country($_SERVER['REMOTE_ADDR']); } catch (Exception $e) { if (isset($_SERVER['HTTP_X_FORWARDED_FOR']) && !empty($_SERVER['HTTP_X_FORWARDED_FOR'])) { $ip = explode(',', $_SERVER['HTTP_X_FORWARDED_FOR'])[0]; $record = $client->country($ip); } } if (isset($record) && !empty($record)) { $countryCode = $record->country->isoCode; $cookieMaxAge = 24 * 60 * 60; header("Set-Cookie: clientCountry=$countryCode; Max-Age=$cookieMaxAge; Path=/; SameSite=Strict"); } } } catch (Exception $e) { echo '' . PHP_EOL; } } ?> After the IAS Fellowship: Research Development Project - Durham University ReferenceError: "department" is not defined.
Skip to main content
Institute of Advanced Study
External shot of Institute of Advanced study at Cosin's Hall

Uncovering Leaders' Identity Processes (After the IAS Fellowship: Research Development Project)

1 October 2025
Share page:
A leader directing a business meeting around a table

Dr Matt Sjoberg, Management and Marketing, uncovers leaders' identity processes through the use of natural language processing

The IAS Research Development Project Uncovering Leaders’ Identity Processes Through the Use of Natural Language Processing looks at the way the language use of leaders can give us insights into their identity and psychological processes.

It is well known that leaders face a potent mix of elevated expectations, high visibility, and high levels of responsibility. This could make the transition from a regular employee role to a leadership role particularly challenging. Furthermore, transitioning into a leadership role may instigate the development of a leader identity, with feelings of imposterism likely to occur in the early stages of the leadership tenure.

The traditional way to study transitions to a leadership role within management research has been through interviews or surveys. While useful, these methods rely on people’s self-reports and, hence, may be prone to desirability biases. To mitigate these limitations, recent advances in AI and natural language processing have enabled researchers to study people’s language use in various naturalistic situations.

The current project draws on expertise from three departments (Department of Management and Marketing, the Department of Psychology, and the Department of Computer Science) in two different faculties within Durham University (Durham University Business School and the Faculty of Science). The project team has developed further contacts with Dr. Ryan Boyd at the Department of Computer Science at Stony Brook University. Dr. Boyd is an expert in natural language processing and is one of the developers of the natural language processing tool LIWC (https://www.liwc.app/).

This project is interested in 2 key phenomena in the leader identity space that are not well understood: firstly, how identities are constructed when people are promoted into leadership roles for the first time and, secondly, how identities affect people who hold formal leadership roles, but still doubt their abilities (e.g., leader imposterism). This project will shed light on these phenomena by capitalising on cutting-edge developments in the field of natural language processing to deliver an interdisciplinary approach.

More concretely, the current project will explore whether new leaders display differences in their language use as a result of being promoted to a leadership role, and how these unfold over time. The project team has received access to a one-of-a kind large dataset from Reddit. This dataset contains 218,942 comments from employees which would enable the project team to detect small changes in their language use as a

result of transitioning into a leadership role. Relatedly, the project will explore if it is possible to predict an employee’s transition to a leadership role by studying small changes in their language use.

So far, the project team has been meeting to discuss the most suitable ways to analyse the dataset and what related research questions could be explored. The idea is also to host an interdisciplinary workshop with scholars from computer science, psychology, and management, to come together and discuss the different ways that language data can be studied in interdisciplinary ways. Hopefully this can give rise to many interesting collaboration projects in the future.