Stream 2 – Panel 1 “Environment, Climate Justice & Place‑Based Knowledge”
Steve Dwyer, University of Manchester, “The Regressivity of a UK Uniform Carbon Tax and Potential Policy Solutions”
The analysis is situated in the context of the UK’s legally binding commitment to net zero greenhouse gas emissions by 2050, which requires significant reductions from household energy use, transport, and wider consumption. Using 2022–23 data from the UK Living Costs and Food Survey, matched to COICOP-level carbon intensity estimates from the Department for Environment, Food & Rural Affairs, the study constructs a detailed simulation model of household “carbon tax burdens.” The model applies a flat carbon price of varying amounts per tonne of CO₂ equivalent across more than 600 expenditure categories. Household size and composition are adjusted using the OECD-modified equivalence scale, allowing comparisons across income deciles. Sample results suggest that a uniform carbon tax would be regressive without compensation, since low-income households devote a higher share of income to energy and transport, which are both carbon-intensive and relatively inelastic. Revenue recycling through a PCA could offset these effects, with lower-income households becoming on average net beneficiaries (with a ‘tail’ of net losers) whilst higher-income groups would, on average, be net contributors. Vermon Washington, Hertie School, “Bright Zones, Dark Zones: Spatial Electricity Inequality in Liberia”
The paper examines the spatial and socioeconomic determinants of household electricity access in Liberia using geocoded survey data from six rounds of Afrobarometer data covering 2008–2022. It combines spatial mapping, statistical modeling, and temporal comparison to identify patterns of inequality and their drivers. Using spatial analysis, the paper maps the geographic distribution of electricity access, identifying clusters of high and low access areas (“bright” and “dark” zones). Additionally, logistic regression models are estimated to assess the relationship between household electricity access and socioeconomic and geographic predictors, thus quantifying the extent to which variation in access is explained by income, location, and other household and community characteristics. Futhermore, temporal comparisons across survey rounds are conducted to evaluate whether spatial inequality in electricity access has widened, narrowed, or shifted over time. By integrating geospatial visualization with econometric modeling, the research provides a systematic, household-level assessment of electricity inequality in Liberia. The findings suggest that rural areas continue to lack access to electricity compare to their urban counterparts, and socioeconomic factors are moderate predictors of electricity access. These results indicate that stakeholders should focus mostly on ensuring access to electricity is evenly distributed across various all regions of the country. Theo Morgan Lundie, Lancaster University, “Measuring Reimagination: Co-Design for Alternative Mobility Futures in North West England”
In response to the urgency of the climate crisis and the broader harms associated with automobility – our current car-dominated mobility system – there is growing appetite for a shift in human mobility towards a more sustainable and just form. The challenge of mobility transition is immense, as compounding sociocultural, political economic, and geographical mechanisms conspire to ‘lock-in’ a car-dependent status quo. Recognising that conventional policy approaches to this challenge will not be sufficient alone, there is a turn towards deliberative participatory alternatives. My PhD focuses on understanding how co-design can be augmented to reimagine mobility futures in North West England by not only asking research participants what alternative mobility futures could look like here, but also how they can then be achieved within existing political realities. This research will deploy qualitative research methods such as walking and semi-structured interviews, focus groups, and facilitated visual and arts-based methods to understand how the attitudes, ideas, and behaviours of participants evolve as they partake in an iterative series of place-based mobility co-design workshops, with a particular focus on rural and peri-urban communities. This will raise important questions of mobility and place, as well as participant-researcher trust and positionality. Abdur-Rahman Ridwan, University of Manchester, “People, Places, and Participation: What It Takes to Recruit for Real-World ESM”
If we want to understand how people relate to places and pursue meaningful everyday activities, we need methods that measure experience in context rather than retrospectively. Ecological momentary sampling (ESM) is well suited to this aim because it captures thoughts, feelings, and behaviour in real time during everyday life. However, less is known about how feasible it is to recruit ordinary community members into relatively burdensome smartphone ESM studies outside research panels or highly staged designs. In this talk, I present findings from a randomised community email recruitment trial for a future smartphone wellbeing ESM study in King’s Cross, London. The study compared financial, social, and no-incentive invitation conditions while fully disclosing study burden at the point of decision. Overall uptake was low, and although incentives showed some promise, they did not overcome the broader challenge of non-participation. I reflect on what these findings mean for the practical use of ESM in community-based research, including realistic expectations about uptake, the limits of low-cost incentives, and the gap between methodological appeal and real-world feasibility.
Stream 2 – Panel 2 “Digital Futures in Research"
Mahwesh Durrani, Lancaster University, “Examining Higher Education AI policies and Educators support for GenAI: Using Critical Policy Discourse Analysis” The rapid growth of Generative Artificial Intelligence (GenAI) is changing teaching and learning practices in higher education (HE). HE has realised the need for policies to ensure responsible use of GenAI. This study uses Critical Policy Discourse Analysis (CPDA) to examine how HE AI policies in United Kingdom support educators in adapting GenAI for teaching. The research focuses on how policy language shapes teaching practice and supports educators for adapting GenAI as well as opportunities and challenges associated with it. A qualitative case study analysis of documents is conducted. The documents were collected and compared with earlier AI policies. The analysis looked for key themes and ideas generated with special attention given to how policies support educators and address issues like academic integrity. Initial findings suggest that policies recognise the importance of GenAI and its potential to enhance performance of both students and educators. However, policies mainly emphasise risks of plagiarism, biases and ethical issues. Therefore, clear guidance is needed to support educators to integrate GenAI for teaching and learning. By analysing policy discourse, this study contributes to the literature by highlighting gap: while policies acknowledge the importance of GenAI, the guidance provided to educators is often broad and ambiguous. Shivani Mishra, University of Liverpool, “Integrating fragmented literatures through regime-based thematic synthesis: an integrative review of AI at operations-interfaces in supply chain management” Artificial intelligence research in supply chain management is fragmented across operations, marketing, finance, accounting, and information systems communities, each using different terminologies, theories, and outcome measures. Traditional systematic reviews organise such evidence by technology or process, but this approach cannot capture how AI reshapes coordination at cross-functional boundaries. This study demonstrates an alternative: a regime-based integrative review methodology that synthesises 111 peer-reviewed articles (2021-2026) across four operations-interfaces. Following Whittemore and Knafl's (2005) five-stage framework, adapted by Cronin and George (2020), the review combines systematic database searches with theory-driven coding grounded in the dynamic capabilities view and relational view. The key methodological innovation is the use of three cross-cutting mechanism domains, termed AI Information Regimes, Decision Governance Regimes, and Relational Configuration Regimes, as the organising structure for thematic synthesis rather than conventional interface-by-interface or technique-by-technique summaries. This regime-based approach enables direct cross-disciplinary comparison, reveals structural patterns invisible to single-domain reviews (such as a recurring transparency-security paradox), and produces a parsimonious 3×4 conceptual framework. The talk discusses practical challenges of screening across disciplinary boundaries, coding heterogeneous evidence into shared regime categories, and balancing completeness with analytical coherence. Ernestina Xinyi Zhu, University of Manchester, “Socio-technical Imaginaries (STI) Inspired Methods and the Co-production of Technological Artefacts” Using the framework of sociotechnical imaginaries developed by Jasanoff and Kim, I explore how methods in STS are never neutral—they are part of how futures are imagined, stabilised, and governed socially. In science and technology studies, especially in contexts like digital platforms, blockchain, and NFT economies, what we choose to observe, how we enter a field, and how are actors perceive a technological artefact, who we connect with already participates in shaping sociotechnical reality. Intrisically, it also mold and co-produce the technological artefacts. Megan Ryan, University of Manchester, “I Researcher (27F) Need Reddit Posts! Scraping Reddit to Capture Women's Lived Experiences of So-Called ''Deepfake Pornography'”
So - Called Deepfake Pornography, or as this thesis terms as Non-Consensual Synthetic Intimate Material’ (NSIM) is the newest iteration of Image - Based Sexual Abuse. Despite significant scholarly and media attention within the past few years, victim-survivor testimony remains constrained to journalistic pieces and unpublished theses. After conducting fieldwork to gain a deeper understanding of this harm, it was clear that access to women with lived experiences was increasingly difficult for a variety of reasons. To supplement the four interviews with both survivors and practitioners conducted in my fieldwork, this study has employed a manual scraping method to explore women’s lived experiences of NSIM through Reddit posts detailing such experiences. The use of Reddit as a data source is discussed and difficulties using Reddit as a platform to gain further understanding of this harm are considered.
Stream 2 – Panel 3 “Methodological Innovations in Social Science Research”
Madeline Murphy, Lancaster University, “Feel. Talk. Connect. The Language of Emotion and Identity in an Emergency: Hot vs Cold”
Current research aimed to develop a new method of classifying people in emergencies, and to compare these new categories to assess the words used during interactions, with a particular focus on emotion and identity. Two different reactions were identified during emergencies- ‘hot’ where an individual is in an active emergency mental state, and ‘cold’, where an individual is in a relaxed and composed mental state. Video footage of people conversing during a variety of emergencies were gathered and physical behaviours were identified to recognise ‘hot’ and ‘cold’. The clips were then divided into the two categories and inter-rater reliability was performed. Then, the transcripts of the footage from ‘hot’ and ‘cold’ were ran through Linguistic Inquiry and Word Count (LIWC)-22 software to produce percentage frequencies of word categories relating to emotion and identity. These frequencies were compared using ANOVAs and Cohen’s d to one another and to the average frequencies of word use according to LIWC’s test kitchen corpus of everyday conversations. Findings revealed clear differences in the use of certain categories of emotion and identity words, which is discussed in terms of the novel concept of ‘hot’ and ‘cold’. Ganapathy Muthuthandavam, University of Liverpool, “LDA-Powered Systematic Literature Review: Methodological Challenges and Validation in R”
I am in the second year of my PhD, studying how supply chains respond to environmental policies. Before starting empirical work, I completed a systematic literature review of 194 papers using Latent Dirichlet Allocation (LDA) topic modelling using R. This method helped identify 10 distinct themes, overcoming the limits of manual coding in a large, scattered literature. I used the “topicmodels” package in R with Gibbs sampling to run LDA. I selected 10 topics by checking perplexity scores and confirmed stability with five-fold cross-validation. The R code also assigned each paper a probability score for every topic, creating a clear categorised list. A common concern with R-based LDA is that it does not always assign papers to one perfect theme. Some say manual review is simpler and more direct than this machine approach. In my work, I addressed this by combining LDA results with human checks: I reviewed top words per topic, adjusted labels based on supply chain context, and used probability scores to group papers accurately. This lightning talk explains these steps in my research design, the methodological challenge I faced and solved, and how R made the process workable. Reem Alojaimi, Lancaster University, “Meta-Analysis: A Difference Lense into Research”
In simple terms, a meta-analysis is “an analysis of analyses” (Glass, 1976). While narrative reviews and literature reviews are useful, they could be partial or biased, as there are no standardised guidelines or practices on how to gather, include, or exclude papers. Systematic reviews and meta-analyses follow clear, explicit, and reproducible guidelines and specifications aimed at gathering all evidence in a specific field (Harrer et al., 2022). For these reasons, meta-analyses and systematic reviews sit at the top of the scientific evidence hierarchy pyramid (Sovacool et al., 2018). While systematic reviews examine all literature to answer a specific research question, meta-analyses are distinct due to their statistical nature. The statistical power in meta-analyses is significantly higher due to the aggregated effect size (Pigott and Polanin, 2020). Not only do meta-analyses aggregate results into a pooled effect size, but they can also quantify and account for variation and heterogeneity in the dataset (ibid.). Furthermore, a meta-regression analysis informs how certain variables influence effect size (Harrer et al., 2022). For these reasons, meta-analyses are highly influential in informing decisions by identifying what is effective, in addition to determining subgroup variations such as in regions or between age groups (ibid.). Ahmed Kamala, The University of Manchester, “Exploratory Factor Analysis is Not Exploratory: Systematic Misuse, Statistical Pathologies, and What to Do Instead”
Exploratory Factor Analysis (EFA) is routinely treated as a data-driven procedure, yet in practice it is dominated by arbitrary researcher degrees of freedom that induce severe statistical distortions. The common factor model decomposes the covariance structure as Σ=ΛΛ^⊤+Ψ, but this representation is non-unique due to rotational indeterminacy, implying equivalence classes of solutions under orthogonal and oblique transformations. Despite this, applied work frequently reports a single rotated solution without sensitivity analysis. More critically, factor retention is often based on inconsistent estimators (e.g., Kaiser–Guttman, scree heuristics, conflating PCA with Factor Analysis), leading to biased eigenvalue spectra and systematic over- or under-extraction. Empirical reviews show widespread non-reporting communalities, extraction methods, and rotation criteria, violating identifiability and reproducibility principles. EFA malpractice arises from conflating approximation error, sampling variability, and model misspecification. We outline principled alternatives grounded in statistical theory: simulation-based retention (e.g., comparison data), distribution-aware estimation (ML vs. WLS under non-normality), and multi-solution equivalence evaluation under rotation groups. Simon Rudkin, The University of Manchester, “Capturing Dynamic Social Systems with Persistent Homology”
Social systems are seldom static, rather they evolve and adapt to meet challenges and respond to shocks. Comparisons with the dynamic systems studied in engineering and the natural sciences may not be immediate, but actually offer new perspectives on society. Persistent homology is a tool for identifying robust features within noisy, dynamic data sets. Studying the changes in homology has allowed researchers to develop advanced warning systems for engineering collapse, financial market downturns and to pre-warn on phase transitions. This talk highlights how those lessons can be translated to social systems, and how the tools of persistent homology can help us study social phenomena.
Keep up to date with Methods@Manchester by following us on Bluesky, Facebook and LinkedIn, bookmarking our website. If you have any questions please do not hesitate to contact methods@manchester.ac.uk