Browsing by Author "Howie, Nicole"

Now showing 1 - 2 of 2
  • Thumbnail Image
    Item
    The Effect of Response Modality on Task Performance When Using an Interactive Voice Response System for Older and Younger Adults
    (2013-04-18) Howie, Nicole; Byrne, Michael D.; Logan, Jessica; Kortum, Philip
    The purpose of this experiment was to gain a greater understanding of how set- level compatibility affects task performance of older adults while using an Interactive Voice Response (IVR) System. Set-level compatibility effects occur when differences in response time are obtained due to characteristics shared between a stimulus and response set. The current study used banking and healthcare IVR tasks to investigate whether manipulating set-level compatibility would differentially affect performance between younger (18 - 39 years old) and older (65 years and older) adults. As expected, subjects performed better in the high set-level compatibility conditions. However, no interaction between set-level compatibility and age was found, which can possibly be attributed to task differences, IVR design and real world context of the task. Overall, the primary contributions of this research are the design guidelines that were created to develop an IVR that suppressed age effects.
  • Thumbnail Image
    Item
    The Generalizability of Cognitive Modeling Parameters
    (2015-07-06) Howie, Nicole; Byrne, Michael; O'Malley, Marcia; Kortum, Philip
    Increased awareness of the importance of usability has stemmed from the realization that customer satisfaction and revenue generation are affected by usability. One method that can be used to evaluate usability is cognitive modeling, which can make quantitative predictions about human performance across different tasks and devices. However, it is unclear if cognitive models can accurately predict older adult performance. Trewin et al. (2012) proposed that parameters used in models for older adults may have to be specific to a task and/or device. The purpose of this research was to determine if task and/or device type must be accounted for in parameters used to model older adult performance in a cognitive architecture. Jastrzembski and Charness (2007) estimated architectural parameters for older adults and were able to successfully predict performance of tasks performed with a mobile phone. The current study investigated if these parameters generalize to different tasks and devices. In the experiment, older (70 years and above) and younger (18 - 39 years old) adults performed two tasks with two different devices. Overall, the results from the behavioral data showed that older adults performed more slowly than younger adults, however older adult’s performance varied across task. Performance differences between older and younger adults due to device were caused by strategy differences. Cognitive models of each task with one device were created using modeling parameters that represented older and younger adults. Then the behavioral data were compared to the models. The models were mainly slower than older and younger adults across each task. The results helped provide evidence that task and/or device type are important and should be incorporated into modeling parameters. However, strategy must be accounted for as well in order to accurately model older adult performance.