Masters Theses

Permanent link for this collectionhttps://hdl.handle.net/10292/5

The Masters Theses collection contains digital copies of AUT University masters theses deposited with the Library since 2002 and made available open access. From 2007 onwards, all theses for masters degrees awarded are required to be deposited in Tuwhera Open Theses & Dissertations unless subject to an embargo.

For theses submitted prior to 2007, open access was not mandatory, so only those theses for which the author has given consent are available in Tuwhera Open Theses & Dissertations. Where consent for open access has not been provided, the thesis is usually recorded in the AUT Library catalogue where the full text, if available, may be accessed with an AUT password. Other people should request an Interlibrary Loan through their library.

Browse

Now showing 1 - 20 of 3433

Exploring Financial Shame
(Auckland University of Technology, 2025) Legaspi, Evan
Households are increasingly indebted, with the financial pressure being associated with stress and depression, worsening physical and mental health, suicide, and divorce. Compounding this is a taboo regarding talking about money, with many individuals concealing their financial problems, avoiding addressing their financial difficulties, and withdrawing and isolating themselves. With all the spiralling negative effects of financial stress and debt on an individual’s well-being, it is critical that we understand financial shame. This research aims to explore the factors that amplify and buffer financial shame. Across Aotearoa New Zealand, 704 Kiwis were asked to answer questions on psychosocial processes, lifetime factors, contextual conditions, and personal resources, identified through prior research on general shame. These factors included role-based identity proxied with household financial responsibilities, group membership based on ethnic and religious affiliations, financial literacy and self-efficacy, and the importance of social comparisons. This research probed the individual’s past financial journey, including childhood socio-economic upbringing and family financial socialization, financial hardship and material deprivation in later life, and explored their current financial circumstances. The univariate analyses showed that most of the investigated factors significantly influenced financial shame. Lack of control over household finances, low self-efficacy, collectivist orientations, and weak ethnic belonging were associated with higher shame, whilst financial literacy and confidence acted as strong buffers. Social comparison emerged as a central amplifier, with financial hardship – both in childhood and adulthood – as a consistent predictor. Subjective financial well-being protected more strongly against shame than objective measures, highlighting the psychological nature of financial distress. Regression results confirmed that self-efficacy (negatively) and social comparison (positively) had the strongest correlations with financial shame, while current financial circumstances is likely a critical trigger towards feeling financial shame, with past hardship likely serving as a key source of shame-proneness. Cluster analysis further reinforced these dynamics, revealing six distinct financial shame personas that differed based on respondents’ agency regarding financial decisions, confidence, cultural orientation, social comparisons, and financial stability. The shame components were also used to establish unique groupings that can be used to understand how background, beliefs and resources shape the emotional experiences of money. Across all models, age consistently moderated the experience of shame – both buffering and amplifying its effects, depending on life stage – demonstrating that financial shame is multi-dimensional, context-dependent, and deeply rooted in both one’s financial trajectory and sense of identity. The research findings reinforce that financial shame is not simply about money, but reflects a deeper judgement of the self, with the meaning people attach to money, such as competence, responsibility, independence, and success, as the core driver. This research presents some opportunities to build financial shame resilience, and contributes to what is currently a very limited literature on a critical issue. Opening up these discussions on financial shame enables developing effective interventions that remove these money and debt taboos, and help those who need help the most, with urgency and purpose.
Best Practices in Intercultural Communication That Support Diversity, Equity and Inclusion in Three New Zealand Media Workplaces
(Auckland University of Technology, 2025) Devgan, Samrita M
This study investigates the application of intercultural communication knowledge in diversity, equity, and inclusion (DEI) practices within three prominent New Zealand media organisations – TVNZ, NZME, and Stuff. Using the Aotearoa Inclusivity Matrix (AIM) framework, the research examines how these organisations use the seven DEI components in the workplace: leadership, diversity infrastructure, diverse recruitment, inclusive career development, bi-culturalism, inclusive collaboration, and social impact. A qualitative thematic analysis identifying the DEI components was conducted on 348 publicly available data items, collected across paid, earned, shared, and owned (PESO) media. Each DEI component was critically assessed using an aligned intercultural communication theory to determine the extent of theory-informed practice. The findings revealed the presence of all seven DEI components in the organisations in varying measures of priority, shaped in part by their ownership structures. The most prioritised practices included social impact and inclusive career development. Inclusive collaboration and diverse recruitment practices showed an overall low presence across the organisations, particularly during periods of restructuring. The study highlights that while intercultural communication theory is implicitly present in many DEI initiatives, its explicit and strategic application is inconsistent. The findings advocate for a more deliberate integration of intercultural communication knowledge into DEI policies to ensure sustainable, culturally responsive practices. These insights contribute to the growing discourse on the value of DEI in workplaces by proposing a strong theoretical framework of intercultural communication knowledge to support the effectiveness of DEI in organisational practices, thereby offering guidance for media and other industries in multicultural contexts.
MLGA - A Modality-Level Graph Attention Architecture for Multimodal Depression Detection
(Auckland University of Technology, 2025) Malika, Malika
Depression is one of the most pressing global health challenges, affecting millions of individuals and placing significant strain on healthcare systems. Early and accurate detection is critical for timely intervention and improving patient outcomes. Traditional diagnostic methods, which rely heavily on clinical interviews and self-reports, are often resource intensive, subjective, and limited in scalability. To address these limitations, this study presents an architectural investigation of Modality-Level Graph Attention (MLGA), a deep learning framework for multimodal fusion in depression detection. The proposed architecture integrates textual embeddings from ClinicalBERT, visual representations from VGG-PCA, and facial behavioral descriptors from OpenFace. These modalities are fused through a modality-level Graph Attention Network (GAT) that explicitly models inter-modality relationships, while a temporal module captures dynamic behavioral patterns over time. To enhance robustness, the framework incorporates Gaussian noise injection, L2 normalization, and modality dropout, thereby encouraging resilience to noise and missing inputs. A comprehensive evaluation was conducted on three benchmark datasets: E-DAIC-WOZ, EATD-Corpus, and D-Vlog. Across these datasets, MLGA achieved competitive performance and, on the E-DAIC benchmark in particular, surpassed several unimodal and late-fusion baselines in terms of precision, recall, F1-score, and ROC-AUC, demonstrating the effectiveness of graph-based multimodal integration under the studied conditions. The results highlight the importance of modeling both intra-modality features and cross-modality dependencies within a unified fusion architecture. Rather than proposing a fully deployable clinical tool, this study advances the field of affective computing by systematically analysing a modality-level graph attention design that is computationally moderate and interpretable, and by quantifying its behavior across heterogeneous datasets. Future directions include expanding modality coverage, applying domain adaptation for cross-cultural and cross setting generalization, and enhancing interpretability using advanced explainable AI techniques.
Improving Developmental Motor Outcomes Through Intensive Early Intervention at 3-months Corrected Age in Preterm Infants: A Pilot Feasibility Study
(Auckland University of Technology, 2025) Pearce, Louise
Background Preterm birth, particularly in moderate to late preterm (MLP) (32+0 to 36+6 weeks gestation), is associated with an increased risk of long-term neurodevelopmental difficulties compared to term infants. These impairments are often under- recognised and consequently low-risk MLP infants, typically receive no developmental follow-up on discharge home or referral to early intervention (EI) services. Although ‘early’ EI is more effective to enhance neuroplasticity in the maturing infant brain, access for MLP is often restricted. Group-delivered EI may offer a solution and may improve accessibility and resource efficiency for service providers. When provided as early as 12-weeks corrected age and at a high intensity, group-based EI has the potential to improve developmental outcomes. Currently there is limited research investigating the feasibility and acceptability of utilising a group-delivery model EI for very young infants born MLP. Objective This study explored the feasibility and acceptability of a 12-week, high-intensity, group-based, early implemented EI programme (NEOgrads Playgroup) for healthy, low-risk MLP infants and their caregivers in Auckland. Methods This non-randomised pilot feasibility study, NEOgrads Playgroup, involved nine MLP infants recruited at 11 weeks corrected age, all with normal general movements and deemed ‘low risk’ of abnormal neurological sequalae. The intervention was delivered weekly via a group class (community/clinic-based setting), over a 12-week period by a neurodevelopmental therapist. In addition to the developmentally supportive play and educational group classes, caregivers implemented a personalised home exercise programme over the duration of the study. Feasibility was assessed via caregiver’s programme compliance (group and home programme), group-based intervention protocol fidelity audit, and caregiver’s acceptability of the programme following intervention completion. Initial efficacy was explored pre and post intervention using the Peabody Developmental Motor Scale – Second edition (PDMS-2) to assess changes in gross motor developmental trajectories. Results The median group programme caregiver compliance rate was 90% (80-91.6%), with mean home exercise programme compliance reported at 4.9/7 (SD =1.18) sessions per week. Protocol fidelity was met for the group programme delivery for core NEOgrads Playgroup principles. The caregiver-rated satisfaction survey showed a strong belief in programme benefits to their infant, themselves and whānau, with a strong support of programme endorsement to other caregivers. Three subthemes emerged with the programme improving: caregiver-infant relationship, caregiver self-efficacy, and infant’s self-efficacy. Initial efficacy findings indicated six infants demonstrated meaningful improvements in gross motor development from pre to post intervention, as measured by the Peabody Developmental Motor Scale (PDMS-2). These changes exceeded the standard error of measurement (SEM = 5.88), suggesting trajectories above normal developmental progression. Whilst two infants showed a reduction in their PDMS-2 scores (a change of 8 points), and one infant demonstrated stable developmental progression (a change of 2 points). Conclusion This initial pilot feasibility study demonstrated that the NEOgrads Playgroup, a group-based early intervention programme, is both feasible and acceptable for caregivers of low-risk MLP infants. While initial efficacy testing suggested positive trends in gross motor development following the intervention, larger population-based studies with longer-term follow-up are needed to assess the sustained impact of the NEOgrads Playgroup on MLP developmental outcomes.
The Effectiveness of Sin Taxes on Public Health and Revenue Generation in Tuvalu: Insights from New Zealand
(Auckland University of Technology, 2025) Lauti, Moitila Brian
Tuvalu faces a dual challenge of high non-communicable disease (NCD) burdens and constrained fiscal capacity. This thesis investigates whether excise taxes on sugar-sweetened beverages (SSBs), alcohol, and tobacco can improve health outcomes and provide a stable revenue base. Using New Zealand’s Pacifica population as a proxy, Household Economic Survey (2006–2019) records were linked to pharmaceutical dispensing, hospitalisation, and mortality data within the Integrated Data Infrastructure. Logistic regression models were estimated to examine whether household expenditure shares on sin goods predict chronic disease medication uptake and mortality. Households in the highest SSB expenditure tier had 36 % higher odds of diabetes-management prescription uptake (OR = 1.36, p < 0.01). Cardiovascular medication outcomes showed modest but statistically significant increases in odds, including diuretics (OR ≈ 1.16, p < 0.05) and beta-blockers (OR ≈ 1.22, p < 0.01). Mortality associations were mixed: SSB-only expenditure was associated with lower odds of mortality among those aged 60+ (OR ≈ 0.75, p < 0.01), whereas broader sin-goods measures, particularly those including tobacco, were linked to higher mortality odds (OR ≈ 1.56, p < 0.01). Within the Pacifica subsample, effect directions were similar but rarely statistically significant, reflecting smaller sample sizes and contextual constraints. These findings suggest that excise taxes can influence health outcomes, particularly for diabetes management, while mortality effects vary by product scope and population group. Expenditure shares were relatively consistent across income groups, implying a stable fiscal base. However, this study does not estimate price elasticities and therefore does not quantify revenue effects. Because excises are financially regressive, earmarking revenues for health promotion and supporting affordable substitutes are recommended to mitigate distributional concerns. Overall, well-designed excise taxes show conditional promise for improving health outcomes in Tuvalu when paired with broad coverage, adequate rates, and ongoing evaluation.
How Does Te Reo Māori Enhance the Wellbeing of Staff at AUT?
(Auckland University of Technology, 2025) Dallas, Trent
This study examines the impact of Te Reo Māori on the holistic wellbeing of Māori staff at Auckland University of Technology (AUT), using a Kaupapa Māori Research methodology approach and Te Whare Tapa Whā health model as guiding frameworks. Through semi-structured interviews with six Māori staff members, the research investigated how engagement with Te Reo Māori enhances cultural identity, emotional resilience, social belonging, and professional empowerment in the workplace. Findings demonstrate that Te Reo Māori extends beyond a linguistic function, acting as a powerful mechanism for cultural reclamation and holistic wellbeing. Participants described enhanced self-confidence, enriched workplace relationships, and a reinforced sense of cultural identity through language engagement. However, whakamā (feelings of shame or embarrassment) and intergenerational language loss emerged as challenges to engagement. The study highlights the critical role of workplace Te Reo Māori initiatives in enhancing staff wellbeing and affirming cultural identity. This research contributes to the broader discourse on language revitalisation and Indigenous wellbeing, emphasising the necessity for continued investment in Te Reo Māori to enhance staff wellbeing. This study advocates for expanded workplace language initiatives and deeper investigation into professional environments to further examine the intricate connections between language, identity, and wellbeing.
Don't Tell My Mother
(Auckland University of Technology, 2021) Agnew, Leonie
The Impossible Story of Hannah Kent is a young adult manuscript focused on 12 – 15-year-old readers. The story has magical realist elements and is set in small town New Zealand. The female protagonist’s journey centers around coming to terms with her adoption and related childhood traumas. For this reason, I chose to use first person, present tense. This narrative style allows the reader clear insights into the protagonist’s thought processes and emotional state, potentially deepening the connection between the reader and a flawed heroine. The accompanying exegesis Don’t Tell My Mother examines two aspects of the writing process. Firstly, how researching the psychological impact of adoption influenced the development of Hannah’s character. Secondly, the thematic importance of mother archetypes and their relationship with fairy tales, which also formed a psychological approach to my creative thesis.
Knocking out ABCC10 in Human Breast MCF7 Cancer Cells Using the CRISPR-Cas9 System
(Auckland University of Technology, 2025) Agyapong, Kofi Effah
Breast cancer continues to represent one of the most common causes of cancer-related morbidity and mortality globally and in New Zealand. Chemotherapy as a routine therapy remains an effective strategy not only in early-stage breast cancer patients but in advanced breast cancer patients when resistance to endocrine therapy occurs, such as oestrogen receptor-positive (ER⁺) disease. Taxane-based agents (e.g., docetaxel) have been extensively applied because they modulate microtubule dynamics and suppress the growth of tumour cells. Yet, the treatment response of docetaxel is often impeded by the development of chemoresistance. Multidrug resistance (MDR) represents a significant obstacle to the success of chemotherapy, which is largely associated with ATP-binding cassette (ABC) transporters that actively remove cytotoxic drugs from cancer cells, thereby inhibiting intracellular drug accumulation. Of these transporters, ABCC10 (also known as MRP7) has been associated with resistance to taxanes such as docetaxel, but its role in chemoresistance in ER⁺ breast cancer remains unclear. Here, we hypothesise that gene knockout of ABCC10 in ER⁺ breast cancer cell line MCF-7 would improve cellular sensitivity to docetaxel. In order to evaluate this, liposomal transfection of CRISPR-Cas9 ribonucleoprotein complexes targeting exon 3 of ABCC10 in MCF-7 cells was used. Polymerase chain reaction and T7 endonuclease I assay indicated successful gene disruption of the gene with ∼36% genomic cleavage efficiency. To limit clonal heterogeneity, the following single knockout clone was chosen (D4-F8) and functionally characterized. Loss of ABCC10 led to increased sensitivity to docetaxel, with a 35-fold reduction in IC50 values compared to wild-type MCF7 cells (p < 0.05). In addition, ABCC10 knockout cells suffered significantly from impaired clonogenic survival and plating efficiency, while basal metabolic activity was still intact, demonstrating impaired long-term proliferative capacity. These observations highlight that ABCC10 is a functional regulator of docetaxel resistance in ER⁺ breast cancer cells. This study is the first CRISPR-Cas9-mediated knockout of ABCC10 in MCF-7 cells and provides mechanistic evidence that ABCC10 is a potential therapeutic target against taxane resistance. This work underlines the utility of precision genome-editing approaches in defining chemotherapy resistance mechanisms, thus providing an understanding of effective cytotoxic resistance mechanisms and informing treatment strategy in ER⁺ breast cancer.
Automated PAM Fluorometry as a Tool to Evaluate the Photosynthetic Response of Undaria pinnatifida to Experimental Warming
(Auckland University of Technology, 2025) Afening, Mary
Ocean heatwaves are increasing in frequency and intensity, posing threats to coastal ecosystems. One potentially affected group is seaweeds, whose photosynthetic performance may be disrupted by elevated temperatures. Pulse-amplitude modulation (PAM) fluorometry is a non-invasive technique commonly used to assess species’ photosynthetic performance. However, its suitability for unattended and automated in situ monitoring of seaweed photosynthesis under heatwave conditions remains untested. To investigate this, three laboratory experiments were conducted using the invasive kelp species Undaria pinnatifida. In each experiment, one blade from each of three specimens (Experiment 1) or six specimens (Experiments 2 and 3) of U. pinnatifida, acclimated in the laboratory under constant temperature and a diel light cycle, was mounted in front of a PAM fluorometer. The blades were held in fixed positions while the fluorometers automatically conducted saturation pulse (SP) analyses at 30-minute intervals—initially during an acclimation period, and subsequently throughout a warming phase. The measuring lights of the PAM fluorometer remained off between analyses, although the blades remained in place for at least 10 days. SP analyses conducted during the first five hours after light onset each day were used to plot the relative electron transport rate (rETR) and photosynthetically active radiation(PAR) curves, with the slope (alpha) serving as a proxy for the efficiency of U. pinnatifida’s photosystem II (PSII). Inspection of the multi-day time series of alpha revealed a significant decline in PSII efficiency under constant temperature, suggesting that the experimental setup was unable to maintain stable PSII performance. Therefore, the second hypothesis remains inconclusive regarding the response of U. pinnatifida’s PSII efficiency to warming. Nevertheless, Experiment 1 indicated the possibility of a positive response to moderate warming from 14 to 19 °C, and a negative response to further warming to 23 °C. This was also evident in Experiments 2 and 3, despite an overall decline in PSII performance including blade discoloration. Our experiments highlight a critical measurement artefact that must be addressed before attempting to monitor PSII performance in the field. Seaweed blades should not remain fixed in position during unattended automated fluorometry. Instead, they should be mounted on the fluorometer only for the duration of an SP analysis. Although more labour-intensive, this approach enables adequate replication when the number of available PAM fluorometers is limited, and reduces confounding effects associated with prolonged blade immobilisation.
The Experiences of Māori Occupational Therapists: Navigating Cultural Safety and Cultural Load in the Mental Health Sector of Aotearoa
(Auckland University of Technology, 2025) Brown Tawhiri, Georgia
There has been a shift towards cultural competence and cultural safety in Aotearoa in the last 30 years, and the Māori health workforce has played a fundamental role in the development and implementation of these concepts (Ramsden, 1993; Hunter & Cook, 2020; Tofi et al., 2023). The knowledge, skillset and experience of cultural values, practices and the lived reality of being Māori in Aotearoa is an attribute that the Māori health workforce possess (Wilson et al., 2022; Te Aka Whaiora, 2024). This skillset being highly sought after by services and organisations has created a demand and pressure on the low number of Māori working in the mental health sector, in particular Māori occupational therapists. Māori occupational therapists are shouldering cultural load and culturally unsafe situations in attempts to ensure tangata whaiora Māori receive the clinically and culturally appropriate care they need and are entitled to. This research explores the experiences of Māori occupational therapists working in the mental health sector in Aotearoa and how they navigate cultural safety and cultural load within this space. Kaupapa Māori methodology was the foundational underpinning of this research, with the interface and coming together of narrative inquiry. The Whakaāria analysis method was used to identify and interpret huahuatau (themes) within the research through the use of whakataukī and whakatauākī. Racism, cultural clinical workload, cultural obligations, internal psychological battles and protective factors contributing to cultural safety, were significant areas of interest in the findings. These huahuatau then informed a discussion around how to best support and mitigate culturally unsafe experiences for Māori occupational therapists in the mental health sector in Aotearoa.
Key Drivers of Consumer Interest and Success Factors in Live-Streaming Influencers with Large Followings
(Auckland University of Technology, 2025) Li, Zihan
Live-streaming has become a dominant force in reshaping online consumer behaviour in today's rapidly evolving digital economy. Platforms like TikTok provide fertile ground for influencers to build large followings and drive real-time commerce. This study investigates why certain live-streaming influencers—particularly those with millions of followers—succeed in capturing and sustaining consumer interest, and which key traits contribute to their commercial and social success. The central focus is on understanding the dynamic interplay between consumer psychology, influencer attributes, and the mechanisms of sustained engagement in the context of live-streaming. While prior studies have examined the extensive impact of social media influencers (SMIs), few have investigated the underlying psychological reasons behind long-term follower behaviour. This study addresses three key gaps in the research: (1) the lack of consumer-centred investigations into why certain influencers attract more sustained attention; (2) the need to explore how real-time interactivity in live-streaming creates a participatory environment in which consumers actively shape their experiences; and (3) the insufficient understanding of how deeper emotional and cognitive gratifications, such as emotional connection or knowledge acquisition, drive long-term follower loyalty. By addressing these areas, this study enhances our understanding of what truly constitutes success for live-streaming influencers with large fan bases. The main research question guiding this study is: What are the key factors that drive consumer interest in live-streaming influencers with large followings? This is supported by three sub-questions: What are the reasons behind the popularity and success of influencers with millions of followers? What consumer preferences do these influencers satisfy? What specific traits differentiate them in a saturated digital market? Methodologically, this research adopts a qualitative approach, conducting in-depth semi-structured interviews with 20 participants who have either purchased or have not purchased items from live streams, but who were attracted as fans from the live broadcasts of influencers with millions of followers. Thematic analysis was employed to examine how various stimuli—ranging from influencer impression and content strategies to emotional expression and product presentation—shape consumer perceptions and behaviours. The S-O-R (Stimulus-Organism-Response) framework provided a guiding structure to trace the transformation of viewers into long-term followers, offering insight into the internal gratification processes that mediate external influencer effects. The findings reveal that successful influencers possess engaging content and promotional strategies and have personal traits that influence consumers’ emotions, such as being brave, healing and knowledgeable. Viewers are attracted to influencers who satisfy their inherent emotional and intellectual needs, such as providing comforting content or practical product insights. These gratifications foster trust and establish long-term and stable relationships between the influencer and their follower. The shift towards more rational consumer behaviour in the live-streaming space was an unexpected yet critical insight. Consumers are no longer driven solely by impulse or price incentives; instead, they increasingly value whether influencer content aligns with their emotional states and cognitive needs. Positive energy has also become a pivotal factor in establishing sustainable engagement. This study contributes to both academic and practical discussions by demonstrating that consumer interest is cultivated through a deeper psychological process, rather than being solely captured through external appeal. The findings offer actionable insights for marketers, influencers, and platforms aiming to foster lasting engagement and brand loyalty in the competitive landscape of live-streaming commerce.
ChatFlags: An AI-Powered Semaphore Interactive System
(Auckland University of Technology, 2025) Huan, Yan
This study presents the development of ChatFlags, an intelligent system for flag recognition and interaction. YOLO11 was selected as the visual backbone based on experiments involving five flag classification tasks. The custom dataset was refined and expanded to address the lack of publicly available resources. An improved model, YOLO-AKEMA, integrating attention mechanisms and adaptive convolution, achieved higher accuracy across 27 flag categories. The user interface was built by using the AI platform Dify, supporting conversational interaction. To mitigate hallucinations in large language models, a retrieval-augmented generation (RAG) framework was constructed by using curated flag documents and the BGE-M3 embedding model. Finally, the DeepSeek language model was integrated via workflow orchestration to complete the system. ChatFlags supports natural language dialogue, flag video analysis, knowledge quizzes, and text-to-image/video conversion. Its multimodal features enhance interactivity, offer a scalable solution for flag language education, and extend the integration potential of vision and language models.
The Effect of Automated vs Non-Automated Advertising on Customer Response to Social Media Ads: The Mediator Role of Perceived Personalisation
(Auckland University of Technology, 2025) Vienings, Daniel Roedolf
The way advertising content is produced, tailored, and distributed has changed dramatically because of the rapid development of artificial intelligence (AI) in digital marketing, especially on social media platforms. AI-generated advertising raises questions about customer trust, perceived manipulation, and emotional discomfort, even though it has advantages, including improved targeting precision and real-time optimisation. These issues are consistent with the personalisation paradox, which holds that emotions of creepiness can coexist with enhanced relevance brought about by personalisation. Even though AI advertising is becoming increasingly popular, little empirical research has been done on how consumers react to and perceive ads produced by AI as opposed to those developed by humans. To address this gap, this study looks at how perceived personalisation, relevance, and creepiness influence consumer engagement (clicks, likes, comments, and shares) and ad avoidance. It also investigates how business experience, being part of a moderated mediation model, influences perceptions of personalisation. A conceptual model that accounts for the direct and indirect impacts of AI-generated advertising on consumer behaviour is tested in this study using a quantitative experimental approach. Three advertising conditions, AI-generated personalised ads, AI-generated non-personalised ads, and human-created non-personalised ads were used in a between-subjects online experiment. A structured survey was used to gather information from a sample of n = 150 participants. The study tested the hypothesised links between ad type, perceived personalisation, relevance, creepiness, and behavioural effects by evaluating the measurement and structural models using Partial Least Squares Structural Equation Modelling (PLS-SEM). The findings indicate that, when compared to non-AI (human-generated) content, AI-generated advertisements did not significantly increase perceived personalisation, indicating that algorithmic targeting by itself does not necessarily provide improved customer perceptions. However, perceived personalisation had a significant impact on both creepiness and relevancy, demonstrating the dual nature of personalised advertising. The personalisation paradox is reflected in this dynamic: while personalised content might increase engagement potential, it can also cause psychological discomfort if it is viewed as being overly intrusive. Additionally, perceptions of personalisation were not significantly moderated by business experience, casting doubt on the idea that exposure to digital marketing in the workplace automatically results in positive assessments of AI-driven content. This thesis makes theoretical contributions by extending the SNS-Post Processing Framework to AI-generated advertising contexts, providing insights into how cognitive and affective responses to personalisation influence behavioural outcomes. Additionally, it also contributes to the limited empirical research on ad avoidance as a defensive mechanism triggered by AI-driven targeting. To promote customer trust and engagement, the study offers practical advice to digital platforms and advertisers on how to strike a balance between AI automation, human creativity, ethical transparency, and emotional resonance.
Migrant Labour Exploitation in New Zealand: A Critical Analysis of the Political Discourse
(Auckland University of Technology, 2025) Kohli, Harsimran Singh
Migrant labour exploitation often occurs in countries where there are significant numbers of seasonal migrant labour (SML) workers. A significant share of New Zealand’s workforce is made up of SML and due to this extensive reliance on imported labour, cases of SML exploitation are common. Earlier studies have documented both the occurrence of SML exploitation and its drivers, such as employer-bounded visas, high recruitment fees, and weak enforcement (Bi, 2016; Collins & Stringer, 2019; Stringer & Michailova, 2019). However, how politicians frame exploitation and how this framing influences the policies that facilitate or control exploitation have received limited attention. Thus, the purpose of this study is to examine how politicians represent the issue of SML and the main themes in the New Zealand political domain related to SML exploitation. In this sense, narratives are not just rhetoric, they are policy-making tools that raise a problem, legitimise some solutions, and “rule out” other options. Politicians – ministers and party leaders – define the rules of migrant labour schemes. Looking at the narratives of politicians about SML shows not only their plan for migrants but how they frame the problem, how they portray stories, and what policies politicians will use to respond exploitation. Using a qualitative research design, secondary data were collected using press releases, Hansard debates, newspapers, and websites of political parties. Data were analysed using thematic analysis. Findings show that politicians – notably those from New Zealand’s two major parties – present narratives in support of continued and increased SML for the New Zealand economy because it helps achieve export targets, industry production demands and benefits for Pacific workers. However, SML impacts local labour and is linked with exploitation by employers. Thus, SML is presented as both beneficial and fraught. Because the beneficial aspect lifts the broader New Zealand economy, successive governments remain reactive – reforms are done when scandals emerge or due to media pressure – to correct compliances instead of addressing migrant exploitation at a structural level.
ChatClothes: An AI-Powered Virtual Try-On System
(Auckland University of Technology, 2025) Zhang, Yuchao
With the advancement of deep learning, latent diffusion models, and large language models (LLMs), virtual try-on (VTON) has emerged as a promising solution for personalized fashion experiences in online shopping, digital design, and augmented retail. This thesis proposes ChatClothes, a modular and multimodal VTON system that integrates controllable diffusion-based generation with dialogue-driven garment interaction. The system architecture is orchestrated by Dify, with ComfyUI managing the visual generation pipeline and Ollama hosting local LLMs. At its core, ChatClothes employs DeepSeek, a customized large language model that interprets natural language instructions and transforms them into structured prompts for image generation and interactive refinement. This prompt-based guidance enhances semantic alignment and enables intuitive user control beyond predefined attribute labels. To improve structural consistency and detail fidelity in image synthesis, this work introduces Low-Rank Adaptation (LoRA) for fine-tuning the original OOTDiffusion model. Without altering the backbone architecture, this strategy focuses on enhancing pose alignment, hand generation accuracy, and garment texture reconstruction. By integrating LoRA modules, the model achieves effective adaptation and fine-grained refinement even under limited training resources. To support garment classification, YOLO12n-LC, a lightweight variant based on YOLO12n, is developed to balance accuracy, speed, and model size. It achieves competitive performance across multiple clothing categories while maintaining feasibility for device-level deployment. A complete system workflow connects image preprocessing, language understanding, garment classification, image synthesis, and output evaluation. Experiments on datasets such as DressCode and VITON-HD demonstrate the system’s initial validation in terms of realism, controllability, structural preservation. This work presents a unified framework bridging vision-language interaction with diffusion-based generation, establishing a foundation for scalable, user-centered, and device-adaptable fashion AI systems applicable across e-commerce, AR fitting mirrors, personalization platforms, and automated outfit design.
Epidemiology of Tuberculosis and BCG Vaccine Uptake Among Pasifika in Aotearoa New Zealand
(Auckland University of Technology, 2025) Schutz, Rhonita
Tuberculosis (TB) has many known risk factors that contribute to its persistence worldwide, particularly in developing countries. Aotearoa New Zealand (NZ), as a high-income country, is fortunate to fall in the low-risk category. However, the incidence of TB is more common in particular ethnic groups such as Pasifika. The Pasifika population exhibit a unique set of vulnerabilities that increase their susceptibility to TB disease. Recent studies have reported a 13.1 per 100,000 notification rate for Pasifika in contrast to 0.5 per 100,000 among the European ethnic group in Aotearoa NZ. The Bacillus Calmette–Guérin (BCG) vaccine is the only licensed vaccine widely used to prevent TB. The World Health Organization (WHO) recommends BCG inclusion in neonatal vaccination schedules, depending on the TB epidemiology of the country. Many high incidence countries in the Pacific have been recommended to administer one dose to all neonates. However, countries with a low incidence of TB, such as Aotearoa NZ, can focus vaccination on high-risk groups only, through recommendation by healthcare workers to parents. Recommendation is based on specific eligibility criteria, meaning that all Pasifika children born in Aotearoa NZ are not necessarily entitled to BCG vaccination. Therefore, this study aims to identify strategies for improving TB prevention efforts by identifying the level of protection against TB for Pasifika. Given their higher TB burden and vulnerabilities, assessing the effectiveness of the BCG vaccination programme to capture at-risk populations is particularly important for these vulnerable ethnic groups. A Pasifika framework, Te Kora, was employed to guide a convergent parallel mixed methods design. To understand the epidemiology of TB and the BCG vaccine uptake among Pasifika in Aotearoa NZ, a quantitative observational study and a qualitative interpretive descriptive study were undertaken. TB and BCG data from 2006 to 2023 were descriptively analysed. Maroro (conversations) was used as the method to generate qualitative data to understand the perceptions of healthcare professionals on the BCG vaccination programme and TB prevention efforts. The qualitative data was then analysed with Conventional Content Analysis (CCA). The quantitative results showed that Pasifika and Asian populations in Aotearoa NZ had the highest TB incidence rates from 2006 to 2023, with average incidence rates of 11.4 per 100,000 (confidence interval (CI): 8.2 – 15.3) and 27.5 per 100,000 (CI: 23.5 – 32) respectively. Further analysis of the Pasifika population showed a higher percentage (average of 55%) of TB incidence among the less dominant Pasifika ethnicities such as Kiribati, Tokelau, Tuvalu and Niue. BCG vaccination rates were steadily high among the Asian population, with an average rate of 1000.7 per 100,000 (CI: 975.8 – 1026.2). BCG rates were also reasonably high among Pasifika, however, there was an observed rapid significant decrease to exceedingly low vaccination rates from 2011 to 2023. The average BCG vaccination rate for Pasifika was 519.4 (CI: 500.3 – 539.3) per 100,000, however, based on the significant decrease which reached vaccination rates as low as 26.6 per 100,000 (CI: 21.7 – 32.2), the Pasifika population were significantly under-vaccinated. Further analysis of the Pasifika vaccination rates indicated that Pasifika ethnicities were proportionately vaccinated relative to their population sizes except for the Cook Island Māori and Niuean ethnicities. This highlights the need for an increase in BCG vaccine uptake for the Pasifika population, especially among the less dominant Pasifika ethnicities, which have the highest proportion of TB notifications. Three main categories were constructed from the qualitative data that outlined barriers for the BCG vaccination programme and TB prevention efforts. These include systemic gaps in identifying at-risk groups, which identified knowledge gaps among healthcare workers and fragmented referral processes. The second category, perceptions of TB disease and BCG vaccine among migrants and Pasifika communities, identified stigma and migrants’ perception of TB risk. The third category is system-based factors that affect BCG uptake and TB reduction, such as the BCG policy and programme changes, the Pasifika umbrella, and effective communication. This study explored the effectiveness of current TB prevention efforts, particularly the BCG vaccination programme among at-risk populations in Aotearoa NZ with a focus on the Pasifika population. The findings highlighted specific barriers from the qualitative data that must be addressed in order to improve the current BCG vaccination programme. The study recommends system-level improvements to progress TB prevention among Pasifika, such as increasing training for healthcare workers to enhance risk assessment, uniform referral processes for all regions, disaggregated data, and health promotion strategies specifically to target TB stigma. The results of this work highlighted particular improvements that are needed to protect the Pasifika populations from TB.
Rainbows in Summer
(Auckland University of Technology, 2025) Powrie, Cherie
The thesis is an excerpt of “Summer Temptation”, a contemporary fiction novel that meets and challenges the romance genre, peppering it with sauciness and characteristics of what could be deemed a tragic romance/ drama featuring the evolving and pressured relationships between Jessica and Lucas Park, a married couple in a seemingly stable and happy marriage. Jessica succumbs to the persistent charms of an old school friend and engages in an illicit affair leading to betrayal and turmoil in her marriage to loyal and loving Lucas. Jessica’s inner circle offer differing viewpoints on her transgression based on their interpretations and differing loyalties, and the end result jeopardises everything she has worked for. The story is told through pivoting points of view; Jessica addresses the reader in the first person while the omniscient narrator interprets Lucas’s behaviour to convey his point of view. Both voices are disrupted as the protagonists and bystanders interrupt the flow with snippets of alternative points of view (POV) in each chapter by either the 3rd person for Jessica or in 1st person for Lucas (shown in italics). An important part of the story are their adolescent twin daughters, Olivia and Tanya who are also ultimately affected by the infidelity and the reader is invited to consider their voices through small chapters dedicated to them in a mix of 1st and 3rd person. The novel is set in Queenstown and surrounds, with iconic New Zealand scenery as a backdrop, and is also brought to life as a character in its own right. Readers of the novel get to take a front seat and make their own decisions on the complexities of romance, led through love, betrayal, hurt and regret. The thesis aims to meet and challenge audience expectations of the traditional romance genres as it canvasses these themes.
Guardians of Matariki: Evaluating the VisionOS for Māori Storytelling
(Auckland University of Technology, 2025) Perese, Thomas
This exegesis presents the design, development, and internal evaluation of Guardians of Matariki, a spatial storytelling prototype for Mixed Reality, developed specifically for the Apple Vision Pro (AVP). The project addresses the research question: How can immersive spatial storytelling enhance users' engagement with Māori culture through a Mixed Reality application? Centred on the Matariki star cluster, the prototype explores how cultural content can be communicated through varying levels of immersive interaction. A structured methodology was applied using the Research Onion model. The research philosophy combined interpretivism, pragmatism, and kaupapa Māori to support culturally grounded, reflective development. The approach employed abductive reasoning and Design-Based Research (DBR) within a mixed methods framework, with qualitative and quantitative self-assessment, journaling, and heuristic walkthroughs used for data collection and analysis. Three levels of immersion were developed and evaluated: a 2D windowed interface, a 3D volumetric interface, and a full space immersive environment. NASA Task Load Index (NASA-TLX) and the Technology Acceptance Model (TAM) were used to assess cognitive load, usability, and engagement. Results indicated that increasing spatial immersion correlated with enhanced narrative embodiment, reduced cognitive load, and improved cultural clarity. The development process followed an Agile (SCRUM) sprint structure, using tools such as Reality Composer Pro, Xcode, SwiftUI, and Meshy AI. Culturally aligned UX heuristics and Apple’s Human Interface Guidelines informed design decisions. Limitations include the absence of external user testing and co-design due to time and hardware constraints. All findings reflect internal evaluation and are not generalisable. This study contributes to Indigenous digital storytelling, spatial computing, and immersive design. It demonstrates a culturally responsive method for developing immersive applications and proposes a foundation for future work involving user collaboration, cross-platform delivery, and expanded narrative design.
Climate Crisis, Tourism and Sustainable Aviation Fuel: A Case Study of Queenstown, New Zealand
(Auckland University of Technology, 2025) Otake, Yuta
The global travel industry is a significant contributor to carbon dioxide (CO₂) emissions, with air travel accounting for a substantial share. As climate change intensifies, the need for sustainable solutions in tourism becomes increasingly urgent. This thesis investigates the potential effects of using Sustainable Aviation Fuel (SAF), an alternative jet fuel that was developed to mitigate carbon emissions, focusing on its implications for Queenstown, New Zealand, a premier tourist destination with one of the busiest regional airports in the country. Queenstown’s reliance on air connectivity for its tourism economy renders it an illuminating case for exploring decarbonization strategies within the travel industry. The research is grounded in the global sustainability agenda, particularly the United Nations Sustainable Development Goals (SDGs), and the net-zero 2050 carbon emissions framework. It addresses a notable gap in the literature: while SAF has been extensively studied in engineering contexts, its application and impact within tourism remain underexplored, especially in New Zealand. This study aims to bridge the gap by examining how the use of SAF could influence the Queenstown travel industry and by capturing the perspectives of key stakeholders involved in its use across the air travel sector. Employing a qualitative case study methodology, the research is framed within an interpretivist paradigm, supported by a relativist ontology and constructivist epistemology. Data collection involved semi-structured interviews with key stakeholders in the Queenstown travel industry, alongside document analysis of relevant policy papers, industry reports, and academic literature. Thematic analysis was used to identify recurring patterns, insights, and key stakeholders’ perspectives, resulting in five major themes and twelve sub-themes that reflect the multifaceted nature of SAF adoption. Findings reveal that SAF holds promise for reducing aviation-related emissions and aligning Queenstown’s travel practices with global sustainability goals. Some findings highlighted that the use of SAF can be a potential way to achieve net-zero emissions and enhance Queenstown’s environmental credibility as a destination. However, some key stakeholders who participated in the semi-structured interview mentioned significant barriers to implementation, including high production costs, limited supply chains, and the absence of robust policy frameworks to support widespread adoption. These challenges underscore the need for coordinated efforts among government bodies, airlines, fuel suppliers, and tourism operators to create enabling conditions for SAF integration. The study concludes that while SAF is not a panacea, it represents a pivotal step toward sustainable air travel. Its successful deployment in Queenstown could serve as a model for other tourism-dependent regions in New Zealand seeking to mitigate their carbon footprint. The research contributes to the broader discourse on sustainable tourism by offering practical insights into the role of alternative fuels in climate action. It also emphasises the importance of stakeholder collaboration, policy innovation, and public awareness in driving the transition to low-carbon air travel. Ultimately, this thesis advances the understanding of SAF within a tourism context and provides a foundation for future research and policy development. By situating SAF within the framework of climate crisis response and sustainable tourism, it highlights the urgency and complexity of decarbonising the air travel industry in a rapidly evolving global landscape.
Fungal Electronics: An Exploration into the Conductivity of Mycelium
(Auckland University of Technology, 2025) Saran, David
This thesis explores the emerging field of Fungal Electronics, a domain that remains vastly underresearched despite fungi outnumbering plant species in global biodiversity. Motivated by this imbalance, the research combines hands-on experimentation, cultivating and monitoring fungal growth, with digital translation methods that convert raw bioelectrical data into sonic and visual formats. Using a PicoLog data logger, voltage signals from various fungi are transformed into MIDI compositions, forming the basis for experimental musical tracks that are further refined in audio workstations. The project also incorporates speculative design through macro videography and environmental soundscapes, presenting fungi as both biological agents and creative collaborators. The final output includes a multi-sensory installation where fungal music synchronizes with color-coded visual overlays and ambient footage, offering audiences a new lens through which to interpret nonhuman intelligences. These processes and findings are critically examined and contextualized within a written exegesis, with the aim of contributing novel insights into the interface between biology, sound, and digital media.

Browse

Recent Submissions