A Review of Data Gathering Methods for Evaluating Socially Assistive Systems

Area of Science:

• Human-Computer Interaction
• Assistive Technology
• Social Robotics

Background:

• Social interactions are crucial for quality of life, yet individuals with special needs often face loneliness and social isolation.
• There is a growing need for socially assistive systems (SAS) to address the social deficits in vulnerable populations.
• Evaluating SAS is challenging due to data collection difficulties with specific user groups, such as those with dementia or autism.

Purpose of the Study:

• To systematically review and analyze data gathering methods used for evaluating socially assistive systems (SAS).
• To identify the most common target user groups and the data collection strategies employed in SAS research.
• To provide insights for future development and research in SAS evaluation.

Main Methods:

• A systematic literature search was conducted across six major academic databases (Scopus, Web of Science, ACM, Science Direct, PubMed, IEEE Xplore).
• Articles published between January 2000 and July 2021 were included, with 65 articles meeting the final selection criteria.
• Data extraction focused on target users, SAS types, and data gathering methods (interviews, questionnaires, observations).

Main Results:

• The most frequent target users for SAS were individuals with visual impairments, older adults, and children with autism.
• SAS commonly aimed to assist users in perceiving social cues, such as facial expressions.
• Interviews, questionnaires, and observations were the predominant methods for SAS assessment.
• Interviews involved target users, secondary users, or stakeholders; questionnaires measured social interaction, emotional state, and usability; observations focused on older adults and children with autism.

Conclusions:

• This review offers a comprehensive overview of data gathering methods applied to diverse user groups in SAS research.
• Findings illuminate how different evaluation techniques are utilized across various populations, informing best practices.
• The insights derived can guide future research and development of more effective and user-centered socially assistive systems.