CLSBE - Dissertações de Mestrado / Master Dissertations
Permanent URI for this collection
Browse
Browsing CLSBE - Dissertações de Mestrado / Master Dissertations by Sustainable Development Goals (SDG) "10:Reduzir as Desigualdades"
Now showing 1 - 10 of 38
Results Per Page
Sort Options
- AI biases and marginalized stakeholders : inquiry into AI implementation practices in dating appsPublication . Lesnikova, Elena; Fioravante, RosaThis study explores how the use of artificial intelligence (AI) in dating apps can reinforce or mitigate biases against marginalized stakeholders groups, including racial minorities, women, and LGBTQ+ users. While AI improves user engagement and safety, it also risks reinforcing systemic biases, such as racial and gender-based discrimination. Through a qualitative thematic analysis of publicly available data, this research examines AI implementation in popular dating apps like Tinder, Bumble, and Grindr. Findings show that AI-driven algorithms often prioritize engagement over fairness, disproportionately excluding marginalized groups. The study highlights AI9s dual role: it can mitigate harm through features like harassment detection but also perpetuates inequalities when unethically designed. Theoretical contributions include applying stakeholder theory and ethical AI frameworks to digital matchmaking, emphasizing companies' ethical responsibilities to marginalized users. Practical recommendations focus on algorithmic transparency, fairness-aware machine learning, and inclusive AI governance. This research highlights the need for ethical AI practices in dating apps to ensure inclusivity and fairness, contributing to broader discussions on AI ethics and digital discrimination.
- Comparing the effectiveness of default, framing, and social norm nudges on charitable giving and donor happiness : an experimental studyPublication . Strziga, Lucas Maximilian; Fernandes, DanielThis study investigates how behavioral nudges can support charitable fundraising to overcome the challenge of gathering funds by comparing their effectiveness on donation behavior and donor happiness. Although the impact of individual nudges has been widely studied, limited research explores their comparison under identical, neutral conditions. Moreover, the isolated effects on donor happiness remains unexplored, carrying potential ethical and long-term retention implications. To address this gap, the study compared default, framing, and social norm nudges under identical neutral conditions exploring their isolated effects on donation amount, participation rate, and donor happiness. A quantitative between-subject online experiment (N=408) was conducted. Participants were randomly assigned to one of four conditions (control, default, framing, or social norm) and asked how much of a 50¬ Amazon gift card prize they would like to donate, followed by a self-report of their post-donation mood. The analyzed results showed no significant difference between conditions on donation amount, participation rates, or donor happiness. Independent of the nudge types a high baseline donation rate and a significant correlation between higher donation amounts and increased happiness was registered. These findings suggest that the isolated effects of nudges are weaker than previously assumed and that their power potentially derives from underlying contextual or psychological mechanisms. Hence, their effect may not originate from the nudge itself but from its ability to combine and convey different contextual and psychological cues.
- Decoding blame : the effect of humanizing AI on responsibility attribution in financial decision-makingPublication . Zwingelberg, Alexander Alfons; Mendonça, CristinaThe steadily growing integration of artificial intelligence (AI) in the financial service industry offers new opportunities for automation of processes, but also challenges traditional human centric advisory models. Robo-advisors are replacing face-to-face customer interactions; thus, it becomes more important than ever to understand how individuals assign responsibility in joint decision-making processes between humans and AI. Perhaps one of the factors that can influence these judgments is anthropomorphism: the human tendency to attribute human-like characteristics, such as emotions or intentions, to non-human agents like AI. This master’s thesis examines how anthropomorphic design cues and decision outcomes (success vs. failure) influence responsibility attributions in AI-assisted financial decisions. Drawing on attribution theory, a 2×2 between-subjects experiment (N = 366) was conducted in which participants interacted with a simulated financial AI advisor applying two levels of human-likeness. After the service encounter, participants rated the extent to which they felt the AI and themselves were responsible for the decision outcome. Findings reveal that participants overall attributed more responsibility to themselves than to the AI. Contrary to expectations, a more human-like AI did not significantly affect the level of responsibility assigned to the AI. The results suggest that responsibility attribution depends on various internal and external factors, such as perceived task complexity and locus of control. In addition, outcome valence only influenced the responsibility assigned to the AI, but not to the participants. Summarized, the results of this dissertation highlight the complex psychological mechanisms of interactive decision-making between AI and humans.
- Democratizing finance : how digital investment platforms have transformed retail investors behaviorPublication . Sansone, Alessia; Rita, MiguelThis dissertation investigates the ways in which digital investment platforms have reshaped retail investor behavior by lowering traditional barriers to market entry and introducing new drivers of participation, risk-taking, and education. Drawing on behavioral finance theories (Prospect Theory, Modern Portfolio Theory, Efficient Market Hypothesis) and technology adoption frameworks (Technology Acceptance Model, Diffusion of Innovations), the study combines a qualitative interview with an industry expert and a quantitative survey on 200 investors. The primary objectives are to (1) assess how digital investment platforms influences market participation and asset allocation, (2) evaluate the role of perceived ease of use and usefulness in technology uptake, (3) identify gaps in financial literacy and their impact on risk exposure, and (4) explore generational differences in digital investment strategies linked to technology adoption. Key findings reveal that digital platforms significantly increase participation among younger cohorts, encourage higher allocation to volatile assets such as cryptocurrencies, and amplify speculative behavior in the absence of structured educational content. Perceived ease of use and usefulness emerge as dominant factors driving adoption, while self-directed learning through informal channels does not always translate into prudent investment decisions. The dissertation concludes that digital platforms hold great promise for democratizing finance but must integrate learning modules, AI-driven risk assessments, and hybrid advisory services to foster informed decision-making and sustainable investor engagement. These insights contribute to both academic discourse and practical platform design, offering recommendations for policymakers and financial institutions to balance innovation with investor protection and long-term market stability.
- Did the period of low interest rates undermine the European Central Bank's ability to effectively lower inflation in the post-COVID-19 economic environment?Publication . Born, Clemens Anton Christoph; Ilseven, EkinThe thesis, "Did the period of low interest rates undermine the European Central Bank’s ability to effectively lower inflation in the post-Covid-19 economic environment" by Clemens Born, analyses how a low-interest-rate environment affected the ECB’s ability to combat inflation. In a normal fiscal environment, the central bank controls inflation through key interest rates and transmission mechanisms, which can lose effectiveness in an expansive monetary environment. While traditional monetary theory suggests a negative relationship between interest rates, lending, and inflation, empirical observations from the pre-pandemic period indicate a contrarian effect. This study examines how prolonged low interest rates and quantitative easing (QE) impacted inflation dynamics in the euro area. Using Ordinary Least Squares (OLS) regressions and Ttests, it assesses monetary transmission before and after the Covid-19 pandemic. The results suggest that excess reserves, rather than supporting credit creation, contributed to inflationary pressures once economic activity resumed. Moreover, interest rate hikes failed to curb inflation effectively in an ultra-loose monetary environment. By examining these dynamics, this thesis contributes to the debate on monetary policy effectiveness by empirically testing whether the ECB’s ability to control inflation was constrained. Unlike previous studies, it considers COVID-19 as a catalyst for inflation, as economic disruptions and fiscal and monetary interventions accelerated price dynamics. The findings suggest that excess liquidity entered the real economy through the credit channel, fueling inflation despite rising interest rates. Consequently, the ECB’s ability to manage inflation post-pandemic was limited, raising concerns about the long-term effectiveness of conventional monetary policy.
- Do investors pay for impact? : an empirical analysis of valuations in private equity and venture capitalPublication . Bauck, Thomas David; Buchner, AxelIn a capital market that is increasingly shaped by concerns around sustainability, impact investing promises to deliver measurable social or environmental benefits alongside a financial return. But do investors in private equity and venture capital actually pay for impact, or is it just a branding tool? This study provides the first large-scale, valuation-based evidence that investors are willing to pay a premium for impact in private markets. Using a unique dataset of 160 “impact” deals matched to a control group of 3,500+ private transactions, I employ propensity score matching and doubly robust regression to isolate the 8impact premium9. Results reveal that impact firms are valued 90-99% higher than comparable non-impact firms, even after controlling for industry, geography, and deal characteristics. This premium is higher for social impact over environmental impact, larger in emerging markets than developed markets and disappears in cross-border deals, suggesting that investor preferences are highly context-dependent. The results challenge the view that impact is a non-financial phenomenon and show that, in practice, social and environmental value is priced into private equity. For fund managers, institutional investors and policymakers, this work offers granular insights and guidance into impact investing in private markets.
- Does financial education improve financial literacy and long-term financial behavior? : evidence from financial education state-mandated requirements in the United StatesPublication . Câmara, Sara Margarida Gomes; Reis, HugoThis dissertation examines the causal impact of financial education on financial literacy and long-term financial behavior in the United States, exploiting policy-driven variations in financial education curriculum requirements across states and years. While research on financial education effectiveness remains somewhat limited and inconclusive, its relevance has grown due to rising concerns about financial literacy levels, particularly among younger individuals, both in the United States and globally. The analysis is based on data from the National Financial Capability Study, between 2012 and 2021, which is a triannual survey assessing financial knowledge and behaviors. Using an Instrumental Variable (IV) approach, accounting for the timing and structure of state-mandated financial education policies, the study first analyzes a full sample including individuals of all ages across states with and without mandates, then narrows the focus to specific age groups and policy contexts for clearer interpretation. The analysis reveals that, among those who only received financial education driven by the state-mandated requirements, the overall impact on financial literacy and long-term financial behaviors is positive. However, statistical significance is only maintained for two outcomes throughout the study: stock diversification knowledge and savings account behavior, though these effects become less pronounced when restricting samples by age. Financial literacy and behavior gaps persist among women, the unemployed, and individuals with lower education levels, even after receiving financial education.
- Emerging challenges in workplace inclusion : addressing bias and stigmatization against neurodivergent employees through organizational trainingPublication . Trindade, Yara Sophia Alves; Fioravante, RosaDiversity and Inclusion literature has focused on organizational initiatives emphasizing gender, ethnicity and sexual orientation. However, recently, a growing attention has been devoted to the inclusion of neurodivergent people, nonetheless remaining an understudied social group. This study seeks to enhance current knowledge on how organizations can support, neurodivergent employees through training programs. More in detail, this study explores how training programs contribute to fostering an inclusive organizational culture. This study deploys an inductive approach, relying on primary and secondary data analyzed through qualitative Gioia methodology. The research focuses on how training programs help address structural, psychological, and cultural barriers faced by neurodivergent employees and the role of targeted training programs in overcoming these challenges. Findings of this study allowed to identify three key dimensions where training programs enhance neurodivergent inclusion: First, strengthening organizational infrastructure and leadership, through modified recruitment models, adapted performance evaluation criteria, and leadership training. Second, fostering psychological safety and authenticity by reducing environmental stressors, supporting peer networks, and promoting inclusive communication styles. Third, driving cultural change through strengths-based inclusion and increasing awareness of neurodivergent capabilities. Based on these findings, the study provides suitable managerial implications for improving workplace inclusion by addressing bias and supporting neurodivergent employees. Furthermore, the study contributes to literature on business ethics, by providing original insights into diversity and inclusion, to overcome current organizational challenges in aiming at ethical goals targeting neurodivergence in the workplace.
- The ethical risks of overlooking human potential in AI recruitment : current unethical outcomes and future mitigation strategiesPublication . Prill-Gonzalez, Johanna Marie; Fioravante, RosaThe rapid integration of Artificial Intelligence (AI) into recruitment processes has revolutionized hiring practices, offering significant opportunities while introducing profound ethical risks. This thesis aims to provide an overview of current debates on the ethical risks and opportunities posed by AI recruitment and their mitigation strategies, with a particular focus on the underexplored risk of AI-Defined Candidate Evaluation Filters Overlooking Human Potential. The central research question is: How can organizations deploy AI recruitment systems without risking overlooking human potential, especially among candidates with more individualistic or non-conforming backgrounds? Utilizing a qualitative systematic literature review, the study synthesizes peer-reviewed articles as well as established books in the field to assess AI's ethical challenges and identify gaps in the literature. Findings reveal that AI-defined evaluation filters can exclude capable candidates from diverse backgrounds, amplify biases, undermine inclusivity, and devalue human judgment. These various outcomes result in workforce homogeneity, stifling innovation and adaptability. To better understand the interconnected risks, the framework categorizes these into three dimensions: Inclusivity & Accessibility, Depth & Accuracy of Candidate Evaluation, and Organizational Adaptability & Long-Term Impact. Finally, the thesis addresses these challenges by proposing a mitigation framework based on two dimensions: resource intensity and implementation horizon. Strategies include expanding outreach channels, conducting algorithmic bias audits, refining data practices to exclude sensitive attributes, and fostering human-AI collaborative evaluation processes.
- Exploring the impact of foreign language use on managerial decision-making : a qualitative study with German-speaking managersPublication . Kürbis, Lea-Marie Annette; Mendonça, CristinaAs global business becomes increasingly multilingual, understanding how foreign language use affects decision-making is essential. This study investigates how managers in a multinational company perceive the influence of the foreign language effect on cognitive processing, risk perception, emotional engagement, and moral judgment. Using semi-structured expert interviews with 15 native German-speaking managers who regularly use English at work, the study employs qualitative content analysis to explore how foreign language use is perceived to shape different aspects of managerial decision-making. While many participants appeared to experience stable cognitive and emotional engagement across languages, others described slower reasoning, greater deliberation, or more restrained communication 3 particularly in unfamiliar or high-stakes contexts. These findings suggest that subtle yet meaningful changes might have occurred in both decision outcomes and the processes by which decisions were made and communicated. Cultural norms may have exerted a more substantial influence on ethical judgments than language, and linguistic confidence appears to be a relevant but insufficiently examined variable. The study expands the current literature by offering context-sensitive insights that challenge the generalizability of experimental foreign language effect findings. It emphasizes the importance of considering language proficiency and organizational context when assessing foreign language use in managerial decision-making in research and practice. Furthermore, the results highlight the importance of raising awareness about these often-overlooked dynamics and encouraging the integration of language-sensitive reflection and training into managerial practice and future research.