This form provides the ability to add variables to the Data Editor that are calculated from sample data. Variables can be generated using a dependent variable indexed by independent variables, or variables can be generated across rows. The options include:
Generate Dispersion Testing Values
Sample statistics:
- Mean, Median, Range, Standard Deviation, Variance, Low, High, and Sum. A single value will be created for each group.
Standarized groups:
- Deviations from the Mean or Median, or
- Combinable Distributions.
- Combinable distributions generate groups each with the overall grand mean plus a random normal value with the mean equal to zero and the standard deviation equal to the standard error (1/sqrt(n)). The overall grand mean is preserved. This attempts to create a distribution closer to the expected population if group-to-group differences were eliminated. If only the mean is subtracted without the random value, the combined groups would have less variation because of the sampling error in the means is not taken into account.
Ranks:
- Ranks are created with correction for ties. The ranking is done from 1 to n, and ties are entered as the average rank.