Functional measurement was originally the brainchild of Norman H. Anderson. It is a methodology that can provide valid estimates of internal quantities when a model of a behavioral process is supported. That is, model testing and psychological measurement go hand in hand.
Most of the measurement techniques employed in the behavioral sciences rely upon untestable assumptions, the chief one being that the respondent does exactly what is requested when asked to report the strength of a feeling. Administrators of surveys are usually willing to assume (because they can see no alternative) that people in effect read internal meters the way we read an external thermometer.
Since Anderson's pioneering work in the 1960's, a mass of evidence has appeared in the literature supporting this innovative technology. Applications have been presented in domains including psychophysics, social psychology, health psychology, and marriage counseling. The biennial Functional Measurement Meeting showcases the latest empirical and methodological efforts.
However, the functional measurement papers have usually been written by and for specialists, so that it can be difficult for the uninitiated to gain the confidence to incorporate functional measurement into their work. The final chapter in Analysis of Variance and Functional Measurement: A Practical Guide is the first elementary treatment of this material. David J. Weiss was a graduate student of Anderson's and has published many papers using the technique.