Un estudio sobre la estabilidad temporal del MMPI-A con un diseño test-retest en estudiantes mexicanos

• Autores: Consuelo Durán Patiño, Jose M. Pérez y Farías, Emilia Lucio Gómez-Maqueo
• Localización: Salud mental, ISSN 0185-3325, Vol. 26, Nº. 2, 2003, págs. 59-66
• Idioma: español
• Enlaces
  • Texto completo (pdf)
• Resumen
  • español

    A raíz de los resultados obtenidos en diferentes investigaciones y de los problemas que enfrentaban los usuarios del MMPI, fue posible detectar algunas fallas en la versión original del instrumento, sobre todo en relación con la muestra normativa, así como en la redacción y el contenido de los reactivos. Después de una labor sistemática de revisión y estandarización iniciada en 1989, en 1992 se publica la versión para adolescentes, el MMPI-A.

    En México, en lo relativo a la construcción de normas, en el campo de la investigación con el MMPI-A existen estudios a partir de los cuales se establecieron normas para la utilización con la muestra de adolescentes mexicanos (21). La presente investigación tuvo como objeto examinar la estabilidad de las Escalas Clínicas y de Validez, de Contenido y Suplementarias del MMPI-A, con un diseño test-retest a siete, catorce y veintiún días. Se utilizaron una muestra de 1056 adolescentes de la Ciudad de México y tres submuestras asignadas aleatoriamente para los retests, con el fin de minimizar las variables extrañas propias del diseño. Las correlaciones entre el test y los retests son altas. Esto permite concluir que el Inventario Multifásico de la Personalidad de Minnesota, en su versión para adolescentes, tiene estabilidad en sus Escalas Clínicas y de Validez, de Contenido y Suplementaria en esta muestra. Estos valores son parecidos a los obtenidos en otros trabajos de investigación con diferentes muestras, tanto en México como en otros países. Además, las correlaciones de Pearson fueron estadísticamente significativas(> 0.75). Podría decirse que las correlaciones tienden a decrecer ligeramente a medida que se amplía el intervalo de los retests, pero se requerirá más investigación al respecto para establecer una aseveración así. Esto permite utilizar el MMPI-A con mayor seguridad de obtener datos confiables y objetivos cuando se utilice para medir muestras similares de adolescentes mexicanos.

    El hecho de que este estudio se haya realizado con una muestra de adolescentes mexicanos permitirá no sólo ofrecer datos sobre su confiabilidad, sino hacer comparaciones con otros estudios, como el que se reporta en el suplemento del manual para la población hispánica dentro de Estados Unidos. Finalmente, esta investigación aporta datos en cuanto a la confiabilidad de este instrumento en una muestra de adolescentes mexicanos. Cabe recalcar que este estudio no sólo añade datos basados en una muestra muy grande a diferencia de los encontrados en la bibliografía, sino que también utiliza un diseño test-retest con una aplicación inicial y tres subsecuentes, mientras que en la bibliografía sólo se encontraron investigaciones que utilizaron un diseño test-retest con dos aplicaciones, que normalmente son de uno a siete días.

  • English

    The Minnesota Multiphasic Inventory (MMPI) is a worldwide applied instrument; it is used in more than 65 countries in clinical and research settings. Surveys indicate that it has become the most widely used personality inventory among practitioners. However, this inventory has been criticized for its limitations in the normative group and for the outdated, sexist, and awkward wording of some items. Therefore, it has recently undergone a revision and a re-standardization. The MMPI has been used in psychiatric and clinical settings with adolescents and adults. It has been used with different groups of adolescents with medical problems, as well as groups in alcohol and drug abuse programs. It has also been widely used for research in correctional or juvenile delinquency programs. The objective of the present study was to assess the test-retest reliability after three different time periods and in three different groups of Mexican students. Even though there are studies with the MMPI-A and a test-retest design is found in the literature with Hispanic samples within the United States, no studies could be found with interval testing ranges of one test and three retests like this one. In addition, most of the other studies were conducted with smaller samples. Thus, data offered by this study could be useful for the MMPI-A research field, and for clinical purposes. Participants included 1056 pathology-free (a exclusion criterion for pathological profiles was included), middle class, high school adolescent students, aged 14 to 18 years (mean age: 16). There were 645 females and 411 males. Their grade average for the previous year was between 70 and 80 in a scale from 0 to 100. In the first group (1-7 days interval testing), there were 147 (38.1%) in tenth grade, 239 (61.9%) in eleventh grade and none in twelfth grade. For the sample of 1-14 days of interval testing, there were 139 (44%) in tenth grade, 106 (33.5%) in eleventh grade and 71 (22.5%) in twelfth grade. For the administration interval of 1-21 days there were 173 (48.8%) in tenth grade, 145 (41.0%) in eleventh grade and 36 (10.2%) were twelfth grades. The MMPI-A Spanish version was administered to all subjects. The instructions contained in the MMPI-A group form test booklet were read aloud by session conductors while participants followed.

    The MMPI-A is a 478 item paper and pencil test. Revised and new features of the test include: a) implementation of new normative data obtained with a sample (N=2,600) that approximates the 1980 US census data, b) replacement and/or rewording of outdated items, c) development of uniform (rather than linear) T scores, d) construction of inconsistency scales and a supplementary F Back (Fb) Scale, and e) development of new content Scales. A total of 1056 students who attended a psychology class were tested from 1200 students registered in two high schools. After participants were initially tested, three subgroups randomly assigned from the original group were retested. Informed consent was obtained, and anonymity was kept. The MMPI-A was thus administered twice to each individual in small group testing sessions (25 students), with a mean retest interval of 7, 14, and 21 days. Twelve subjects were excluded from data analyses because they were 19 years old. A total of 132 students did not attend the first testing session, making up a total of 144 students who were not included in this study. The present study examined the reliability of the MMPI-A within the Clinical and Validity, Content and Supplementary Scales with a Test-Retest design, with an interval range of administrations of 1-7, 1-14, and 1-21 days with randomized subgroups in an attempt to prevent extraneous variables in the design. When mean comparisons (first time vs. second time) of the scales were made, no significant differences were found in most of them. Nevertheless, in all cases the mean differences were less than 4 T-score points, which would be interpreted as clinically insignificant. Moderate to high stability coefficients were obtained for the Validity and Clinical, the Content and the Supplementary Scales; most alpha coefficients ranged from 0.56 to 0.91, falling most of them in the higher end. Concerning sex, the results obtained revealed no significant differences between the test and the three retests. There were, however, differences in the profiles for each sex group, which may be explained by differences between genders. These values are similar to those reported in the manual for the Hispanic version of the MMPI-A for the US, and in the manual for the MMPI-A. In Mexico, the results are similar to those reported for the Mexican adolescents sample (14). Since the Pearson correlation coefficients were statistically significant, we can rely on the fact that the MMPI-A correlates substantially in any of the interval time ranges used as “temporality” variables. In the present study different retest intervals were employed, whereas a one-week interval was used in the studies previously mentioned. A larger sample was used in this study as compared to those in others, and this may give more reliability to the data. Likewise, other information may be added to the background of this instrument that can be useful with Mexican-American samples within the US and other countries. Such additional information can also be useful with Latin American samples because of their similar cultural background. This study may also suggest that MMPI-A reliability may decline slightly as the retest interval increases. Given the limited data base presently available, further studies are expected to follow.