OPTIMIZING THE EVALUATION ON TEXT COMPREHENSION

A theoretical model of reading comprehension was validated, including ten tasks and two informationrecall contexts, with responses of 414 upper secondary students. Logistic calibration and 2 ^nd order confirmatory factor analysis were used. 16 items, which showed adequate discrimination indices (a > 0.65), a desirable continuum of facility-difficulty (from -2 to 2), and a lambda > 0.30, were accepted. The theoretical model showed statistical fit, given no difference between the theoretical and saturated models was found (χ ² = 130.59, p =.00). Practical fits showed appropriate fit values (CFI = 1.00, RMSEA =.00). The hypothesized structure was validated through achieved factor loadings, allowing to accept the underlying theoretical sense, as well as to trust in used measures.