Science, Arts, and National Mental Health: A Cross-National Panel Study of Intellectual Capital and Anxiety Prevalence
Description
The dataset supports the manuscript “Science, Arts, and National Mental Health: A Cross-National Panel Study of Intellectual Capital and Anxiety Prevalence.” It is a country-year panel dataset covering 149 countries over 26 years from 1995 to 2020, yielding approximately 3,874 observations. The dependent variable is national anxiety prevalence (NAP), measured as the population prevalence of anxiety and depression. The main explanatory variables capture national intellectual capital development in two domains: the sciences, measured through science, technology, engineering, and mathematics publication intensity per capita, and the arts, measured through social sciences, humanities, and creative arts publication intensity per capita. The dataset also includes their interaction term to examine whether arts-based intellectual capital moderates the association between science-based intellectual capital and national mental health. Additional variables include education attainment among the population aged 15 and above, GDP per capita, R&D expenditure, health expenditure, internet diffusion, population size, economic freedom, political corruption, military spending, individualism, and OECD/developing-economy status. The data were assembled from internationally comparable sources, including the World Health Organization, Clarivate Journal Citation Reports, the World Bank, SIPRI, and institutional/economic indicators, and were prepared for longitudinal mixed-effects, robustness, lagged-effect, and sensitivity analyses.
Files
Steps to reproduce
Steps to reproduce Compile the country-year panel dataset for 149 countries covering the period 1995–2020. Collect the dependent variable, national anxiety prevalence (NAP), from the World Health Organization mental health data. Standardize the measure across countries and years so that higher values indicate greater anxiety/depression prevalence. Construct the two main intellectual-capital variables using Clarivate Journal Citation Reports publication data: The sciences: publications per capita in science, technology, engineering, and mathematics fields. The arts: publications per capita in social sciences, humanities, and creative arts fields. Add the moderating variable by multiplying the sciences variable by the arts variable to create the sciences × arts interaction term. Add education and country-level controls, including education attainment among the population aged 15 and above, GDP per capita, R&D expenditure, health expenditure, internet diffusion, population size, economic freedom, political corruption, military spending, individualism, and OECD/developing-economy status. Transform and standardize variables where required. Publication variables and several country-level indicators should be log-transformed and rescaled to improve comparability across countries and reduce skewness. Estimate mixed-effects panel models with national anxiety prevalence as the dependent variable. Include country-level fixed effects and random effects to account for unobserved country heterogeneity and within-country change over time. Test the main hypotheses by estimating models for: the independent association between the sciences and NAP; the independent association between the arts and NAP; the interaction between the sciences and the arts. Conduct robustness checks, including lagged models for the sciences and the arts, sensitivity tests with simulated 10%, 20%, and 30% increases in science and arts productivity, and inter-model comparisons using mixed-effects and GLM specifications. Test nonlinearity by estimating models that include the squared interaction term to examine whether the arts have a nonlinear moderating effect on the relationship between the sciences and national anxiety prevalence. Generate figures and tables corresponding to the manuscript, including descriptive statistics, correlation matrix, main regression tables, lagged-effect tables, sensitivity analyses, nonlinearity tests, and coefficient plots. Compare the reproduced outputs with the reported results in the manuscript, focusing on the direction, magnitude, and significance of the sciences, arts, and sciences × arts interaction coefficients.
Institutions
- Liaoning UniversityLiaoning, Shenyang