VUIs: A Two-Part Study

An exploratory qualitative study on Voice User Interface (VUI) use, followed by a derivative quantitative study focused on Internet of Things (IoT) devices' relationship with VUIs.

I collaborated with a group of graduate-level researchers at Rochester Institute of Technology to produce and execute two related HCI studies. The studies specifically focused on consumer voice assistant technologies and different patterns of use. Each study took place over the course of a 4-week period and culminated in an ACM-style research article describing our methods, results and conclusions. A general overview of each study is broken down below:

    (1) “A Phrase in Mind”: A Case Study of Multitasking with Voice User Interface — An exploratory qualitative study following a grounded theory analysis model. Our research goal for the study was: "Explore how individuals employ voice technologies to multitask and understand their preferences for these technologies within different contexts of use." Each researcher conducted three VUI task walthroughs and semi-structured interviews, and afterwords we collectively coded the transcripts semantically. We then collaboratively visualized and grouped our semantic codes into 11 latent themes, and further distilled those themes into 6 over-arching categories: (1) Options/Choices; (2) User Inputs; (3) Limitations; (4) Physical Interfaces; (5) Chosen Technology; and (6) System Outputs.
    (2) A Rich, Device-Rich Life: A Quantitative Analysis of Voice Interface Usage and Perception — A quantitative, survey-driven study. Our research question for this study was: "How does a user’s quantity of connected smart home devices affect their frequency of use, perceived utility, and perceived accuracy of voice assistant technologies?" We created a brief survey using the Qualtrics platform consisting of 17 close-ended questions that yielded ordinal and continuous quantitative data. We closed the survey after receiving 64 responses (59 usable responses, after cleaning the data). We then used R to carry out multiple rounds of nonparametric inferential statistics tests, attempting to identify a significant relationship between a user's quantity of home IoT devices and their frequency and perception of VUI use. Our analysis concluded with 4 significant test results, of which we were able to triangulate 3 of these using additional statistical tests. We found a very strong, positive relationship between a user's quantity of IoT devices and their frequency of VUI use; additionally, we identified a significant difference in perceived accuracy of familiar commands, perceived effect on life and users' likelihood to recommend VUIs between respondents with few IoT technologies (zero to three) vs. those with several IoT technologies (four or more).

You can read the full research papers resulting from the two studies below.

  • — category: UX/UI, Information Design / Research
  • — class: Research Methods
  • — year: 2021
  • — for: M.S. in HCI