In this paper we describe the Subnational Human Development Database. This database contains for the period 1990–2017 for 1625 regions within 161 countries the national and subnational values of the Subnational Human Development Index (SHDI), for the three dimension indices on the basis of which the SHDI is constructed – education, health and standard of living --, and for the four indicators needed to create the dimension indices -- expected years of schooling, mean years of schooling, life expectancy and gross national income per capita. The subnational values of the four indicators were computed using data from statistical offices and from the Area Database of the Global Data Lab, which contains indicators aggregated from household surveys and census datasets. Values for missing years were estimated by interpolation and extrapolation from real data. By normalizing the population-weighted averages of the indicators to their national levels in the UNDP-HDI database, values of the SHDI and its dimension indices were obtained that at national level equal their official versions of the UNDP.

To gain insight into the duration and quality of the scientific peer review process, we analyzed data from 3500 review experiences submitted by authors to the SciRev.sc website. Aspects studied are duration of the first review round, total review duration, immediate rejection time, the number, quality, and difficulty of referee reports, the time it takes authors to revise and resubmit their manuscript, and overall quality of the experience. We find clear differences in these aspects between scientific fields, with Medicine, Public health, and Natural sciences showing the shortest durations and Mathematics and Computer sciences, Social sciences, Economics and Business, and Humanities the longest. One-third of journals take more than 2 weeks for an immediate (desk) rejection and one sixth even more than 4 weeks. This suggests that besides the time reviewers take, inefficient editorial processes also play an important role. As might be expected, shorter peer review processes and those of accepted papers are rated more positively by authors. More surprising is that peer review processes in the fields linked to long processes are rated highest and those in the fields linked to short processes lowest. Hence authors’ satisfaction is apparently influenced by their expectations regarding what is common in their field. Qualitative information provided by the authors indicates that editors can enhance author satisfaction by taking an independent position vis-à-vis reviewers and by communicating well with authors.

Until now, little was known about the variation in incidence of twin births across developing countries, because national representative data was lacking. This study provides the first comprehensive overview of national twinning rates across the developing world on the basis of reliable survey data. Data on incidence of twinning was extracted from birth histories of women aged 15–49 interviewed in 150 Demographic and Health Surveys, held between 1987 and 2010 in 75 low and middle income countries. During the interview, information on all live births experienced by the women was recorded, including whether it was a singleton or multiple birth. Information was available for 2.47 million births experienced by 1.38 million women in a period of ten years before the interview. Twinning incidence was measured as the number of twin births per thousand births. Data for China were computed on the basis of published figures from the 1990 census. Both natural and age-standardized twinning rates are presented. The very low natural twinning rates of 6–9 per thousand births previously observed in some East Asian countries turn out to be the dominant pattern in the whole South and South-East Asian region. Very high twinning rates of above 18 per thousand are not restricted to Nigeria (until now seen as the world's twinning champion) but found in most Central-African countries. Twinning rates in Latin America turn out to be as low as those in Asia. Changes over time are small and not in a specific direction. We provide the most complete and comparable overview of twinning rates across the developing world currently possible.

STUDY QUESTION: How many twins are born in human populations and how has this changed over recent decades? SUMMARY ANSWER: Since the 1980s, the global twinning rate has increased by a third, from 9.1 to 12.0 twin deliveries per 1000 deliveries, to about 1.6 million twin pairs each year. WHAT IS KNOWN ALREADY: It was already known that in the 1980s natural twinning rates were low in (East) Asia and South America, at an intermediate level in Europe and North America, and high in many African countries. It was also known that in recent decades, twinning rates have been increasing in the wealthier parts of our world as a result of the rise in medically assisted reproduction (MAR) and delayed childbearing. STUDY DESIGN, SIZE, DURATION: We have brought together all information on national twinning rates available from statistical offices, demographic research institutes, individual survey data and the medical literature for the 1980–1985 and the 2010–2015 periods. PARTICIPANTS/MATERIALS, SETTING, METHODS: For 165 countries, covering over 99% of the global population, we were able to collect or estimate twinning rates for the 2010–2015 period. For 112 countries, we were also able to obtain twinning rates for 1980–1985. MAIN RESULTS AND THE ROLE OF CHANCE: Substantial increases in twinning rates were observed in many countries in Europe, North America and Asia. For 74 out of 112 countries the increase was more than 10%. Africa is still the continent with highest twinning rates, but Europe, North America and Oceania are catching up rapidly. Asia and Africa are currently home to 80% of all twin deliveries in the world. LIMITATIONS, REASONS FOR CAUTION: For some countries, data were derived from reports and papers based on hospital registrations which are less representative for the country as a whole than data based on public administrations and national surveys. WIDER IMPLICATIONS OF THE FINDINGS: The absolute and relative number of twins for the world as a whole is peaking at an unprecedented level. An important reason for this is the tremendous increase in medically assisted reproduction in recent decades. This is highly relevant, as twin deliveries are associated with higher infant and child mortality rates and increased complications for mother and child during pregnancy and during and after delivery. STUDY FUNDING/COMPETING INTEREST(S): The contribution of CM was partially supported by the European Research Council (ERC) under the European Union's Horizon 2020 Research and Innovation Programme (grant No 681546, FAMSIZEMATTERS), Nuffield College, and the Leverhulme Trust. The contribution of GP was partially supported by the French Agence Nationale de la Recherche (grant No ANR-18-CE36-0007-07). The authors declare no conflict of interest. TRIAL REGISTRATION NUMBER: N/A.