El objetivo de este estudio instrumental fue determinar las propiedades psicometricas del test de cribado de demencias Pesotest en muestras clinica y no clinica de adultos mayores de Bogota. Los participantes fueron 213 adultos mayores de 65 anos. Los instrumentos utilizados fueron el Pesotest, el Mini Mental State Examination (MMSE), la Encuesta Sociodemografica y la Escala de Depresion Geriatrica de Yesavage (GDS). La consistencia interna del Pesotest obtenida con el Alpha de Cronbach fue de 0.86. Se uso el MMSE para establecer la validez convergente y se encontro una correlacion significativa de 0.596. Se hallaron diferencias estadisticamente significativas entre las muestras clinica y no clinica. El analisis factorial arrojo cuatro factores: mediante la curva ROC se encontro que la sensibilidad o proporcion de personas diagnosticadas con demencia que dieron positivo fue de 0.83; la especificidad o proporcion de personas sanas que segun la prueba puntuan sin demencia fue de 0.78, y el punto de corte fue 18. El analisis con el modelo de Rasch evidencio que solo un item no se ajusto al modelo, y que el item con mayor dificultad era el que evaluaba memoria y calculo, cuya alteracion indica evolucion a demencia. Con base en los resultados favorables del analisis psicometrico, adecuada consistencia interna, validez convergente, validez de criterio y validez de constructo, se recomienda utilizar el Pesotest en servicios de atencion primaria. Abstract The aim of this instrumental study was to establish the psychometric properties of the screening test for dementia Pesotest in clinical and non-clinical samples of elderly people. The participants of this study were 213 elderly people over 65 years old. The instruments used were the Pesotest, the Mini Mental State Examination (MMSE), the Sociodemographic Survey and the Geriatric Depression Scale Yesavage (GDS). A Cronbach’s alpha of 0.86 for internal consistency of the Pesotest was found. The MMSE was used to establish convergent validity, finding a significant correlation of 0.596. Statistically significant differences were also found between the clinical and nonclinical samples. The factor analysis yielded four factors and using the ROC curve it was found that the sensitivity, or the proportion of people diagnosed with dementia, was 0.83; and the specificity, or the proportion of healthy people that were diagnosed with dementia by the test, was 0.78, with a cut-off point of 18. The analysis with the Rasch Model showed that only one item was not fit and that the most difficult item was the one which assessed calculation and memory, whose alteration indicates an evolution towards dementia. Based on the favorable results of the psychometric analysis, suitable internal consistency, convergent and construct validity and reliability, the use of the Pesotest in primary care services is recommended.