A common struggle for novice statisticians is choosing a measure of central tendency that is most appropriate in a given situation. It depends on the type of data you have and some basic data distribution characteristics.

1. If your data values are categories or descriptive words (e.g., political party, eye color, the country where you were born), the mode is your best option. Categorical data is information that fits into finite options in which only one option can be true at a time. These categories hold no numeric value and usually have no order in which they are typically placed (other than potentially alphabetical). For example, if you want to discover the most representative entree order at a wedding (beef, chicken, fish or vegan), you count the orders from each category. If beef was the most frequently ordered entree, beef would be the mode.
2. The most appropriate central tendency measure is the median if a data set has extreme scores (i.e., skewed). If you’re measuring something that may contain extreme scores, like income in New York City, the best way to describe the typical income would be the median since it is not affected by very high or very low-income values. The median is the preferred value when your data is a ranking of some type (e.g., level of agreement or satisfaction).
3. If the data in your set is not categorical and does not contain extreme scores, the mean is the most appropriate description of the sets’ central tendency.

Other important characteristics of these measures:

1. the mean is a more precise measure than the median, and the median is a more precise measure than the mode. Therefore the mean is the preferred measure when circumstances permit.
2. If a data set can be described using the mean, it can also be described using the median and mode. If a data set can be described using the median, it can also be described using the mode.

Watch this video for a quick description of determining the measures of central tendency using Excel.