NaN stands for Not a Number and indicates that the data is missing.
Observations containing a value of NaN are automatically filtered from any analyses if the question they appear in is numeric. That is:
- NaN values have no impact upon analyses based only upon questions of type: Pick One, Pick One – Multi, Pick Any and Pick Any – Grid, except where numeric statistics, such as Average and Average are being displayed using Statistics – Right or Statistics – Below. To automatically filter a value of these questions requires selection of the category as Missing Data in the Value Attributes dialog box.
- NaN settings are ignored in Pick Any, Pick Any - Compact and Pick Any - Grid questions and any variables in such questions only have values of 0 and 1.
- NaN values in variables used as weights are given a weight of 0.
- Observations with values of NaN on numeric variables used as filters are excluded from the filter.