Percent of Valid Observation

The percent of valid observation is a measure of the proportion of valid observations for a selected variable to the number of cases reported for that variable. If there are no missing data for the selected variable, then the percent of valid observations will be 100%.

Percent of valid observation can be useful in the initial stage of data analysis (exploration). Variables that are almost empty, e.g., have less than 1% (or some user-specified threshold) of valid observation, can be examined and removed from analysis specification.

Note: Percent of valid observation is the valid N divided by the number of cases that passed case selection criteria such as Case States, Selection Conditions, or Case Weights. Note that invalid weight should be treated as a failed selection criterion.

If conducting Case-wise deletion, valid N and the percent of valid observation will be the same for all variables.

When case multipliers are in effect, valid N and the percent of valid observation are computed as if the cases were repeated multiple times. Weighted moments do not affect valid N and the percent of valid observation.