A new Open Source web Statistical Tool

Professionals of education have to develop new instruments of evaluation. These tools have different goals (e.g., questionnaires, examinations, tests) and different forms (paper and pencil, computer assisted testing). Even though the use of all sort of evaluation tools increased during the last decades, academics, teachers and practitioners of the educational field have often difficulties to verify the psychometric qualities of their instruments. This is also the case for evaluation tools used via e-Leaning (e.g., pre-test skills or learning potential assessment before training, skills acquisition testing after learning, etc.). Many reasons can explain this: there are no user-friendly softwares available, they are expensive, and the interpretation of the results is difficult. To provide a solution to this problem, educational and psychological labs of the Laval University (Quebec) and of the University of Luxembourg have developed a user-friendly open source web tool. Files created in different formats (e.g., Excel) can be downloaded and data are analysed quickly and without the requirement of being a computer expert. Classical Test Theory (e.g., difficulty and discrimination indices, Cronbach's alpha) as well as Item Response Theory (e.g., item characteristic curves, test information function) are used. With a minimum of tables and graphics, the neophyte psychometrician will be able to choose good items for testing groups varying on age, abilities or cultural aspects. An interpretation support is currently being implemented. The objective of this proposal is to present a new instrument aimed to verify the psychometric properties of an evaluation tool. The paper also concerns the difficulties of choosing good items and response options. Different empirical data sets from cognitive tests, attitude questionnaires and multi-choice questionnaires will be used for demonstration. The procedure used to analyze the data and the interpretation of the results will be discussed.