DOT

Design of Optimal Test (DOT) is a Windows (Windows 95 or higher) program to select items into a test (or more tests) that comply the best possible way to the required test design criteria. The DOT program uses an item bank which can be calibrated using the Rasch model, the OPLM model, the 2-parameter logistic model or the 3-parameter logistic model. The program can handle dichotomous and/or polychotomous items.

DOT is a program that uses a variety of techniques and routines based on linear programming and optimization technology as found in industrial-, technological- and financial sectors. The program also uses genetic algorithms. Given test design criteria DOT will always find the smallest set of items complying to the test design criteria. Such design criteria can include:

• test difficulty;
• test reliability;
• test length (in time, in number of items);
• content based criteria (test blue print);
• and combinations of these criteria.

DOT can also create multiple tests, e.g., parallel tests, with or without overlap items. The DOT program translates the design criteria into a series of mathematical expressions which then are solved by effective algorithms. Since 1985 several publications have appeared in international journals (Psychometrika, APM ) about the theory, methodology and algorithms used in DOT.