Which kind of data is typically used to compute median values?

Median values are typically computed using ordinal data because the median is a measure of central tendency that requires a meaningful order among the data points. In ordinal data, the values can be ranked in a specific order, allowing for an assessment of the middle point when the data is arranged. For example, in a survey with a Likert scale (e.g., "strongly disagree" to "strongly agree"), you can determine the median response based on the ranked levels of agreement.

While interval and ratio data can also be used to compute a median, they are more suited for calculating means and other statistical measures that rely on the magnitude of the data values. For nominal data, the concept of median is not applicable since it does not have an inherent order or ranking, making it unsuitable for such calculations. Thus, ordinal data is the most correct choice for computing median values due to its ordered nature.