I have panel data with countries and daily data on covid-19 outcomes at the country level (such as covid-19 cases, deaths due to covid, hospitalization, etc...). I am trying to do a preliminary analysis to see if the political system of a country (federal vs unitary states) leads to worse and better covid outcomes. Here the country political system is identified by the variable country_system = 1 if the country is a federal state, 0 otherwise. This is a data sample:
Code:
* Example generated by -dataex-. To install: ssc install dataex clear input str32 country str10 date double(total_cases_per_million new_deaths_per_million hosp_patients_per_million) float country_system "France" "2020-03-04" 4.226 0 . 0 "Australia" "2020-03-04" 2.039 .039 . 1 "United States" "2020-03-04" .323 .012 . 1 "Netherlands" "2020-03-05" 4.786 . 3.793 0 "Italy" "2020-03-05" 63.809 .678 35.411 0 "Belgium" "2020-03-05" 4.314 . . 1 "India" "2020-03-05" .022 . . 1 "South Africa" "2020-03-05" .017 . . 1 "South Korea" "2020-03-05" 118.746 0 . 0 "Luxembourg" "2020-03-05" 1.598 . . 0 "Brazil" "2020-03-05" .019 . . 1 "Japan" "2020-03-05" 2.886 0 . 0 "Israel" "2020-03-05" 2.426 . 1.502 0 "Greece" "2020-03-05" 2.974 . . 0 "Spain" "2020-03-05" 5.54 .021 . 1 "United States" "2020-03-05" .556 .003 . 1 "China" "2020-03-05" 55.955 .022 . 0 "Lithuania" "2020-03-05" .367 . . 0 "Russia" "2020-03-05" .027 . . 1 "Saudi Arabia" "2020-03-05" .144 . . 0 "Chile" "2020-03-05" .68 . . 0 "Germany" "2020-03-05" 5.753 . . 1 "New Zealand" "2020-03-05" .622 . . 0 "Denmark" "2020-03-05" 1.899 . . 0 "Portugal" "2020-03-05" .785 . 0 0 "Finland" "2020-03-05" 2.166 . . 0 "Canada" "2020-03-05" .98 . . 1 "Austria" "2020-03-05" 4.552 . . 1 "Latvia" "2020-03-05" .53 . . 0 "Mexico" "2020-03-05" .039 . . 1 "Estonia" "2020-03-05" 2.262 . .754 0 "Indonesia" "2020-03-05" .007 . . 0 "Iceland" "2020-03-05" 99.634 . 2.93 0 "Switzerland" "2020-03-05" 13.172 .116 . 1 "Hungary" "2020-03-05" .207 . . 0 "Slovenia" "2020-03-05" .962 . . 0 "United Kingdom" "2020-03-05" 4.346 . . 0 "Poland" "2020-03-05" .026 . . 0 "France" "2020-03-05" 6.251 .044 . 0 "Sweden" "2020-03-05" 8.614 . .594 0 "Norway" "2020-03-05" 16.048 . . 0 "Argentina" "2020-03-05" .022 . . 1 "Ireland" "2020-03-05" 1.215 . . 0 "Australia" "2020-03-05" 2.157 0 . 1 "Czechia" "2020-03-05" 1.121 . 0 0 "Israel" "2020-03-06" 4.275 . 1.849 0 "New Zealand" "2020-03-06" .829 . . 0 "Australia" "2020-03-06" 2.353 0 . 1 "Ireland" "2020-03-06" 3.645 . . 0 "Japan" "2020-03-06" 3.321 0 . 0 "China" "2020-03-06" 56.061 .02 . 0 end
At this stage of the research, I would like to investigate the correlation between the covid-19 outcomes and the dummy variable country_system. I saw in previous posts that - pwcorr - is an excellent command to do that. However, it is not clear to me how to see the correlation differentiated for the two dummy variable conditions (namely country_system == 1 and country_system ==0) in a separate way, but in the same table. In case it is not possible to distinguish the correlation for country_system ==1 and ==0 separately, I would ask how to interpret the correlation coefficients with respect to the dummy variable.
Thank you in advance for your time
Best regards
Alessio Lombini
0 Response to Correlation between dummy and continuous variables
Post a Comment