서브메뉴
검색
A Computational Approach to Measuring Identity and its Applications in Organizations
A Computational Approach to Measuring Identity and its Applications in Organizations
상세정보
- 자료유형
- 학위논문 서양
- 최종처리일시
- 20250211152118
- ISBN
- 9798384341581
- DDC
- 306
- 저자명
- Yang, Ruo Ying.
- 서명/저자
- A Computational Approach to Measuring Identity and its Applications in Organizations
- 발행사항
- [Sl] : Stanford University, 2024
- 발행사항
- Ann Arbor : ProQuest Dissertations & Theses, 2024
- 형태사항
- 113 p
- 주기사항
- Source: Dissertations Abstracts International, Volume: 86-03, Section: B.
- 주기사항
- Advisor: Goldberg, Amir.
- 학위논문주기
- Thesis (Ph.D.)--Stanford University, 2024.
- 초록/해제
- 요약This dissertation presents a computational measure of identity and explores the interplay between identity and social structure. Identity is central to the human experience. Correspondingly, identity has also been a central concept in the social sciences. While significant theoretical progress has been made in terms of understanding the formation and development of identity over the last few decades, the methodology for measuring identity has been lagging behind. This dissertation attempts to address this issue. In three interrelated chapters, I seek to accomplish two goals. First, I seek to formulate a computational approach to measuring identity that is rooted in extant theory of identity. Second, I seek to illustrate the utility of such an approach by highlighting two separate applications of this methodology in the organizational context.In the first chapter, I introduce a computational approach to measuring identity. I describe the existing work in sociology, psychology, and computational linguistics that has laid the foundation for the conceptualization of my measure. Namely, using the definition of identity as a set of selfreferential meanings from Identity Theory (Stryker and Burke, 2000), I conceptualize identity as the semantic meanings associated with the first-person singular pronoun "I". I quantify these meanings using a class of machine learning models from Computer Science, known as word embedding models, by retrieving the word embedding associated with the word "I". In parallel, I also highlight the potential for using finetuning techniques to address the problem of small training data in the social sciences. I end the chapter with describing how this methodology can be applied and modified in two separate applications, first, to measure identity similarity, and second, to measure group identification. This methodology, as well as the application to group identification, is first established in coauthored work with Dr. Amir Goldberg and Dr. Sameer Srivastava.In Chapter 2, I expand on the use of my computational identity measure to study homophily. Homophily is a fundamental principle that orders and structures social ties. Existing work conceptualizes homophily as a static phenomenon. In the commonly studied case of gender homophily, for instance, two individuals either share the same gender or they do not. However, a core insight in the identity literature is that identities are dynamically enacted as a function of social contexts and interactions. Integrating this insight, I maintain that homophily is also a dynamic, interactional, and contextualized process. Building on prior work, I theorize that similarity in enacted identity predicts tie existence and strengthens existing ties. I further deconstruct enacted identity similarity into its intra-relational and extra-relational components. That is, for each pair of individuals, I distinguish between identity enacted within and outside of the purview of their relationship. Under the contextualized view of identity, intra- and extra-relational enacted identities should diverge, and only intra-relational enacted identity similarity should strengthen social ties. Finally, I contend that the effect of intra-relational enacted identity similarity is amplified when enacted in private contexts, as privacy renders enacted identities more authentic and intimate. I apply my computational identity measure to a proprietary corpus of Slack data. Through analyzing channel membership on Slack, I identify the intra-relational and extra-relational components of enacted identity similarity. Combining this approach with responses from a network survey, I find consistent support for my hypotheses.
- 일반주제명
- Culture
- 일반주제명
- Computer science
- 일반주제명
- Social interaction
- 일반주제명
- Homeless people
- 일반주제명
- Identity
- 일반주제명
- Sociology
- 일반주제명
- Linguistics
- 일반주제명
- Social sciences
- 일반주제명
- Semantics
- 일반주제명
- Logic
- 일반주제명
- Social psychology
- 기타저자
- Stanford University.
- 기본자료저록
- Dissertations Abstracts International. 86-03B.
- 전자적 위치 및 접속
- 로그인 후 원문을 볼 수 있습니다.
MARC
008250123s2024 us c eng d■001000017162969
■00520250211152118
■006m o d
■007cr#unu||||||||
■020 ▼a9798384341581
■035 ▼a(MiAaPQ)AAI31460342
■035 ▼a(MiAaPQ)Stanfordrd969sd3691
■040 ▼aMiAaPQ▼cMiAaPQ
■0820 ▼a306
■1001 ▼aYang, Ruo Ying.
■24512▼aA Computational Approach to Measuring Identity and its Applications in Organizations
■260 ▼a[Sl]▼bStanford University▼c2024
■260 1▼aAnn Arbor▼bProQuest Dissertations & Theses▼c2024
■300 ▼a113 p
■500 ▼aSource: Dissertations Abstracts International, Volume: 86-03, Section: B.
■500 ▼aAdvisor: Goldberg, Amir.
■5021 ▼aThesis (Ph.D.)--Stanford University, 2024.
■520 ▼aThis dissertation presents a computational measure of identity and explores the interplay between identity and social structure. Identity is central to the human experience. Correspondingly, identity has also been a central concept in the social sciences. While significant theoretical progress has been made in terms of understanding the formation and development of identity over the last few decades, the methodology for measuring identity has been lagging behind. This dissertation attempts to address this issue. In three interrelated chapters, I seek to accomplish two goals. First, I seek to formulate a computational approach to measuring identity that is rooted in extant theory of identity. Second, I seek to illustrate the utility of such an approach by highlighting two separate applications of this methodology in the organizational context.In the first chapter, I introduce a computational approach to measuring identity. I describe the existing work in sociology, psychology, and computational linguistics that has laid the foundation for the conceptualization of my measure. Namely, using the definition of identity as a set of selfreferential meanings from Identity Theory (Stryker and Burke, 2000), I conceptualize identity as the semantic meanings associated with the first-person singular pronoun "I". I quantify these meanings using a class of machine learning models from Computer Science, known as word embedding models, by retrieving the word embedding associated with the word "I". In parallel, I also highlight the potential for using finetuning techniques to address the problem of small training data in the social sciences. I end the chapter with describing how this methodology can be applied and modified in two separate applications, first, to measure identity similarity, and second, to measure group identification. This methodology, as well as the application to group identification, is first established in coauthored work with Dr. Amir Goldberg and Dr. Sameer Srivastava.In Chapter 2, I expand on the use of my computational identity measure to study homophily. Homophily is a fundamental principle that orders and structures social ties. Existing work conceptualizes homophily as a static phenomenon. In the commonly studied case of gender homophily, for instance, two individuals either share the same gender or they do not. However, a core insight in the identity literature is that identities are dynamically enacted as a function of social contexts and interactions. Integrating this insight, I maintain that homophily is also a dynamic, interactional, and contextualized process. Building on prior work, I theorize that similarity in enacted identity predicts tie existence and strengthens existing ties. I further deconstruct enacted identity similarity into its intra-relational and extra-relational components. That is, for each pair of individuals, I distinguish between identity enacted within and outside of the purview of their relationship. Under the contextualized view of identity, intra- and extra-relational enacted identities should diverge, and only intra-relational enacted identity similarity should strengthen social ties. Finally, I contend that the effect of intra-relational enacted identity similarity is amplified when enacted in private contexts, as privacy renders enacted identities more authentic and intimate. I apply my computational identity measure to a proprietary corpus of Slack data. Through analyzing channel membership on Slack, I identify the intra-relational and extra-relational components of enacted identity similarity. Combining this approach with responses from a network survey, I find consistent support for my hypotheses.
■590 ▼aSchool code: 0212.
■650 4▼aCulture
■650 4▼aComputer science
■650 4▼aSocial interaction
■650 4▼aHomeless people
■650 4▼aIdentity
■650 4▼aSociology
■650 4▼aLinguistics
■650 4▼aSocial sciences
■650 4▼aSemantics
■650 4▼aLogic
■650 4▼aSocial psychology
■690 ▼a0290
■690 ▼a0984
■690 ▼a0626
■690 ▼a0800
■690 ▼a0395
■690 ▼a0451
■71020▼aStanford University.
■7730 ▼tDissertations Abstracts International▼g86-03B.
■790 ▼a0212
■791 ▼aPh.D.
■792 ▼a2024
■793 ▼aEnglish
■85640▼uhttp://www.riss.kr/pdu/ddodLink.do?id=T17162969▼nKERIS▼z이 자료의 원문은 한국교육학술정보원에서 제공합니다.


