Search results for: “ai in education”

Beginning from the Inside Out: The Value of Pilot Studies in Decolonial Research

Kate Matzopoulos, PhD researcher in Education at the University of Bath

This presentation explores the value of conducting a pilot study prior to larger-scale research, particularly when working with Indigenous and decolonial methodologies that are still in experimental phases and not widely documented in education research. My context, with the Ju/’hoansi San community in Namibia, is not widely written about outside of anthropology, which meant that I had to draw from other Indigenous scholarship to piece together meaningful approaches.

My pilot study, carried out during my master’s research, was foundational in forming authentic relationships with the community. Relationality became the hinge on which the work rests: research emerges in and through relationships, shifting with daily life rather than fitting into neat methodological boundaries. The pilot also revealed how to engage respectfully across difference and avoid reproducing colonial knowledge systems that obscure reality.

Ultimately, the study affirmed the importance of beginning research from the “inside out,” allowing Indigenous people to act as co-researchers working from their own onto-epistemologies, rather than being framed by external categories or outsider lenses. This informs my ongoing fieldwork in my PhD studies.

 

Conducting a pilot to trial photo diary instructions and assess the data generated

Louise Toller, SWDTP alumni in Sociology at the University of Exeter

My PhD research investigated the experiences of chronically ill young adults using a combination of verbal and visual methods, specifically participant-generated photographs and photo elicitation interviews. In this presentation, I describe how I used a pilot study to develop my visual method, essential due to both my lack of prior experience and the lack of detailed information in published studies. I conducted a pilot with two main objectives: to trial the draft photo diary instructions I would give to participants, and to assess the sort of data the photographs generated. I reflect on what I learned from the pilot, and the modifications I then made to the photo-based phases of the main study, highlighting lessons that I would not have gained without carrying out a pilot.

This session is part of the SWDTP Data Analysis Webinar Series. Visit the following link for further information and registration: https://www.tickettailor.com/events/swdtp/1956811

Higher Education in Nigerian public discourse

Yusuf Olaniyan, SWDTP-funded PhD researcher in Education at the University of Bath

This presentation will discuss the methodological approach and data analysis techniques used in my study of how higher education (HE) is constructed within Nigerian public discourse. The research combines media analysis with a discourse-historical approach (DHA) to explicate how HE is framed in Nigerian national news medias. It focuses specifically on the post-COVID period, and the introduction of the National Universities Commission’s (NUC) Core Curriculum and Minimum Academic Standards (CCMAS) polices. The study used 331 media articles from three major Nigerian news media (Punch, Daily Trust, and Daily Post), published between 2021 and 2022, which focused on the intersections of vocational education, entrepreneurship, skills training, and apprenticeship-like pathways in the context of HE reforms. Through a two-phase analytical process, the first phase involved a quantitative scan to map the frequency and thematic emphasis of HE discussions in the media, highlighting the economic framing of HE around employability and market-driven outcomes. The second phase employed DHA to visibilised the ideological functions of these discourses. It also investigates how HE policies were legitimized, how social actors were positioned, and how historical continuities were reproduced in the media. Through this methodology, the presentation will illustrate how media discourses serve both framing and gatekeeping functions to determining which perspectives on HE are amplified or marginalised. It will highlight the challenges of conducting discourse analysis on publicly available media data, as well as the methodological and theoretical benefits of combining quantitative content analysis with a qualitative DHA framework. This session will be useful for doctoral researchers like me and early-career researchers interested in media analysis, discourse studies, and the intersection of policy, education, and public opinion in national contexts.

How do social media influencers legitimise veganism on YouTube?

Adrianna Jerzierska, PhD researcher in Business Studies at the University of Bristol

In this presentation, I will outline the methodological framework used in my PhD project, which aims to understand how social media influencers have legitimised veganism over time, by analysing approximately 8,000 YouTube videos posted between 2014 and 2024.

During this workshop, I will discuss the computational theory construction and the qualitative techniques used to study the discursive legitimation strategies by which influencers mainstream veganism. The first phase of the analysis involves generating the Dynamic BERT topic model (DTM). BERT is a large language model (LLM) that performs language tasks, such as distilling clusters of documents with similar semantic topics. By incorporating temporal analysis of topics, researchers can effectively track lexical changes across topics, leading to a deeper understanding of discursive legitimacy over time. The second stage involves the qualitative interpretation of the machine output by adopting a qualitative content analysis of the videos to contextualise the discursive strategies used by online actors.

Given that BERT performs best with short texts compared to long and noisy YouTube video transcripts, this was a significant challenge to overcome. In this webinar, I will discuss the decisions made to work within these limitations, as well as provide a brief overview of the analytical framework, from querying and downloading videos using the YouTube API to discussing AI tools for converting large audio files into transcripts.

This session is part of the SWDTP Data Analysis Webinar Series. Visit the following link for further information and registration: https://www.tickettailor.com/events/swdtp/1956811

The Messy Realities of Anticolonial Research: Reimagining Language Practices in Secondary Education in Nigeria

Mercy Martins, PhD researcher in Education at the University of Bath

My research explores how language practices in Nigerian secondary schools can be reimagined using anticolonial concepts of agency and resistance. The research design was indigenous, weaving qualitative, creative, and participatory methods that were flexible and aligned with the daily experiences of my collaborators. Structured in two phases, the first focused on the voices of marginalised students, while the second included teachers, parents, school management, and leadership.

After securing ethics approval, I conducted a small online pilot with local collaborators which informed my choice of a school with linguistic, ethnic, and religious parity. I prioritised depth in one all-male secondary school rather than breadth across two schools due to time constraints. Almost five months of fieldwork produced extensive data. To preserve Nigerian Pidgin and avoid rough translations, I transcribed and cleaned the data manually. Journalling and a participatory analysis session, where students directly engaged with the research questions, guided the analysis.

While I set aside some methods (such as Theatre of the Oppressed and parent interviews) for future work, I sought to include multiple perspectives, centre participant voices, and clearly distinguish between their ideas and mine. Ethical tensions included deciding what to share, avoiding reductive categorisation, and ensuring authenticity.

 

Storytelling, Embodiment, and the Ethics of Analysing Bedouin Ethnography

Wesam Wekhyan, PhD researcher in Social Policy at the University of Bath

My PhD research explores Bedouin identity, knowledge, and community praxis in central Jordan through a decolonial, feminist ethnographic approach. The dataset is multi-modal: focus groups, one-to-one interviews, oral histories, WhatsApp notes, photos, videos, fieldnotes, memos, and observations of ceremonies and everyday life (weddings, funerals, dinners, walkabouts). Much of this material was produced in a Bedouin dialect and a culture where history is preserved orally rather than textually, requiring translation across dialect, Arabic, and English, as well as from embodied, performative expression into written form. Virtual data (voice notes, chats, images) extended this storytelling into digital space.

Writing was not separate from analysis, but a continuation of storytelling: through memos, vignettes, and embodied descriptions I traced how narratives lived, moved, and transformed across settings and generation. Analysing such material meant making decisions about segmentation and representation without erasing relational meaning. Ethical and personal challenges were constant: anonymity in small communities, safeguarding oral/visual data sovereignty, and navigating my insider–outsider position and the embodied emotional labour of researching from within.

This session is part of the SWDTP Data Analysis Webinar Series. Visit the following link for further information and registration: https://www.tickettailor.com/events/swdtp/1956811

“Entanglement makes all the categories of humanist qualitative research problematic.” (Lather and St. Pierre, 2013)

How does our analysis shift when we stop asking ‘What does this data show?’ and instead ask ‘Who and what is this data entangled with, and how can we stay accountable to those relations?’

This webinar invites pondering of the joys and challenges that emerge with data analysis when social researchers embrace their entangled inseparability from the world being researched. We explore what becomes possible when we analyse data as an experiment in relations rather than attempting to represent a universal reality. By considering the posthuman ontological turn and its implications for social sciences research, we, a group of posthumanist, feminist materialist, post-qualitative educational researchers, share some of the processes and techniques that we employed as we grappled with the process of meaning making in our research projects. Through matter, waters, bodies, silences, we consider what our more-than-human data might say, if it could speak, and how such voices might re-shape the ways we engage with the worlds around us, when we listen more expansively.

By showcasing examples of ‘how did we do it’, this webinar is suitable for researchers at any career stage, including ECRs, who are either working with posthuman data analysis, or are curious about practices that challenge conventional ways of doing social sciences research. In addition, the webinar will provide an interactive opportunity for participants to engage in posthuman data analysis through practical examples. Anonymised data examples will be provided, or participants are welcome to bring their own.

This session is part of the SWDTP Data Analysis Webinar Series. Visit the following link for further information and registration: https://www.tickettailor.com/events/swdtp/1956811

Understanding home deathcare through narrative.

Sam Hooker, SWDTP-funded PhD researcher in Social Policy at the University of Bath

My PhD project used a narrative methodology to capture people’s experiences caring for the dead at home. Two interviews were conducted with participants. The first consisted of asking them to tell me their experience, while the second was used to ask questions to fill in gaps in the narratives. Then, in the participant’s own words, a chronological narrative of the event (core story) was formed, which reads like a short story of their experience. This process was an emotionally taxing experience for me as a researcher, due to the depth of immersion in the data. The core stories are included in the thesis, allowing the data to remain situated, capturing the complexity of a very personal and challenging experience for participants. While additional analysis is included, the core stories allow the reader to draw conclusions from the data itself. They also provide a valuable resource for people who wish to take care of their own dead to learn from.

 

Putting a light in the window: using creative writing to focus analysis in relationally engaged research

Luci Gorell Barnes, PhD researcher in Education at UWE Bristol

My PhD investigates how relationally engaged arts-based research methods can support minoritised children to express, reflect on, and amplify their lived experiences and perceptions. I understand meaning-making as happening through the process of engagement, and gathered data to consider the ‘double hermeneutic’ (Smith and Eatough, 2021) of not only what the children expressed, but how they did this through their conversations, interactions, and demeanours.

This study focuses on the children’s unique experiences and I wanted to examine what I had learned from each child, foregrounding how they had ‘got under my skin’. Having familiarised myself with the data I then drew on less conscious or orthodox ways of knowing (Hammond and Fuller, 2024) to write what I think of as ‘fictional portraits’ of my encounters with individual children. These were informed by the issues that child explored, the metaphors and images they used, and how they had interacted with me and each other. Using this process helped me get close to each child’s ‘internal narrative’ Mannay (2010, p.10) and I came to understand each ‘portrait’ as being like a light in the window guiding me through the dark woods of data, focusing my analysis and supporting me to write interpretive accounts that kept the relational nature of the study at their heart. McNiff (2019) suggests that fact and fiction can work together to deepen our understandings, and by engaging with my data in this highly subjective and ‘interruptive’ way (Clark, 2024, p. 3) I brought a level of relational accuracy to my analysis that I might not have otherwise found.

This session is part of the SWDTP Data Analysis Webinar Series. Visit the following link for further information and registration: https://www.tickettailor.com/events/swdtp/1956811