Obviously, then, 50% of the observations are higher than the median.

e.g., Among our customers, the median number of purchases is 71.

e.g., In our sales database, the median number of customers assigned per salesperson is 147.2.

Decimals are possible with medians even if all the data are integers because a mean is calculated to break ties.

Although the analysis of medians is less common in marketing research than the analysis of means, the median has one important advantage: It is robust (i.e., insensitive) to the presence of outliers. To choose an extreme example, even if LeBron James buys your product, adding his single household into your set of customers will have almost no effect on the median of your customers' household income. This can be contrasted to a large effect of LebBron on the mean of your customers' income.