Epistemic institutional sclerosis and the middle-income trap: evidence from China’s intellectual capital imbalance
Description
The dataset is a cross-country panel-style comparative dataset covering 62 countries and 58 academic disciplines, aggregated into three epistemic domains—sciences, social sciences, and creative arts—and then further collapsed into two broad categories, sciences versus arts, to examine how national intellectual capital composition relates to development outcomes. The main independent variables are per-capita measures of discipline-based publication output, alongside country dummy variables for China and the USA relative to the rest of the world, while the dependent variables are GDP per capita as a measure of economic quantity and anxiety prevalence (reversed in some models as a proxy for mental well-being) as a measure of economic quality. The study also reports strong internal consistency for these knowledge domains, with Cronbach’s alpha of 0.93 for sciences, 0.97 for social sciences, 0.91 for creative arts, and 0.99 for the combined arts category, supporting the aggregation strategy. The data are analysed using GLM with robust standard errors and multivariate regression to compare whether China and the USA differ from the world average in the extent to which their growth and mental-health outcomes are associated with science-heavy or arts-inclusive intellectual capital structures.
Files
Steps to reproduce
To reproduce the data, first define the unit of analysis as the country, with a sample of 62 countries, and collect publication-output data across 58 disciplines as the core measure of national intellectual capital. Next, classify those disciplines into three groups—sciences, social sciences, and arts/humanities—and test whether each group is internally coherent using Cronbach’s alpha; in the paper these are reported as 0.93 for sciences, 0.97 for social sciences, and 0.91 for arts/humanities. Then merge social sciences plus arts/humanities into a single broader arts category and confirm the combined reliability (alpha = 0.99), producing the two final epistemic variables used in the main analysis: sciences and arts, both measured on a per-capita basis. After that, assemble the country-level outcome variables by collecting GDP per capita as the economic-quantity indicator and anxiety prevalence per 100,000 people as the economic-quality or mental-well-being indicator, with the paper also noting a combined trade-off measure computed as log GDP per capita minus scaled anxiety. Finally, create dummy variables for China and the USA, leaving the rest of the world as the reference category, and merge all variables into one country-level dataset containing: country identifier, China dummy, USA dummy, sciences per capita, arts per capita, GDP per capita, anxiety prevalence, and any transformed versions such as logged GDP or reversed anxiety used in robustness checks.
Institutions
- Liaoning UniversityLiaoning, Shenyang