A Principled Account of the Older Adult in ACT-R: Age-Specific Model Human Processor Extensions in a Mobile Phone Task

Abstract

In previous research, Jastrzembski & Charness (2007) estimated weighted mean values for nine information processing parameters for older adults using the Card, Moran, & Newell (1983) Model Human Processor (MHP) model, and successfully validated a subset of those parameters with age-specific GOMS models for two mobile phone tasks across two mobile phone devices. The current research extends the mapping of MHP parameters to the ACT-R cognitive architecture (e.g., Anderson, 2007), and transitions age-specific MHP parameters into theoretically-plausible ACT-R models of older adult performance. We argue that the incorporation of theoretically-motivated MHP parameters into ACT-R will produce better fits to empirical data than default ACT-R parameters, and the use of those theoretically-principled parameter values will provide designers with more precise insight concerning why hardware/software designs/devices work or fail as a function of age. We test these hypotheses in the same mobile phone domain used to validate the age-specific MHP parameters in previous research and our findings demonstrate that age-specific ACT-R models using elemental MHP parameter values achieve better fits than default ACT-R parameters. This research may provide designers and human factors engineers the ability to inspect details of human performance at a finer grain of resolution than is currently available, help determine the cause of specific errors resulting from hardware/software design, cognitive workload, or user characteristics, and support the development of age-sensitive technologies.