Technology-Enabled Ecstasies Moderated by National Anxiety Prevalence: Digitalised Diffusion of the Global Psychedelics
Description
Abstract / Description This dataset supports a country-year panel study of 149 countries from 1996 to 2021 examining the relationship between internet diffusion, national anxiety prevalence (NAP), and ecstasy diffusion. The dependent variable is ecstasy prevalence per 100,000 population. The main predictors are internet diffusion and NAP, alongside economic, institutional, and sociocultural controls. Source data were compiled from WHO, the World Bank, Heritage Foundation, SIPRI, Oxford-based health expenditure sources, and Hofstede-based cultural indicators. Variables were harmonized into a country-year panel and transformed using a log–shift–expand procedure for estimation comparability.
Files
Steps to reproduce
Steps to reproduce This study uses a country-year panel covering 149 countries from 1996 to 2021. The unit of analysis is the country-year. Collect the raw annual country-level data for all variables used in the study. The main variables are ecstasy diffusion (national prevalence per 100,000 population), internet diffusion (internet use/penetration), and National Anxiety Prevalence (NAP). Also collect the control variables: economic freedom, healthcare spending (% GDP), R&D expenditure (% GDP), population, political corruption, military spending (% GDP), GDP per capita, literacy rate, and individualism–collectivism. Harmonize country identifiers and year values across all sources. Merge all series into one panel dataset with one observation per country-year for 1996–2021. Transform all continuous variables using the same procedure as in the manuscript: apply a log transformation (using a small constant where needed for zero values), shift the transformed series so the minimum equals zero, and multiply by 10. Descriptive statistics and model estimates are based on these transformed variables. Construct the interaction term between transformed internet diffusion and transformed NAP. Estimate fixed-effects panel models with country fixed effects, year fixed effects, and standard errors clustered at the country level. Run the baseline models progressively: internet diffusion only; NAP only; internet diffusion + NAP + interaction; then add the full set of controls. Compute marginal effects and predicted values from the interaction model to examine whether the association between internet diffusion and ecstasy diffusion becomes stronger at higher levels of NAP. Replicate the robustness checks reported in the study, including lag and lead specifications, sensitivity analyses, nonlinear specifications, and the instrumental-variable model. Regenerate the descriptive statistics, regression tables, and figures from the transformed analytical dataset and compare coefficient signs, significance patterns, and interaction results with those reported in the manuscript.
Institutions
- Liaoning UniversityLiaoning, Shenyang