Briefly |
Each case can have a different value in a case attribute. But the value of a measure applies to the entire collection. |
Fathom uses these two kinds of attributes very differently. You'll use case attributes most of the time, but there are some things you can't do without measures, especially related to sampling and simulation .
Attributes
In a collection of people from a Census, you might have age, income, and maritalStatus. Since each case is a separate person, each case has a separate value for each attribute.
How to make a new case attribute
Measures
Measures apply to the whole collection.
You usually calculate measures. For example, you might have a measure averageAge that has a formula, mean(age). That value is the same for all cases in the collection. It will change as you add new people to the collection, always reflecting the mean age in your data set.
Occasionally, you will use measures to store collection-wide parameters, such as the year the Census was taken. Then you would make a measure called censusYear, with a value like 1990.
Using measures
If you're learning about sampling distributions, you need measures like averageAge. Every sample "inherits" its source's measures and recomputes their values. So each sample will compute a new averageAge, which you can record in a measures collection.
You could use the measure censusYear to create a new case attribute, yearOfBirth, with the formula censusYear - age.
Other distinctions