Wright, Benjamin D., "A History of Social Science Measurement."
Educational Measurement: Issues and Practice (Winter 1997) 33-45, 52.

Some reading notes--Adam Kissel

The gist of it: if you're a social scientist analyzing raw scores without using the Rasch model, your analysis is probably flawed. It is possible to have really solid analysis of raw data, but only if the Rasch model is used. Data needs to be reproducible as in the other sciences. If the data doesn't fit the Rasch model, you need better data!--don't just manipulate your model.