Psychological identity in a digital world: Detecting and understanding digital traces of our psychological self

Lead Research Organisation: UNIVERSITY OF EXETER
Department Name: Psychology

Abstract

Boundaries between digital technologies and ourselves become blurred as technology is integrated into our work, home and even our bodies. Interdisciplinary research is needed to understand how our sense of self - our psychological identity - affects and is affected by technology use. During this fellowship, I will lead an interdisciplinary team of psychologists and computer scientists to explore how our identity shapes and is shaped by technology in the fields of security and healthcare. This work will be underpinned by a programme of collaboration with industry partners, Polaris Consulting, the National Crime Agency (NCA), Dstl, Milton Keynes University Hospital (MKUH) and EDP Drugs and Alcohol Services to explore applications that support rather than replace human analysts and consultants.

Sophisticated technologies are rapidly becoming part of our lives. Research on digital technology needs to urgently address concerns about privacy, trust, and ethical implications. I will extend current research in this area by considering how our different psychological identities shape what we find acceptable in different situations. For instance, a person might be less concerned about the tracking of personal information when thinking of themselves as a patient rather than as a parent. Throughout the fellowship, I will work closely with user groups (e.g., patients in the Lived Experience Group Exeter), industry partners and the general public to understand privacy concerns and privacy behaviour.

I will also continue to develop the capacity to detect psychological identities from naturally occurring digital data (e.g., forum posts, blogs, e-mails). This research will allow us to understand which psychological identity (e.g., parent, addict, criminal network identity) is relevant in a particular situation. I will extend my current work to test whether it is possible to distinguish between several identities by analysing text data, whether detecting identities in text is robust to deception, and whether it is possible to tell how committed an individual is to their group from the way in which they communicate online. I will work closely with my industry partners, Polaris, NCA and Dstl, to explore how findings can enhance current machine learning capabilities and analytic approaches in defence and security.

Finally, building on the identity detection work, I will examine how individuals develop new psychological identities (e.g., becoming a parent) and leave identities behind (e.g., leaving behind an addict identity during therapy), and the consequences of such transitions for mental health (e.g., post-natal depression, addiction recovery). I will work closely with industry partners MKUH and EDP to explore how these findings can be translated into diagnostic and monitoring solutions of the future that augment the work of therapists and medical consultants.

The project will be integrated with research on software engineering through the EPSRC SAUSE platform grant. It will be conducted at the psychology department of the University of Exeter, which has a long and successful history of high-impact research. The project will draw on the strengths of the Social, Environmental and Organisational Psychology Research Group (SEORG), which is world-leading in research on social identity, privacy, and well-being, and the Clinical and Cognitive Research Groups, which are world-leading on depression and addiction. The University of Exeter also fosters interdisciplinary work, for instance through the co-supervision of EPSRC students across colleges, access to world-leading experts on machine learning and data science at the Alan Turing Institute and experts in healthcare at the Wellcome Centre for Cultures and Environments of Health and the EPSRC Centre for Predicitive Modelling in Healthcare. Taken together this project will establish Exeter as a key centre for EPSRC work on psychological identity and digital technologies.

Planned Impact

The main impact aim of this fellowship is to create social and economic benefits to the UK in healthcare and security by combining social psychology with computational linguistics and machine learning.

In particular, the proposed research will contribute to the development and application of AI by understanding how psychological identities shape privacy and, in turn, are shaped by technology use. Importantly, the proposed research will ultimately provide human analysts in security and healthcare with additional information on the psychological identities of target populations, thus augmenting rather than replacing humans. The research therefore addresses the EPSRC/NPIF industrial strategy area "High productivity services through artificial intelligence, data and digital technologies".

Psychological identities are closely related with several mental health problems. The proposed research aims to contribute to digital technology that transforms the delivery of healthcare by anticipating and preventing specific mental health problems (e.g., post-natal depression) and providing opportunities for monitoring patients outside clinical settings (e.g., during addiction therapy), thereby enabling independent living. Poor mental health has effects well beyond the individual: A recent government report estimates the costs to employers between £33-42 billion per year and annual costs to the UK economy at around 4.5% of GDP. The fellowship research addresses the EPSRC/NPIF area "Leading edge healthcare and medicines".

In the security and defence sector, I will work closely with Polaris Consulting to provide UK defence and security analysts with additional information based on the proposed psychological identity detection work to enable them to prioritise and act on relevant information. Outputs generated within this fellowship are anticipated to form part of future software packages. Work with Polaris as well as Dstl serve as a potential route to market for some of the project outputs - with potential to licence IP to Polaris for exploitation of my work through their software. The project will extend current work with the NCA on criminal identities as part of an EPSRC-NPIF studentship that started in September 2017. Working closely with the NCA from the start of the project will ensure that the findings can be translated into appropriate support for their operational and preventative work.

In the healthcare sector, I will work closely with Milton Keynes University Hospital (MKUH), EDP Drugs and Alcohol Services and the Lived Experience Group in Exeter to work towards detecting psychological identities relevant to the healthcare context (e.g., addiction identity) from written text. This research ultimately aims to help healthcare providers to achieve cost savings through early detection of (mental) health issues and cost-effective ways to identify and monitor patients remotely.

The planned research will impact on a variety of stakeholders including industry, professional communities of practice and the general public. Throughout the research I will engage with these stakeholders through:
- Patient and public involvement including focus groups, e.g., MKUH, Lived Experience Group
- Steering groups of stakeholders and academic experts in the field
- Research exchange visits with industry partners Polaris and MKUH
- Joint funding applications with industry partners Polaris and MKUH
- Exploration of different routes to market with industry partners, including IP licencing
- Development of proof of concept solutions with software engineering contacts from EPSRC platform SAUSE
- Public engagement, e.g., Soapbox Science, Pint of Science, @Bristol etc. and social and mainstream media
Additional impact will arise through career development for early career researchers (i.e. the fellow and her team of a PDRA and PhDs) including training and visits to leading industry and academic research labs (e.g., University of Queensland.

Publications

10 25 50
 
Description The key findings from this award are:
(1) a toolkit and tutorial that allows us to detect which group membership (identity) is active in a particular moment through the analysis of writing style;
(2) a validated model (based on 4 studies) that can detect whether a parent identity is active in a particular moment;
(3) a longitudinal study of forum posts (N = 326 mothers, across 13 months) that indicates that a parent identity develops differently in first-time mothers with perinatal mental health difficulties than those without such difficulties;
(4) a model that can detect whether a depression or a drug user identity is active in a particular moment;
(5) a longitudinal study of forum posts (N = 443 recovering addicts, 3 years from beginning of recovery) that shows that the risk of relapse during opioid recovery is lower for those who increase their participation in a larger number of online groups;
(6) a longitudinal study of forum posts that shows that engagement with different groups (drug-related, recovery-related, non-addiction groups) affects relapse in different ways;
(7) a longitudinal study of forum posts that shows that engagement with different groups (drug-related, recovery-related, non-addiction groups) affects the development of a recovery identity in different ways;
(8) a set of four studies that shows that privacy behaviour in terms of information disclosure differs when different social identities are active;
(9) a longitudinal study of parents of autistic and non-autistic children (N>200) that shows differences in parent identity across the child's ages of 3-5 years;
(10) a set of three online experiments examining whether those who pretend to have a particular social identity ("impostors") write in a different style and change content systematically compared with those who write in their actual social identity;
(11) a longitudinal study of a pro-eating disorders forum examining how the writing style of participants shows their level of identification with the forum, and how this, in turn, relates to self-esteem, mental wellbeing, and weight loss;
(12) a longitudinal study of the shift in writing style in the Republican party Reddit forum in relation to Reddit forums on the conservative to right-wing spectrum (incl. The Donald) during the Trump candidacy and presidency
Exploitation Route Academic:
(1) the toolkit and tutorial allow other researchers to create classification models for specific identities they are interested in
(2) collaborations with clinical and cognitive psychologists and computer scientists (e.g, University of Queensland, Nottingham Trent, UCL, Imperial) that have led to new research and funding applications
(3) presentations at international conferences, incl. EASP 2020 (social psychology), IC2S2 2020 (computational social sciences), ICSIH 2020 (social and clinical psychology) have increased the visibility of the PsyID lab and our research internationally
(4) invitation by the European Commission Joint Research Centre as an expert on social media and identity leading to a new research project

Non-academic:
(1) Collaboration with National Crime Agency (impact partner) with an Impact Accelerator Award application in preparation
(2) Meetings with dstl (impact partner)
(3) Investment from Awakn Life Sciences via an MRC Confidence in Concepts grant in our research to help evaluate their innovative alcohol therapy programme with the tool we have created to detect a drug/recovery identity
Sectors Communities and Social Services/Policy

Digital/Communication/Information Technologies (including Software)

Healthcare

Security and Diplomacy

URL http://blogs.exeter.ac.uk/identitrack/
 
Description Findings from this research have led to more applied research projects with close stakeholder/industry partnerships: 1) Security: the findings have been used to create a toolkit for understanding hybrid identities such as eco-fascism in online forums through a CREST follow-on project 2) Health: the addiction/depression classifier from this project is now used to help evaluate a stage II randomised controlled trial for the treatment of alcohol use disorder (with Awakn Life Sciences) 3) Wellbeing: integration of automatic identity detection into an online meeting tool for wellbeing (with Astral Dynamics via an ESRC impact accelerator awar 4) Computational Social Psychology: The project has led to interest in computational social science techniques being used in traditional areas of social psychology; the pioneering influence of this project has been recognised in my promotion to Associate Professor for Computational Social Psychology as well as my appointment as Interim Director of the Institute for Data Science and AI (IDSAI) at the University of Exeter
First Year Of Impact 2022
Sector Digital/Communication/Information Technologies (including Software),Healthcare,Security and Diplomacy
Impact Types Societal

 
Description Detecting Hybrid Social Identities: A Computational Analysis Of Influence And Resilience In Online RWE Communities
Amount £124,971 (GBP)
Organisation Centre for Research and Evidence on Security Threats 
Sector Academic/University
Country United Kingdom
Start 12/2021 
End 11/2022
 
Description Imperial Confidence in Concept
Amount £198,292 (GBP)
Funding ID MC_PC_18050 
Organisation Medical Research Council (MRC) 
Sector Public
Country United Kingdom
Start 06/2019 
End 06/2022
 
Description MORE-KARE: A multi-centre investigation of increasing alcohol abstinence with ketamine-assisted psychological therapy in severe alcohol use disorder
Amount £1,038,054 (GBP)
Funding ID NIHR150193 
Organisation National Institute for Health Research 
Sector Public
Country United Kingdom
Start 03/2023 
End 11/2026
 
Description Social Identity Switching: How effective is it and how much control do we have over social identity switches?
Amount £73,887 (GBP)
Funding ID 2074877 
Organisation Engineering and Physical Sciences Research Council (EPSRC) 
Sector Public
Country United Kingdom
Start 09/2018 
End 05/2022
 
Title Automated Social Identity Assessment (ASIA) - tutorial 
Description Automated Social Identity Assessment (ASIA) allows researchers to build a classifier that detects whether a specific identity is active (salient) from written text (e.g., online posts). It includes a step-by-step tutorial on an example tool (parent v feminist identity). 
Type Of Material Model of mechanisms or symptoms - human 
Year Produced 2020 
Provided To Others? Yes  
Impact The tool allowed us to study - differences in parent identity development for perinatal mental health - predictors of recovery and relapse in recovering opioid users 
URL https://github.com/enaserianhanzaei/Tutorial-on-salient-social-Identity-detection-model
 
Description Awakn Life Sciences 
Organisation Awakn Life Science
Country Canada 
Sector Private 
PI Contribution Support with validating alcohol therapy trials
Collaborator Contribution Access to data
Impact Industry partners on a grant (MRC Confidence in Concepts Translational Research Grant)
Start Year 2021
 
Description Care Circles online platform 
Organisation University of Sussex
Department School of Psychology Sussex
Country United Kingdom 
Sector Academic/University 
PI Contribution We supported the development of a novel online communication platform using social identity principles to allow individuals to switch their social identity away from a work/learner identity towards one focussed on mental wellbeing.
Collaborator Contribution Provided initial online communication platform, software development and server.
Impact ESRC impact accelerator account (University of Exeter) Two journal publications in prep Multidisciplinary: software engineering, cognitive psychology, social psychology
Start Year 2022
 
Description Dstl workshop 
Form Of Engagement Activity Participation in an activity, workshop or similar
Part Of Official Scheme? No
Geographic Reach National
Primary Audience Industry/Business
Results and Impact Workshop between dstl group leads and Exeter academics. Agreement was reached to establish a best practice network between dstl Influence group and my research group.
Year(s) Of Engagement Activity 2018
 
Description European Commission Joint Research Centre Expert Workshop on Social Media and Identity 
Form Of Engagement Activity A formal working group, expert panel or dialogue
Part Of Official Scheme? No
Geographic Reach International
Primary Audience Policymakers/politicians
Results and Impact 2-day workshop between academics and members of the European Commission Joint Research Centre
Year(s) Of Engagement Activity 2021
 
Description FUTURES European Researchers Night stall 
Form Of Engagement Activity Participation in an activity, workshop or similar
Part Of Official Scheme? No
Geographic Reach Regional
Primary Audience Public/other audiences
Results and Impact A stall at a pop-up public science event for FUTURES European Researchers Night (an all-day event on the weekend). We provided an interactive demo where members of the general public could try out which identity they conveyed through their writing style; the stall was voted "most inspiring research" by visitors to the event.
Year(s) Of Engagement Activity 2021
URL https://www.agile-rabbit.com/futures-2021/
 
Description Futures 2020 - European Researchers Night 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach International
Primary Audience Public/other audiences
Results and Impact Online interactive talk with 30 participants from the general public about our latest research findings, with sparked questions and discussion afterwards and very positive feedback about a change in view and behaviour.
Year(s) Of Engagement Activity 2020
URL https://futures2020.co.uk/events/who-do-we-think-you-are-online-social-identity-detection/
 
Description IDSAI Day of Ideas industry engagement event 
Form Of Engagement Activity Participation in an activity, workshop or similar
Part Of Official Scheme? No
Geographic Reach National
Primary Audience Industry/Business
Results and Impact IDSAI day of ideas was a high-profile industry-facing event about data science at Exeter with invited industry and business partners, journalists, charities and politicians. My lab group provided information at a stall about computational social science (1 of 4 stalls at the event).
Year(s) Of Engagement Activity 2022
 
Description PSY ID research blog 
Form Of Engagement Activity Engagement focused website, blog or social media channel
Part Of Official Scheme? No
Geographic Reach International
Primary Audience Industry/Business
Results and Impact Lab website featuring short descriptions of our research and links to open access publications - it has generated enquiries from businesses and postgrad students interested in working with us
Year(s) Of Engagement Activity 2020
URL http://blogs.exeter.ac.uk/identitrack/
 
Description Pop-up curiosity school visit 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach Local
Primary Audience Schools
Results and Impact Workshop at a Devon grammar school as part of a wider outreach initiative to get about 50 students excited in computational social science and data science more widely.
Year(s) Of Engagement Activity 2023
 
Description Press release for Koschate, Naserian, Dickens, Stuart, Russo & Levine (2021) 
Form Of Engagement Activity A press release, press conference or response to a media enquiry/interview
Part Of Official Scheme? No
Geographic Reach International
Primary Audience Media (as a channel to the public)
Results and Impact Press release for a key paper of the project which was picked up by 11 news outlets (according to Altmetric) including the Times of India, Hindustan Times, and several Spanish daily newspapers.
Year(s) Of Engagement Activity 2021
URL https://www.exeter.ac.uk/news/homepage/title_844552_en.html
 
Description eHealth industry engagement conference 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach Regional
Primary Audience Industry/Business
Results and Impact Digital Health Conference (Exeter) on 6/6/18 with around 60 attendees, particularly SMEs in eHealth, health practitioners (incl. from NHS). The conference led to several meetings with two tech entrepreneurs (Rupa Chilvers, Charlie Young) to discuss options for integrating our methods into apps for pregnant women.
Year(s) Of Engagement Activity 2018
URL https://www.swahsn.com/event/digital-health-conference/