Prediction of the difficulty level in a standardized reading comprehension test: contributions from cognitive psychology and psychometrics.

This research seeks to identify possible predictors of the difficulty level of reading comprehension items used in a standardized psychometric test for university admission. Several potential predictors of difficulty were proposed, namely, propositional density, negations, grammatical structure, vocabulary difficulty, presence of enhancement elements (words highlighted typographically), item abstraction level and degree of similarity between correct option and relevant text to resolve the item. By Linear Logistic Test Model (Fisher, 1973) it was found that the number of propositions, the syntactic structure, and fundamentally, the presence of difficult words contributed to the prediction of the item difficulty level.