| A variable that can only take distinct values. Examples might included the locality within a service (as in Chapter 8) or employment status. By contrast, in principle a continuous variable can take any of an infinite or large range of ordered values: age, time from referral to assessment. |
Details #
By definition nominal/categorical variables are discrete. Names, given or family names, are an interesting example of discrete variables that can in theory take an infinite number of values (not in reality but the real number of options is huge).
Yes/no, and other dichotomous variables are at the other extreme and can only take one of two values. Counts are discrete variables that can only take the value of zero or of a positive integer. (In theory, though again not in practice, are an infinite number of values a count can take.)
In reality variables that are actually discrete are often treated as continuous, e.g. we tend to measure age as age at the most recent birthday so it’s a discrete, integer valued, variable. Similarly, scores on questionnaires are generally discrete not truly continuous. However, for above ten values, for most statistical methods, treating age and scores as continuous works fine.
Try also #
Categorical/nominal measures
Ordinal measures
Interval measures
Ratio measures
Stevens’ hierarchy of measurement scales
Count data
Variable
Observation
Chapters #
Mostly chapter 5.
Dates #
Created 5/11/21.
