alsosDV: Alternating Least Squares Optimal Scaling

In davidaarmstrong/damisc: Dave Armstrong's Miscellaneous Functions

Alternating Least Squares Optimal Scaling

Description

This is a wrapper for the newer alsos function which allows optimal scaling of both dependent and independent variables. I retain the old operationalization of alsosDV for backward compatability purposes.

Usage

alsosDV(form, data, maxit = 30, level = 2, process = 1, starts = NULL, ...)

Arguments

form	A formula for a linear model where the dependent variable will be optimally scaled relative to the model.
data	A data frame.
maxit	Maximum number of iterations of the optimal scaling algorithm.
level	Measurement level of the dependent variable 1=Nominal, 2=Ordinal
process	Nature of the measurement process: 1=discrete, 2=continuous. Basically identifies whether tied observations will continue to be tied in the optimally scaled variale (1) or whether the algorithm can untie the points (2) subject to the overall measurement constraints in the model.
starts	Optional starting values for the optimal scaling algorithm.
...	Other arguments to be passed down to lm.

Value

A list with the following elements:

result	The result of the optimal scaling process
data	The original data frame with additional columns adding the optimally scaled DV
iterations	The iteration history of the algorithm
form	Original formula

Author(s)

Dave Armstrong and Bill Jacoby

References

Jacoby, William G. 1999. ‘Levels of Measurement and Political Research: An Optimistic View’ American Journal of Political Science 43(1): 271-301.

Young, Forrest. 1981. ‘Quantitative Analysis of Qualitative Data’ Psychometrika, 46: 357-388.

Young, Forrest, Jan de Leeuw and Yoshio Takane. 1976. ‘Regression with Qualitative and Quantitative Variables: An Alternating Least Squares Method with Optimal Scaling Features’ Psychometrika, 41:502-529.

davidaarmstrong/damisc documentation built on Oct. 1, 2023, 3:05 p.m.