Research in environmental psychology demonstrates a persistent gap between environmental attitudes and observed pro-environmental behavior, challenging linear models linking awareness, knowledge, and action. Drawing on Gifford’s “Dragons of Inaction” framework, this study quantitatively examines psychological barriers to sustainable behavior using large-scale cross-national data.

Using the International Social Survey Programme (ISSP) Environment Module 2020, presented work aims to develop and empirically a dragon proxy index (DPI) that operationalizes psychological barriers with predictive capacity at the individual level. Data preparation, scale construction, and inferential analyses are conducted using Stata, enabling transparent, reproducible, and cross-nationally comparable analytical workflows. Exploratory factor analysis with rotation is applied to evaluate the dimensional structure of proxy indicators and their correspondence to the seven theoretical barrier categories. Construct validity, internal consistency, and overall model fit are assessed using factor loadings and related diagnostics within Stata.

A key contribution of this study is addressing challenges of reliability and reproducibility in index development across countries. Expected results indicate statistically significant variation in psychological barrier profiles across national contexts and outcome measures. By integrating ISSP data with a Stata-based psychometric framework, this study proposes a robust quantitative tool for diagnosing behavioral barriers and informing targeted sustainability interventions.

This study presents an applied Stata-based analysis of non-take-up in Mexico's old-age pension during the transition from targeted eligibility to near-universal coverage. Using harmonized microdata from the National Household Income and Expenditure Survey (ENIGH) and poverty estimates from the National Council for the Evaluation of Social Development Policy (CONEVAL) for 2016 to 2024, I estimate a two-step Heckman selection model to separate two related outcomes: the probability that an eligible older adult receives the pension and the pension income observed among beneficiaries. This approach addresses sample-selection bias because benefit amounts are observed only for participants, and it allows territorial differences in implementation to be examined within a single framework. The results show that non-take-up declined markedly after universalization and that cross-state differences in access became more compressed over time. However, exclusion did not disappear. Indigenous-language speakers and older adults below the poverty line remained less likely to receive the benefit, and important differences persisted in the amounts actually received, especially in southern states. The study shows how Stata can be used to combine harmonized survey and poverty data, estimate selection models, and visualize territorial variation in social policy implementation.

Introduction: Atrial fibrillation (AF) is common in critically ill patients and may worsen hemodynamic instability. Cardioversion is often required for rhythm control. Factors influencing the choice between electrical and pharmacologic cardioversion remain unclear. This study evaluated whether clinical variables, including age, acute coronary syndrome, acute respiratory failure, chronic heart failure, and hypertension, are associated with cardioversion strategy.

Methods: A retrospective observational study was conducted using the Atrial Fibrillation in Critically Ill dataset. Critically ill adult patients with preexisting AF admitted to the ICU were included. The primary outcome was cardioversion strategy (electrical versus pharmacologic) during hospitalization. Predictor variables included age, ACS, ARF, CHF, and HTN. Multivariable logistic regression analysis was performed using Stata (BE 19) to estimate adjusted odds ratios with 95% confidence intervals.

Results: Among 1,705 critically ill patients with AF, 179 (10.5%) underwent electrical cardioversion and 290 (17.0%) received pharmacologic cardioversion. ARF was present in 397 patients (23.3%), ACS in 212 patients (12.4%), CHF in 693 patients (40.6%), and HTN in 838 patients (49.1%).

Discusion: Hypertension was associated with increased electrical cardioversion, possibly representing a more stable patient subset. Overall, these findings support a context-driven approach to cardioversion in critically ill patients, prioritizing physiologic stability and reversibility of illness.

This research compares the socioeconomic background of college students in fraternities with those who are not. Using data from the Panel Study of Income Dynamics (PSID) and its Transition into Adulthood Supplement (TAS), we compare family income and net worth between students who participate in Greek life and those who do not. After adjusting for inflation and transforming distributions using natural logarithms, we find that fraternity-affiliated students come from families with significantly higher income and wealth levels. On average, Greek participants report family net worth over 60% higher and family incomes over 40% higher than their non-Greek peers. These gaps are shown across kernel density estimates, violin plots, and box plots and are proven through the use of regression and t-test analysis. We then contextualize these findings within the institutional framework of Greek life. Fraternities benefit from federal tax exemption under their IRC 501(c)(7) designation, under which they operate as nonprofit “social clubs” while avoiding taxation on dues and property.

Simultaneously, numerous studies have documented the public health risks associated with fraternity membership, including increased rates of sexual assault, substance abuse, epidemiologic outbreaks, and hazing. This creates a situation in which highly privileged students receive access to both institutional resources and public subsidies, despite generating social externalities and restricting membership in practice. Taken together, this work suggests that Greek life explicitly or implicitly discriminates based on socioeconomic status and is protected by public policy. By analyzing data within a broader institutional framework, this research contributes to the works on educational inequality and calls for reconsideration of how fraternities are regulated, subsidized, and integrated into university systems.

Reliable measurement of patient outcomes requires robust data validation and standardized reporting processes. This presentation introduces a Stata-based framework for data reporting, cleaning, and validation developed to support patient-reported outcome measures (PROMs) for coronary artery disease and stroke patients within a national health insurance system. The framework standardizes PROMs data collection across multiple healthcare providers by automating data cleaning, validation, and linkage procedures. Using automated loops, merge routines, and customized validation algorithms, Stata scripts were developed to harmonize variable structures, verify national identifiers, remove duplicates, and prepare PROMs datasets for analysis.

In parallel, a national population segmentation model was developed using Stata to stratify patients based on diagnosis profiles, comorbidity burden, and healthcare utilization patterns. The validated PROMs datasets were then linked to this segmentation model, enabling analysis of patient-reported outcomes across clinically and utilization-based population segments. This integration enabled the development of a PROMs segmentation dashboard for coronary artery disease and stroke patients, providing segment-level benchmarking and improved visibility of outcomes for high-need and high-cost populations. Findings highlight the potential of Stata as an end-to-end platform for national-scale outcomes measurement, supporting performance management and equitable care for high-need populations.

Background: Stata's ltable and sts list commands provide point estimates for cumulative mortality but do not produce confidence intervals for cumulative failure or excess mortality between groups. This limits their use for hypothesis testing and comparative analysis.

Methods: We developed a workflow that extends Stata's life table capabilities by adding confidence intervals using a simplified delta method. The approach: (1) obtain interval mortality probabilities using ltable, failure; (2) apply lowess smoothing (bandwidth = 0.4) to stabilize estimates; (3) propagate uncertainty using a delta method approximation, where the variance of cumulative mortality is computed as the cumulative sum of the squared mortality contributions; (4) calculate cumulative mortality and excess mortality with 95% confidence intervals. The workflow was implemented in R and validated against Stata's point estimates using HIV cohort data from Uganda (AFRICOS, 2013–2023).

Results: The workflow reproduced Stata's point estimates exactly while adding confidence intervals that revealed when excess mortality became statistically significant (e.g., overall excess: 3.4%; 95% CI: 1.3–5.6% by year 9). These intervals are not available from standard Stata commands. The method assumes independence between intervals, which provides a practical approximation for hypothesis testing.

Conclusion: This workflow enhances Stata's life table functionality by providing confidence intervals for cumulative and excess mortality. It can be implemented as a Stata ado-command or used as a template for researchers needing hypothesis tests for survival differences. The simplified delta method offers a practical solution while future work could develop a full gradient-based approach.

This poster presents practical and engaging methods for teaching statistics using Stata, with a focus on improving student understanding through applied learning. Many students find statistical concepts abstract and difficult to grasp when taught theoretically. This work demonstrates how integrating Stata into teaching can bridge the gap between theory and practice. The approach emphasizes hands-on learning, where students actively engage in data analysis using real and simulated datasets. Step-by-step demonstrations using do-files are employed to guide students through data management, statistical modeling, and interpretation of results. In addition, problem-based learning is introduced by presenting students with real-world research questions, encouraging them to select appropriate statistical methods and interpret outputs independently.

The poster also highlights key resources for teaching Stata, including official documentation, online tutorials, and instructor-developed materials such as guided exercises and recorded sessions. Visual tools such as graphs and tables are used to enhance understanding and interpretation. This approach has been found to improve student engagement, confidence, and practical statistical skills. The poster provides insights and strategies that can be adopted by educators, students, and new professionals interested in teaching or learning statistics using Stata.

This presentation introduces catllm, a Stata package for LLM-powered text analysis that requires no NLP or AI experience. The package provides four commands. cat-llm classify assigns open-ended text responses to user-defined categories, with options for chain-of-thought reasoning and multimodel ensemble classification with configurable consensus thresholds. cat-llm extract inductively discovers a parsimonious set of categories from a corpus without requiring the researcher to specify them in advance. cat-llm explore performs raw category extraction across repeated passes for saturation analysis. cat-llm summarize generates concise text summaries row by row. All commands operate directly on Stata string variables, returning results as new variables or r() macros.

The package supports all major LLM providers—OpenAI, Anthropic, Google, and Mistral—as well as locally hosted models via Ollama. A single cat-llm setup command handles all installation. Model selection, temperature, prompting strategy, and ensemble configuration are all controlled through standard Stata options syntax. Default parameter settings are informed by systematic empirical validation against human-coded survey data, striving for maximal accuracy and efficiency within the package framework. The package is freely available and installable from GitHub.