Show simple item record

dc.contributor.advisorNunamaker, Jay
dc.contributor.advisorBurgoon, Judee
dc.contributor.authorPentland, Steven J.
dc.creatorPentland, Steven J.
dc.date.accessioned2018-10-12T01:01:34Z
dc.date.available2018-10-12T01:01:34Z
dc.date.issued2018
dc.identifier.urihttp://hdl.handle.net/10150/630161
dc.description.abstractThe human body provides a wealth of information that, when captured and analyzed, offers deep insight into the mind and its processes. At no other point in history has this information been as accessible as it is today. Using various sensor systems, researchers can now efficiently capture fine-grain behavioral data and leverage it for scientific insight. This dissertation begins by reviewing the capture and use of human-data from an information systems perspective in which the objective is to provide organizational value. The dissertation then proposes a scalable interview system for the collection and analysis of verbal and nonverbal human behaviors. Following design science principles, a proof-of-concept prototype system is created and evaluated in the context of personnel-selection. The prototype system comprises a highly structured interview paradigm and uses a standard web-camera to record interviewees. Experiment 1 evaluates the system’s ability to replicate subjective human judgements of source credibility. Experiment 2 then assesses the system’s ability to predict objective measures of general mental ability and job knowledge. During each experiment, study participants conduct mock job interviews using the prototype system. Participants respond to a series of interview questions related to a mock-job description. Behavioral features are extracted from facial displays, voice characteristics, and language usage captured by video recordings. In Experiment 1, participant performance is assessed using third-party raters. In Experiment 2, participants complete computerized assessments of general mental ability and job skills following the interview. Assessments and behavioral measures are then processed with predictive machine learning algorithms. The results indicate that subjective and objective measures of job performance can be inferred at rates considerably above chance using automated analysis of human behaviors. This research provides insight into the design principles that allow for human-analytics to become part of organizations’ day-to-day processes.
dc.language.isoen
dc.publisherThe University of Arizona.
dc.rightsCopyright © is held by the author. Digital access to this material is made possible by the University Libraries, University of Arizona. Further transmission, reproduction, presentation (such as public display or performance) of protected items is prohibited except with permission of the author.
dc.subjectbehavioral informatics
dc.subjectbehavioral sensing
dc.subjectemployee sifting
dc.subjecthuman-analytics
dc.subjectpersonnel selection
dc.titleHuman-Analytics in Information Systems Research and Applications in Personnel Selection
dc.typetext
dc.typeElectronic Dissertation
thesis.degree.grantorUniversity of Arizona
thesis.degree.leveldoctoral
dc.contributor.committeememberNunamaker, Jay
dc.contributor.committeememberBurgoon, Judee
dc.contributor.committeememberValacich, Joe
thesis.degree.disciplineGraduate College
thesis.degree.disciplineManagement
thesis.degree.namePh.D.
refterms.dateFOA2018-10-12T01:01:34Z


Files in this item

Thumbnail
Name:
azu_etd_16434_sip1_m.pdf
Size:
4.167Mb
Format:
PDF

This item appears in the following Collection(s)

Show simple item record