Criar personas automáticas ou personas manuais? Um estudo comparativo na geração de personas de crianças com transtorno do espectro autista com base em dados reais.

Abstract

The use of the persona technique has been a common practice in user-centered design. Due to this, some studies have been proposing the automatic generation of personas in order to optimize the creation processes, which can be slow and expensive. However, these studies do not explore the possibility of differences in relation to the quality of the personas generated compared to the results obtained in manual methods. Therefore, this paper presents a comparative study of methodologies for automatic and manual persona generation in relation to the quality of the created personas. The study was carried out in the context of creating personas of children with Autistic Spectrum Disorder, created based on real data. The quality of the personas was evaluated by HCI researchers and professionals in the field of health and education. The results indicated that the quality of the automatic personas was greater than the manual personas. However, both methodologies should be improved in order to generate personas with greater completeness, willingness to use, clarity, and empathy.