Examples (PLS command) 

PLS Response WITH Price.
  • PLS estimates a univariate prediction model that regresses Response on Price.
  • Measurement level is not explicitly specified for dependent variable Response. Thus, if Response is categorical (nominal or ordinal), a classification model is estimated; otherwise, if Response is scale, a regression model is estimated.
  • Variable Price is treated as a continuous predictor (covariate).

Classification Model (Explicitly Specified)

PLS Response MLEVEL=N REFERENCE=5 BY Region WITH Price.
  • PLS estimates a univariate prediction model.
  • Since dependent variable Response is specified as nominal (N), a classification model is estimated. Value 5 is used as the reference category.
  • Variable Region is treated as a categorical predictor and Price is treated as a continuous predictor (covariate). By default, a main effects model is estimated.

Multivariate Model

PLS Q1 MLEVEL=S Q2 MLEVEL=S BY Region Edlevel 
  /MODEL Region Edlevel Region*Edlevel.
  • PLS estimates a multivariate regression model.
  • Both dependent variables, Q1 and Q2, are treated as scale (continuous) variables.
  • Variables Region and Edlevel are treated as categorical predictors.
  • MODEL specifies a fully factorial ANOVA model that includes main effects for Region and Edlevel as well as the interaction between the two predictors.