In SPSS, it’s important to choose the right level of measurement for each variable.

## About the SPSS Nominal Ordinal Scale options menu

SPSS gives you three choices for levels of measurement: Nominal, Ordinal, and scale. Each of those levels gives you different amounts of analyzable information in SPSS. Various procedures like hypothesis testing, require that your data is collected with specific measurement levels. The level is partially determined by the nature of your variables. However, you do have a little leeway when choosing an option.

To find the options, click the variable view tab at the bottom of the screen, then look at the “measure” column:

After entering your data, the SPSS default option is scale, but you can change this.

## Choices for SPSS Nominal Ordinal Scale

- Nominal variables can be placed into categories. The categories have no numeric value. These also have no order and are often displayed on a pie chart. If you appear to have categories with an order, you may have ordinal variables instead.
- Ordinal variables: These have an order, like hottest to coldest, 1st to last or lightest to heaviest. If you can rank your data 1st, 2nd… then you have ordinal data.
**Scale variables**: In SPSS, you’ll use the “scale variable” option for variables on the interval scale or ratio scale. The interval scale has meaningful divisions, like temperature. Variables on the ratio scale also have meaningful divisions, but the zero has a specific meaning: it doesn’t exist (like a height of zero).

Do not rely on SPSS to choose your scale. It doesn’t know what your intent was, or how the data was collected.

## The Hierarchy of Levels

The three levels should be treated as a hierarchy:

- Scale
- Ordinal
- Nominal

This hierarchy refers to the amount of analyzable information contained in the data. It is always possible to change scale data to nominal (i.e. move down the hierarchy), but you should never move in the other direction, changing nominal variables to scale. It’s mathematically unsound to do this. For example, you could change “height” from scale to nominal, making categories for “short to tall”. But it would make no sense to change a nominal variable like modes of transportation to scale, because there are no meaningful divisions between *car & plane* or *bike and train.*