El objetivo de este trabajo es comparar los resultados obtenidos en estudios de validez y confiabilidad de una Escala de Utilidad de la Matemática cuando varían los formatos Likert de los ítems. El test mide las creencias de estudiantes de Psicología respecto de la importancia atribuida a la Matemática para la carrera y el futuro desarrollo profesional. Participaron 939 estudiantes de Psicología (81% mujeres) quienes respondieron a los ítems usando escalas de 3, 5 y 6 categorías. Se controló el efecto del orden de exposición de los individuos a cada formato y se incluyeron además otros instrumentos para reducir la memorización de las respuestas. Las escalas Likert con más categorías incrementaron la precisión del instrumento en los niveles extremos del rasgo, pero a costa de comprometer las evidencias sobre la estructura interna (Análisis Factorial Confirmatorio y Modelo ...

Happiness is defined as a positive emotion focusing on the present, which is highly desirable but difficult to achieve. To estimate the level of happiness in adults, we used the Argentinian version of the Happiness Scale of Lima (EFL, Alarcón, 2006). The EFL measures three dimensions: External Perspective of Sense of Life (SVPE), Internal Perspective of Sense of Life (SVPI) and Life Satisfaction (SCV). Three independent samples of 200 individuals each (50% female) were selected considering different educational levels (from high school graduates to college graduates). A two-way analysis of variance (ANOVA) was applied to each dimension taking into account gender and educational level as fixed factors. In all three dimensions, the interaction between having a higher educational level and being female favored significantly the highest scores of happiness (SVPE, F = 4.14, p = .02, partial...

Problem and objectives: To find out if the structure of the prosocial behavior measured by the Prosocial Behavior Scale (PBS) can be represented by a general factor, i.e. prosociality, combined with specific factors: the subscales of the PBS. Method: The sample consisted of 692 subjects (65% females) recruited by a convenience sampling method and living in Greater Buenos Aires and inland areas of Argentina. The assumptions of unidimensionality and local dependence were checked. Then the relative fit was compared between three models of the Item Response Theory: Graded Response Model (GRM), Multidimensional Graded Response Model (MGRM) and Two-Factor Graded Response Model (TGRM). Results: The TGRM showed the best relative fit. This model was implemented by calculating the conditional and marginal parameters. Likewise, the ECV index was calculated. Conclusions: The PBS items were primarily...

This study seeks to update Argentine psychometric studies. A Computerized Adaptive Test (CAT) was developed from a Verbal Analogies Item Bank (IB); it was programmed with FastTEST Pro version 2.0. A sample of 108 undergraduate students was assessed. Results were compared according to three different CAT termination criteria (fixed length of 32 items, variable length when reaching an estimation error of less than or equal to .40 and to .30). Significantly high correlations were obtained (r > .90; p < .001) between the level of trait estimated from the BI and each of the three CAT variants. Given the BI characteristics, the fixed-length of 32 items for the CAT presented the optimal balance between precision and length.

Este trabajo busca actualizar los estudios psicométricos en la Argentina. Se desarrolló un Test Adaptativo Informatizado (TAI) a partir de un Banco de Ítems (BI) de Analogías Verbales y se programó con FastTEST Pro versión 2.0. Se contó con una muestra de 108 estudiantes universitarios. Se compararon los resultados obtenidos a partir de tres criterios de finalización distintos del TAI (longitud fija de 32 ítems, longitud variable al alcanzar un error de .40 y un error de .30). Las tres variantes obtuvieron correlaciones significativamente altas (r > .90; p < .001) con el nivel de rasgo estimado a partir del BI. El TAI de longitud fija de 32 ítems presentó el balance óptimo entre precisión y longitud dadas las características del BI.

This study seeks to update Argentine psychometric studies. A Computerized Adaptive Test (CAT) was developed from a Verbal Analogies Item Bank (IB); it was programmed with FastTEST Pro version 2.0. A sample of 108 undergraduate students was assessed. Results were compared according to three different CAT termination criteria (fixed length of 32 items, variable length when reaching an estimation error of less than or equal to .40 and to .30). Significantly high correlations were obtained (r > .90; p < .001) between the level of trait estimated from the BI and each of the three CAT variants. Given the BI characteristics, the fixed-length of 32 items for the CAT presented the optimal balance between precision and length.