Briefly |
Numeric attributes are usually numbers. |
Fathom treats these two different kinds of attributes differently. To get the most out of the program, you should understand the distinction.
In Fathom, an attribute is treated as either numeric or categorical.
Numeric |
The values are numbers. Height, age, speed, and price are good examples. Sometimes we call numeric attributes continuous. |
Categorical |
The values are not numbers. Sex, school, favorite vegetable, and marital status are good examples. |
Both kinds are important. Here are some implications of the difference:
You can do arithmetic with numeric values, but not with categorical. So you can't calculate mean or standard deviation for favorite vegetable, but you can with height.
The kind of attributes you have determine what kind of graphs you can make. See about how to make a graph for a description of the different types.
Forcing treatment of attributes
When you use an attribute in a graph, summary table or analysis, Fathom looks at all the values of that attribute and decides whether to treat it as numeric or categorical. If every value is either blank for a number, Fathom treats the attribute as numeric; otherwise it treats the attribute as categorical. Generally this behavior is what you want, but you can override it by holding down the key (Windows) or key (Mac) when you release an attribute in a graph, summary table, or analysis to force the attribute to be treated numerically. Similarly, hold down the key to force the attribute to be treated categorically.
A tricky way to force a numeric attribute to be treated categorically, is to create a new attribute that will look identical to the original attribute but which is defined by the formula concat(" ", oldAttribute).
When it's not so clear
Sometimes there are numbers that act like non-numbers. For example, suppose you rated peaches as Class 1, Class 2, and so forth. You might make an attribute called PeachClass and give it values of 1, 2, 3, and so forth. Is it numeric or categorical?
It depends on what questions you'll ask. If you ask, what's the average weight of the Class 2 peaches, you're treating it as a categorical variable, just as if you'd asked, what's the average weight of the people who like broccoli.
On the other hand, if you ask, what's the average class of the peaches we picked on Thursday, it's acting more numeric. Even so, you should be extremely careful when you do arithmetic like this to ensure that it's appropriate.
Other distinctions