Based on an original hypothesis that differentiates the critical attributes of educational games and simulations (Sauvé
et Chamberland, 2003), an analysis grid for articles and research reports was developed and validated by the
triangulation method which ensured reliability of data coding (Lincoln et Guba, 1985). Then researchers, other than
those coding the information, proceded by counter coding the units and their identifiers for the categories and subcategories
created by the coder. This operation was applied to several articles until a higher than 90% agreement was
obtained. More precisely, each time the coding was not identical from one researcher to another, it was discussed in
order to understand the points of divergence and to readjust the categories, if necessary. This validation allowed us to
establish the relevance of the categories and to refine the interpretation of certain units of meaning to make sure that
mutual comprehension was relatively identical.