Hasil untuk "Demography. Population. Vital events"

Uswatun Miftahul Jannah, Maria Angelina Sarventa Melani Dawi, Riztha Ferryanthi Maulinda et al.

This literature study examines the development, challenges, and opportunities of implementing Indonesia’s Digital Population Identity (in Indonesian: Identitas Kependudukan Digital/IKD) and its potential contribution to strengthening the national demographic statistics system. Drawing on national and international publications, government reports, and empirical findings from regional case studies, the review highlights that IKD represents a significant step in modernizing population administration through real-time data updates, system integration, and improved accuracy of civil registration. However, implementation remains constrained by uneven digital infrastructure, low digital literacy, varying administrative capacity, and data security concerns. Comparative insights from Estonia, India, Singapore, and South Korea show that Indonesia faces challenges similar to global digital ID initiatives, particularly regarding interoperability, privacy protection, and governance. Despite these obstacles, the literature indicates substantial opportunities for IKD to enhance demographic statistics through improved accuracy of vital event recording, cross-agency data integration, and the development of big-data-driven demographic analysis. Strengthening infrastructure, capacity building, and inter-agency collaboration will be essential for maximizing the role of IKD in supporting a modern, efficient, and integrated demographic statistical system.

Talita Zupo, D. F. Escobar, G. S. Santos et al.

Demographic compensation occurs when reductions in some vital rates are offset by increases in others, allowing populations to maintain similar performance across varying environments. This mechanism may help explain species' ecological distributions and range limits, yet its role at microenvironmental scales remains poorly understood. We investigated demographic compensation in Ipomoea cavalcantei, a narrow‐range but locally abundant species endemic to Amazonian ironstone outcrops, by comparing populations in two contrasting habitats: open‐ and shrubby‐canga. Using 3 yr of demographic data, we built matrix population models and conducted a life table response experiment. We also carried out germination and seedling establishment experiments under different temperature and light conditions simulating both habitats to identify the potential environmental drivers and their effects on key life‐cycle events. Despite contrasting environmental conditions, both populations exhibited similar population growth rates (λ), with opposing contributions of growth and fecundity – evidence of demographic compensation. The open‐canga population had lower growth but higher recruitment, driven by favorable temperature regimes for seed dormancy release and germination. Reduced growth was associated with physiological stress under high irradiance and shallow soils. Our results show that demographic compensation allows I. cavalcantei to persist across microhabitats, highlighting the importance of fine‐scale environmental heterogeneity in shaping species distributions.

Barnali Chakraborty, Priyanka Dey

J. Pavía, J. Lledó

The big data revolution has greatly expanded the availability of microdata on vital statistics, providing researchers with unprecedented access to large and complex datasets on birth, death, migration, and population, sometimes even including exact dates of demographic events. This has led to the development of a novel methodology for estimating sub-annual life tables that offers new opportunities for the insurance industry, also potentially impacting on the management of pension funds and social security systems. This paper introduces the qlifetable package, an R implementation of this methodology. It begins by detailing how basic summary statistics are computed by the package from detailed individual records, including the length of age years, which should be observed as relative (subjective) to ensure congruency between age and calendar time when measuring exposure times and exact ages of individuals at events. This is a new result that compels the observation of time as relative in the disciplines of actuarial science, risk management and demography. Afterwards, the paper demonstrates the use of the package, which integrates a set of functions for estimating crude quarterly (and annual) death rates, calculating seasonal-ageing indexes (SAIs) and building quarterly life tables for a (general or insured) population by exploiting either microdata of dates of births and events or summary statistics.

Sam Fallick, Josh Dixon, Silvia Colonna

This report investigates the number of individuals who successfully transition into treatment services after a Community Sentence treatment Requirement (CSTR) for an alcohol treatment requirement or a drug rehabilitation requirement in Wales. It also examines the length of time taken to transition and the outcome of treatment journeys. The cohort consists of individuals sentenced to a CSTR between 2014 and 2019. Data linking was used to identify these individuals’ substance misuse treatment journey. Multilevel logistic regression models investigate an individuals’ transition into treatment services and their outcomes. The role played by socio-demographic and offender related characteristics are also considered. Odds ratios show the importance of various characteristics on the likelihood of a successful transition from probation to treatment services and their outcomes. Time between the date of the CSTR and engagement with treatment services is analysed to ascertain the proportion of individuals entering treatment at different time intervals. The study shows the percentage of individuals sentenced to CSTRs in Wales who transitioned into treatment services. This includes the proportion of individuals who were either already in treatment on the date the CSTR was given or engaged with treatment within three weeks as well as how this varied between alcohol treatment requirements and drug rehabilitation requirements. It also reports the percentage of successful treatment outcomes and the reasons for non-successful outcomes. Finally, it identifies the impact of individual socio-demographic and offending- related characteristics on transitioning into treatment and successful completion rates such as age, gender and main problem substance. The results are preliminary as the study is unpublished, but it will be in time for the conference, so detailed findings can be shared then. The analysis included in this report provides detailed evidence on the pathways between probation and treatment services for CSTRs in Wales and an overview of the following treatment journey. Considerations are made on aspects that would have the biggest impact on improving service delivery.

Romana Burgess, Andy Boyd, Oliver Davis et al.

Background Linking digital footprint data into longitudinal population studies (LPS) presents an opportunity to enrich our understanding of how digitally captured behaviours relate to health traits and disease. However, this linkage introduces significant methodological challenges that require systematic exploration. Objectives To develop a robust framework for successful digital footprint linkage into LPS, informed by discussions from a workshop from the Digital Footprints Conference 2024. Methods We propose a structured, four-stage framework to facilitate successful linkage of digital footprint data into LPS: (1) understand participant expectations and acceptability; (2) collect and link the data; (3) evaluate properties of the data; and (4) ensure secure and ethical access for research. This framework addresses the key methodological challenges identified at each stage, discussed through the lens of two LPS case studies: the Avon Longitudinal Study of Parents and Children and Generation Scotland. Results Key methodological challenges identified include privacy and confidentiality concerns,\textbf{ }reliance on third-party platforms, data quality issues like missing data and measurement error. We also emphasize the role of trusted research environments and synthetic datasets in enabling secure, privacy-sensitive data sharing for research. Conclusions While the linkage digital footprint data to LPS remains in early stages, our framework provides a methodological foundation for overcoming current challenges. Through iterative refinement of these methods there is significant potential to advance population-level insights into health and wellbeing.

Nick Parr

Abstract The Total Fertility Rate (TFR) is the below replacement level for a population without migration throughout Europe. The population growth implications of low fertility combined with non-zero migration remain widely misunderstood. This paper proposes new measures which may enhance understanding of the relationships between TFRs and population growth for open populations. For 22 European countries in 2019, the method adjusts the familiar ‘typically just below 2.1’ replacement level for effects of constant non-zero immigration counts and emigration rates. The long-run perspective on zero growth the ‘Migration-Adjusted Replacement TFR’ provides is supplemented with near-term perspective by also presenting the TFR that would produce zero population growth immediately. The Migration-Adjusted Replacement TFR for 2019 ranges between 0.86 for Spain and 2.44 for Croatia. The variation is associated with the differences in migration between these countries. Its value is below 2.1 in 18 countries. For nine countries, the 2019 TFR is above the Migration-Adjusted Replacement level. The ‘Immediate Population Replacement TFR’ ranges from 0.26 for Sweden to 2.83 for Bulgaria, and for most countries lies below the Migration-Adjusted Replacement TFR. For most of the European countries, the TFRs that are coherent with zero population growth, immediately and long run, are below 2.1. A major advantage of this paper’s version of the ‘Migration-Adjusted Replacement TFR’ is its applicability to contexts with negative current net migration. The new measures proposed in this paper can better guide assessment of the relationships between fertility levels and population growth for European and other countries with non-zero immigration and emigration.

Abdul Wahab, C. Indriani, L. Lazuardi et al.

Determining the cause of death (CoD) is crucial for effective health policy and decision-making, particularly in population health programs. The World Health Organization (WHO) developed the Verbal Autopsy (VA) tool to ascertain CoD through verbal information, particularly in countries lacking comprehensive vital registration systems. In Indonesia, the health landscape is shifting from communicable diseases to non-communicable diseases (NCDs), highlighting the need for updated mortality surveillance. This study aimed to determine the major causes of death among adult mortality in Sleman Health and Demographic Surveillance System (HDSS) Indonesia. A demographic surveillance system was employed to monitor vital events, including mortality, with verbal autopsy interviews conducted for each death. A total of 279 adult deaths (ages 15 years and older) from the first two cycles of surveillance were analyzed. Trained enumerators conducted verbal autopsy interviews with informants close to the deceased. The InterVA program was used to process the verbal autopsy data, identifying the CoD for 274 adult deaths. Descriptive analysis was performed to determine the proportion of each cause of death, and Chi-square tests were used to assess differences in proportions. The findings revealed that 68.2% (95% CI: 62.38–73.72) of deaths were due to NCDs, including stroke, heart disease, diabetes, asthma, and chronic liver disease. Infectious diseases accounted for 24.8% (95% CI: 19.82–30.37), while injuries (primarily accidents) contributed to 6.6% (95% CI: 3.94–10.18). Stroke was the leading cause of death, particularly in individuals aged 50–64 years (21.2%, 95% CI: 16.48–26.49), followed by acute respiratory infections, including pneumonia (10.6%, 95% CI: 7.2–14.85). The study concluded that NCDs, particularly stroke, are the leading causes of adult mortality in Sleman HDSS, with significant contributions from acute respiratory infections and injuries. It is recommended for future research to further develop Verbal Autopsy technology, such as AI-based applications that can improve the accuracy of determining the cause of death.

Lee Williamson

ObjectiveThe Scottish Longitudinal Study (SLS) is a largescale research ready record-linkage study created and supported by the SLS Development and Support Unit (SLS-DSU). It links Census through time 1991-2022 to administrative data on major life events, maps changing residential location and for children, their progress through the educational system. MethodsThis paper will introduce the SLS as a data resource for researchers, the datasets held as part of it, along with the application process for using it. Census data are the building blocks of the SLS from 1991 onwards, for a 5% representative sample of the Scottish population (about 270,000 sample members each Census). The SLS links together a wealth of information from routinely collected administrative data, including vital events registrations (births, deaths and marriages), migration data, Scottish education data, and with appropriate additional permissions can be linked to NHS health data including cancer registry and hospital admission data. ResultsThe size and scope of the SLS make it an unparalleled research resource in Scotland for analysing a range of socio‐economic, demographic and health questions. Additionally, the longitudinal nature of the SLS is particularly valuable, allowing an exploration of causality in a way that cross‐sectional data collected at a single point in time does not. In this way, the SLS can provide insights into the health and social status of the Scottish population and, crucially, how it changes over time. The Scottish Census was a year behind the rest of the UK, with the 2022 Scottish Census data to be linked by SLS summer 2025. The 2025 ADR Conference is an excellent opportunity to showcase the 2022 Census data linkage results. ConclusionsThe paper will report on preliminary 2022 Census linkage results and SLS-DSU plans to extend research user access arrangements outside of Scotland. As part of the wider UK Census Longitudinal Studies (UKcenLS), the SLS aligns with the ONS-LS for England and Wales and the Northern Ireland LS, facilitating cross-UK analysis.

Girmatsion Fisseha Abreha, S. Fadel, Erica Di Ruggiero et al.

In Ethiopia, the reduction in perinatal mortality rates is still falling short of national and global targets set for 2030. Additionally, accurate recording is challenging, as many births occur at home. This study aimed to assess the trends and determinants of perinatal mortality using population-based longitudinal data from 2009 to 2016 across three Health and Demographic Surveillance Systems (HDSS) in Ethiopia: Gelgel-Gibe, Dabat, and Kilite-Awlaelo. Data on vital events and pregnancies were continuously collected at these HDSS sites. The study utilized follow-up data from prospective linked pregnancy and birth cohorts from January 2009 to December 31, 2016. Perinatal mortality was defined as deaths occurring from 28 weeks of gestation until six days after birth, measured per 1000 live births. Relevant health, demographic, and socioeconomic data were included in the analysis. Poisson regression was employed to assess factors associated with perinatal mortality. Out of 38,691 pregnancies that led to births, there were 1214 perinatal deaths (456 stillbirths and 758 early neonatal deaths), resulting in a perinatal mortality rate of 31 deaths per 1000 total births. The early neonatal death rate was higher, at 19.6 deaths per 1000 total births, compared to the stillbirth rate of 11.8 per 1000 total births. The perinatal mortality rate declined from 40.6 in 2009 to 29.1 per 1000 total births in 2016, reflecting an average annual rate reduction of 2.4%. Determinants of perinatal mortality included being a male newborn, multiple births, first-time pregnancies (primi-gravidity), lack of antenatal care visits, absence of delivery services, and residing in tropical zones. The primary causes of death were asphyxia, sepsis, and preterm birth. Overall, perinatal mortality rates were high in the three HDSS sites, with slow reductions over time and significant variations between them. Addressing the issue of stillbirths and improving the availability and quality of emergency obstetric care are crucial. Continuous home visits in rural communities to prevent stillbirths and newborn deaths, are also essential.

A. Amberbir, M. Boah, M. Semakula et al.

INTRODUCTION: This protocol outlines the establishment and implementation of a Human Development and Demographic Surveillance System (HD2SS) in Butaro, Rwanda. The HD2SS will facilitate prospective, continuous monitoring of the population, tracking vital statistics, social events, and key health and demographic indicators in a defined population. The system will enable accurate and validated assessment of the impact of health and related population-level interventions, supporting evidence-based decision-making and data-driven improvements in healthcare and socioeconomic outcomes at a population level. METHODS: The HD2SS will be implemented in the Butaro sector, home to 38,013 individuals across 68 villages. The Butaro sector was purposively selected due to the existence of the University of Global Health Equity and Partners In Health-supported Butaro healthcare delivery. Data, including location, demographic, socioeconomic, and health-related variables, will be collected using the annual household census and stored using the Survey Solutions system for real-time electronic data capture, ensuring data quality, security, and confidentiality. Data analysis will enable the identification of emerging trends, the development of interventions, and the evaluation of related policies and programs. CONCLUSION: The HD2SS will provide currently limited but much-needed data to inform improvements in public health programming and socioeconomic development and strengthen local health research capacity. Regular dissemination of findings will ensure stakeholders, including local health authorities and development partners, are informed and able to use the results to improve health and social development outcomes in Rwanda.

Merlène Saunier, C. Barbraud, Maxime Amy et al.

Understanding demographic processes is crucial in ecology and conservation biology to assess how populations respond to environmental pressures. Although demographic parameters often react to disturbances, populations may remain in unstable age structures (transient dynamics) following such events. To explore the influence of transient dynamics in masked booby and red‐footed booby populations on Tromelin Island 15 years after rat eradication, we compared observed, transient, and asymptotic growth using count and capture–recapture (CR) data. CR data were used to estimate survival probability from Cormack–Jolly–Seber models and other vital rates such as breeding frequency and age at first breeding. We coupled asymptotic growth with a density‐dependent model for breeding site availability to project long‐term population growth. Both populations exhibited high adult survival (0.91–0.96) and breeding success (0.46–0.54). The masked booby juvenile survival was also high (0.86). Populations are expected to grow until breeding sites are saturated, in about a century. Comparison of asymptotic and transient growth suggests that population growth is linked to an intrinsic process and that both populations have undergone disruption in their age structure likely due to past predation by rats. Although the observed population growth appears to align more closely with transient growth, we could not entirely exclude the possibility that differences between asymptotic and observed growth rates may be due to data scarcity or age‐related variation in demographic parameters. Nevertheless, the study of transient dynamics allowed a better understanding of the differences observed between field counts and estimated asymptotic growth rates based on matrix population models.

Thi Phuong Anh Phan, Thi Minh Tuan Dang, Minh Hiệp Ngô et al.

Introduction. The Mekong Delta of Vietnam (VMD) is a key agricultural region vital for national food security. In recent years, the region has faced significant outmigration driven by economic challenges and climate change impacts, such as droughts, saltwater intrusion, and extreme weather events. Goals. This study examines the relationship between climate change and migration in the VMD and its implications for sustainable regional development. Methods. Using a political-environmental perspective, the paper analyzes quantitative and qualitative data from the 2019 Population and Housing Census and the 2022 Mekong Delta Economic Annual Report. Migration patterns, demographic shifts, and regional development indicators are assessed to understand how environmental stressors influence human mobility. Results. Findings reveal that climate change is a major driver of migration reshaping population structures, labor supply, and the region’s long-term development capacity. Over 1.3 million people migrated from the VMD in the past decade, posing challenges to demographic stability and socio-economic resilience. Conclusions. The study recommends strengthening climate-adaptive sustainable development policies and enhancing local human resource capacity. These measures aim to stabilize migration, promote inclusive development, and ensure climate-resilient growth in the Mekong Delta.

E. McLean, Rebecca Sear, E. Slaymaker

Rachel McCarter, Michael Rosato, Gerard Leavey

Objective and Approach Using census returns from three waves of NILS (1991, 2001, 2011) linked with mortality data, we examined morbidity/mortality outcomes by socio-economic disadvantage over time (using an individual-level deprivation index derived from educational attainment, social class, housing tenure and household car availability at each census). These were combined to generate two indicators of Intercensal social mobility: measuring no change, and upward/downward mobility between censuses. The relationship between mobility (2001-2011) and (separately) self-reported mental ill-health (MIH) at the 2011 Census and all-cause mortality (2011-2015) were examined using logistic regression and Cox PH modelling respectively. Results   Population comprised 288,262 individuals aged 25-74 in 2011, recording 19,318 deaths (2011-2015) and 23,959 (8.3%) reporting MIH. MIH: those with no intercensal mobility followed the standard pathway associated with persisting disadvantage (comparing with the most advantaged the least advantaged recorded OR=16.13:95%CI=13.50,19.28); and while both the upwardly and downwardly mobile generally recorded higher ORs than those consistently most advantaged, the magnitude of ORs increased with social distance traversed; and ORs for downward mobility were consistently higher than with upward mobility. Patterns associated with all-cause mortality were similar. Conclusions This suggests: upward mobility retains something of the social patterning pertaining in social disadvantage levels left behind; downward mobility may be connected to ongoing health issues. Poor health outcomes remain strongly associated with socio-economic circumstance.  Implications   Research focusing on relationship between social mobility, disadvantage and health outcomes in NI is limited: access to more precise administrative data will enhance possibilities associated with evidence-based policy formation.

Xiaoyue Ni, O. Wei, Hyoyoung Jeong et al.

Significance Continuous measurements of health status can be used to guide the care of patients and to manage the spread of infectious diseases. Conventional monitoring systems cannot be deployed outside of hospital settings, and existing wearables cannot capture key respiratory biomarkers. This paper describes an automated wireless device and a data analysis approach that overcome these limitations, tailored for COVID-19 patients, frontline health care workers, and others at high risk. Vital signs and respiratory activity such as cough can reveal early signs of infection and quantitate responses to therapeutics. Long-term trials on COVID-19 patients in clinical and home settings demonstrate the translational value of this technology. Capabilities in continuous monitoring of key physiological parameters of disease have never been more important than in the context of the global COVID-19 pandemic. Soft, skin-mounted electronics that incorporate high-bandwidth, miniaturized motion sensors enable digital, wireless measurements of mechanoacoustic (MA) signatures of both core vital signs (heart rate, respiratory rate, and temperature) and underexplored biomarkers (coughing count) with high fidelity and immunity to ambient noises. This paper summarizes an effort that integrates such MA sensors with a cloud data infrastructure and a set of analytics approaches based on digital filtering and convolutional neural networks for monitoring of COVID-19 infections in sick and healthy individuals in the hospital and the home. Unique features are in quantitative measurements of coughing and other vocal events, as indicators of both disease and infectiousness. Systematic imaging studies demonstrate correlations between the time and intensity of coughing, speaking, and laughing and the total droplet production, as an approximate indicator of the probability for disease spread. The sensors, deployed on COVID-19 patients along with healthy controls in both inpatient and home settings, record coughing frequency and intensity continuously, along with a collection of other biometrics. The results indicate a decaying trend of coughing frequency and intensity through the course of disease recovery, but with wide variations across patient populations. The methodology creates opportunities to study patterns in biometrics across individuals and among different demographic groups.

M. Mezzavilla, Federico De Pizzol, L. Vallini et al.

The intersection of historiography and archaeology has long pondered over the impact of known historical events on census size. In recent times, genetic methods have successfully traced changes over time in the genetic size of a given population. Moreover, the correlation between genetic and census sizes of a population is contingent on several demographic assumptions that are relatively simple for our species. Our research endeavours to examine the changes in effective population size (Ne) in all human populations in the 1000 Genomes Project over the past two millennia. We compared our findings with estimates from historical censuses where available. Our investigation confirms what was already observed in France and reveals a common pattern found in most European populations, which manifests as a drastic population decrease beginning around the year 1300 and growth after the year 1600. This profile aligns well with known wars, famines, and epidemics that characterized these trying times in Europe. The most notable among them being the second plague epidemic, caused by Y. pestis, which in Europe commenced in 1347/8 and is also known as the "Black Death". Our findings demonstrate that changes in genetic population size through time can serve as a dependable proxy for census size, which is independent of potential biases in the written historical record. Consequently, we provide a robust estimate of the impact caused by the population crisis that followed the year 1300 on the European genomic landscape in light of previous results. Our study offers a new paradigm for interpreting the past and underscores the potential of genetic methods in reconstructing historical events.

Piotr Teodorowski, Elisa Jones, Saiqa Ahmed et al.

Objectives Public involvement and engagement have been suggested as a way to establish public support for big data research, yet there has been no review exploring how these activities could facilitate this. Therefore, this scoping review aimed to explore public involvement and engagement in big data research. Methods Following Arksey and O’Malley’s methodology, we systematically searched the following databases: CINAHL, Health Research Premium Collection, PubMed, Scopus and Web of Science for papers published between 2010-2021. Additional manual searches took place. These included the first 100 hits in Google search, journals (BMC Research Involvement and Engagement, International Journal of Population Data Science and Health Expectations) and grey literature (Patient Outcome Research Institute database, first 100 hits were screened). We extracted data using a standardised form. We then organised it in a descriptive and narrative way. A system logic model was developed to understand the complexity of this topic. Results Fifty-three papers were identified as eligible for inclusion in our review. The findings indicate that public involvement and engagement have the potential to improve public trust and accountability for data resharing for research. However, there is limited literature actually evaluating these activities. The findings suggest that the public can be meaningfully involved and engaged in big data research, both in terms of individual research projects and data governance, but there is no one standardised approach to do it. Therefore, we developed an initial system logic model to map relevant aspects of the involvement and engagement activities. These include which communities to reach, the context (e.g. ethical, legal aspects or public views), the design and delivery of activities, and outcomes. Conclusion Despite the growing literature on public involvement and engagement in big data research, more research is needed as there are few primary empirical studies exploring involvement and engagement. We suggest using the system logic model we developed when reflecting on issues that might be relevant in organising these activities.

Lu-hai Yu, Tingting Wang, Huidong Bai et al.

Background Acute coronary syndrome (ACS) has become a vital disease with high mortality in the Uygur populations. Clopidogrel plays an important role in reducing the risk of recurrent cardiovascular events after ACS; however, it is a prodrug that requires biotransformation by cytochrome P450 ( CYP450 ). Objectives To determine the effect of genetic polymorphisms in CYP2C19*2 , *3, and *17 , and along with clinical, demographic factors, on variation in response to clinical outcomes in Uygur patients. Methods A total of 351 patients with ACS were treated with clopidogrel and aspirin for at least 12 months; we recorded major adverse cardiovascular events (MACE) or bleeding within 1 year. Multivariable logistic regression analyses were carried out to identify factors associated with MACE or bleeding. Results We analyze risk factors include age, BMI (body mass index), smoking, alcohol intake, NSTEMI (non-ST-segment elevation myocardial infarction), hypertension, dyslipidemia, concomitant medication, CYP2C19*2 carriers, CYP2C19*17 carriers and metabolizer phenotype. CYP2C19*2 carriers had an odds of having MACE of 2.51 (95% CI: 1.534–4.09) compared with noncarriers ( P   0.05). Conclusion The CYP2C19*2 gene polymorphism contributes to the risk of MACE in dual clopidogrel—treated Uygur population with ACS with or without PCI (percutaneous coronary intervention). These data may provide valuable insights into the genetic polymorphisms affecting clopidogrel metabolism among minority groups in China.

Andrea Priulla, Nicoletta D’Angelo, Massimo Attanasio

Abstract This paper investigates gender differences in university performances in Science, Technology, Engineering and Mathematics (STEM) courses in Italy, proposing a novel application through the segmented regression models. The analysis concerns freshmen students enrolled at a 3-year STEM degree in Italian universities in the last decade, with a focus on the relationship between the number of university credits earned during the first year (a good predictor of the regularity of the career) and the probability of getting the bachelor degree within 4 years. Data is provided by the Italian Ministry of University and Research (MIUR). Our analysis confirms that first-year performance is strongly correlated to obtaining a degree within 4 years. Furthermore, our findings show that gender differences vary among STEM courses, in accordance with the care-oriented and technical-oriented dichotomy. Males outperform females in mathematics, physics, chemistry and computer science, while females are slightly better than males in biology. In engineering, female performance seems to follow the male stream. Finally, accounting for other important covariates regarding students, we point out the importance of high school background and students’ demographic characteristics.