Winsteps: history and "steps"

What is the origin of Winsteps and to what does "steps" refer?

 

Winsteps is an outcome of this process of development:

 

In 1983, Benjamin D. "Ben" Wright of the University of Chicago and John "Mike" Linacre released the first Rasch analysis program for personal computers (IBM PCs). It was also the first Rasch program to allow missing data. It was called Microscale, "Rasch scaling by microcomputer". Since MS-DOS was limited to 8-character program names, the actual execution name was "MSCALE".

 

1987: Mscale (dichotomies and Andrich rating scales) + Msteps (for partial credit "steps"). Ben implemented the Microscale algorithm on a Unix minicomputer using his own staff, but kept the PC execution name, "Mscale".

 

1989: Bigscale (back to PCs). Again under MS-DOS but with much larger datasets. Mike takes over software development again.

 

1991: Bigsteps (the functionality of Msteps was included in Bigscale). Ben interpreted this to mean "big steps forward in social science measurement"

 

1998: Winsteps (Windows-native Bigsteps). Ben interpreted this to mean "winning steps forward in social science measurement".

 

1999: The University of Chicago ejected Winsteps from its computers for "encouraging gambling". They seemed to be unaware that probability theory originated in gambling. Winsteps.com was launched as a separate endeavor under the direction of Mike Linacre.

 

2001: Ben Wright was incapacitated. Winsteps became an independent activity of Mike Linacre.

 

For developments since 2001, see www.winsteps.com/wingood.htm

 

When talking about Rasch measurement, Ben Wright used "step" to mean:

 

(a) the category number counting up from 0 at the bottom. The bottom step, for dichotomies or polytomies, was the lowest category, always numbered 0. Ben would talk about going up and down the steps as one moved up and down the latent variable.

 

(b) the location of the transition from one category to the next higher category on the latent variable. Now called the Rasch-Andrich threshold for polytomies and the item difficulty for dichotomies.

 

(c) the process of moving from one category to the next as one's amount of the latent variable changes. A low negative threshold below a category indicates that the category is easy to step into as one moves up the latent variable. A high positive threshold below a category indicates a category that is hard to step into. So "disordered" thresholds around a category (high below, low above) indicate a category that is "hard to step into and easy to step out of" as one moves up the latent variable, i.e., a narrow category. The extreme of this is an infinitely-narrow, i.e., unobserved, category. It is infinitely hard to step into and infinitely easy to step out of.