Three tasks for the exploration of verbal fluency: evidence of test-retest reliability in Argentine adults
*Correspondencia: Dr. Pablo Martino. Laboratorio de Investigaciones en Ciencias del Comportamiento. Facultad de Psicología. Universidad Nacional de San Luis. Avenida Ejército de los Andes, 950. C.P. 5700 San Luis, Argentina.
E-mail: p.martino@hotmail.com
Introduction: The phonological and semantic verbal fluency tasks are frequently used in neuropsychological assessment due to their easy application and good sensitivity to dementia. In Argentina, the psychometric evidence for these tasks is limited, with a special lack of knowledge of the temporal stability of its measurements. The psychometric production is even lower for the action fluency variant (emission of verbs in the infinitive in one minute). In effect, this research analyzes the test-retest reliability of three verbal fluency tasks in Argentine adults.
Subjects and methods: The sample was made up of 85 Argentine (average age, 63.7 years), 75,3% women and with a medium-high educational level. A prospective longitudinal design was carried out, administering phonological, semantic and action fluency tasks at two different times with an interval of up to four months. The intraclass correlation coefficient (ICC), a statistical method suggested for test-retest reliability studies, was analyzed. For the interpretation of the ICC, the Fleiss criteria were adopted.
Results: The phonological and semantic fluency tasks showed good reliability, with ICCs of 0.77 and 0.79. The fluidity of action variant yielded ICC of 0.90, indicating excellent reliability.
Conclusions: All fluency tasks have appropriate temporal stability, and their use is recommended when prospective neuropsychological research is planned (with language evaluation at different times) or as a method of monitoring the evolution of aphasic patients undergoing neurorehabilitation. Based on its excellent reliability, it is recommended to use the action variant more frequently.
Sujetos y métodos La muestra se compuso de 85 argentinos (medida de edad, 63,7) de población general no clínica, un 75,3% mujeres, de nivel de instrucción medio-alto. Se efectuó un diseño longitudinal-prospectivo administrando tareas de fluidez fonológica, semántica y de acción en dos momentos distintos con un intervalo hasta de cuatro meses. Se analizó el coeficiente de correlación intraclase (CCI), método estadístico sugerido para estudios de fiabilidad test-retest. Para interpretar el CCI se adoptaron los criterios de Fleiss.
Resultados Las tareas de fluidez fonológica y semántica demostraron buena fiabilidad, con un CCI de 0,77 y 0,79. La fluidez de acción obtuvo excelente fiabilidad, con un CCI de 0,9.
Conclusión Las tareas de fluidez relevadas poseen apropiada estabilidad temporal, por lo que se sugiere su uso en investigaciones neuropsicológicas prospectivas (cuando se evalúe el lenguaje en distintos momentos) o cuando se requiera un seguimiento de la evolución de pacientes afásicos en neurorrehabilitación. Basándose en su excelente fiabilidad, se recomienda utilizar con más frecuencia la variante de acción.