Chinese

• Gov_data/CH_policy_zky
  • from here, manually collected: http://120.27.61.185:8081/NCOVSYS/index?MENU_ID=27&MENU_URL=template%2Fload%2F48

International

• need to put JHU data folder in the root folder

Evaluating Governmence by Country

• data issues
  • owid-covid-data.csv is from https://github.com/owid/covid-19-data/tree/master/public/data
  • JHU data

What question to ask?

• influence of type of government on controlling the virus
  • type of gov -- virus
• Focus on policy
  • how is effectiveness of policy affected by polity
  • Reference/Which Measures are Effective in Containing COVID-19?
• explore a three-layer model: type of gov --affect--> direction/execution of policy --affect--> virus transmission
  • get to know the coefficients -- how gov affect policy, how policy affect virus
  • hierarchical linear model? Just add another variable?
  • ML modelling?
• Or answer the question: is the form of government a good determinant?
  • Alternatives: state-society relation -- any quantification work available?
  • culture score
    • https://www.hofstede-insights.com/product/compare-countries/
    • https://geerthofstede.com/research-and-vsm/
    • power distance, individualism, masculinity, uncertainty avoidance, long term orientation, indulgence

Technical Issues

• choose proper variable to quanitfy virus severeity
  • Virus severity
    • avg. growth rate (current)
  • Policy**
    • pace of action -- time interval between first occurance and total lockdown (control)

Other

• Related researcher
  • Prof. Christopher Berry, Assoc. Prof. Anthony Fowler (field relevant!)
    • UChicago news article
    • (a US version of measuring governmental response )
  • Dali Yang & Regina?
  • NYU Shanghai -- Prof Jenq
• Peter Meeting 1
  • log regression -- growth rate
  • maybe try exclusion criteria for days to calculate growth rate (num of days between 10 - 1000)
  • Diff-in-diff (utilize panel data)

Dali Yang update!

Prof Jenq May 22

May 22

• underreporting / measurement error
• effect is in terms of counterfactual -- measure effect in terms of that
  • Book: (undergrad) Mastering metrics: the path from cause to effect, Josh Angrist, et al
  • (grad) mostly harmless econometrics
• find a research design to approximate the experiment to get the causal effect
  • treatment: 'democracy'
  • control should account for the confouding factors: those affecting virus except democracy
  • natrual experiment appraoch
  • diff-n-diff: similar countries that differ in democracy
  • other treatment variable: public health? Taiwan/S. Korea vs. US UK UCLA '' https://stats.idre.ucla.edu/stata/seminars/stata-survival/

Synthetic control

• whether counterfactual
• history of that counterfactual -- look at history
• matching based on important characteristics of the country -- use other control variables
• ability to enforce lockdown (might )
  • Taiwan? no lockdown but did well
  • Japan
• cultural score

dependent variable

• effectiveness: change in line
• manually construct: smoothing, limit it to countries with more data
• messy data: don't have
• develop an algorithm how to select data
• number of death, num of confirmed cases

literature review

• mass media articles -- collective/individualistic articles
• political science literature: how effective government reacts