how to make a descriptive statistics like the picture (including Jarque‐Bera test for normality) and output the result

[roy zhong]

"sum,de" did not include Jarque‐Bera test for normality, and how to use the panel data to achieve the picture result.

clear
input str9 province byte id int year double(netrate urban)
"shanghai" 1 1997 .34 71.64
"shanghai" 1 1998 .62 72.83
"shanghai" 1 1999 5.37 73.19
"shanghai" 1 2000 12.3 74.62
"shanghai" 1 2001 19.21 75.28
"jiangsu" 2 1997 .05 29.9
"jiangsu" 2 1998 .16 31.5
"jiangsu" 2 1999 .6 34.9
"jiangsu" 2 2000 1.67 41.5
"jiangsu" 2 2001 3.67 42.6
"zhejiang" 3 1997 .05 39.04
"zhejiang" 3 1998 .22 42.26
"zhejiang" 3 1999 .88 45.48
"zhejiang" 3 2000 3.24 48.7
"zhejiang" 3 2001 4.82 50.9
"jiangxi" 4 1997 .01 25.32
"jiangxi" 4 1998 .08 26.05
"jiangxi" 4 1999 .09 26.79
"jiangxi" 4 2000 1.12 27.69
"jiangxi" 4 2001 1.45 30.41
end
[/CODE]

[roy zhong]

o

[Sven-Kristjan Bormann]

Something along the following lines could work:

putexcel set some-file.xlsx, replace
putexcel A1="Province" B1="Variable" C1="Mean" D1="SD" E1="Skewness" F1="Kurtosis" G1="J-B"
mat results=J(8,5,.)
local count 0
levelsof province, local(prov) clean
forvalues i=1/4{    
    putexcel A`=`count'+2'="`=word("`prov'",`i')'"
    foreach var in netrate urban{
    putexcel B`=`count'+2'="`var'"
    local ++count
    qui sum `var' if id==`i',d
    mat results[`count',1]=r(mean)
    mat results[`count',2]=r(sd)
    mat results[`count',3]=r(skewness)
    mat results[`count',4]=r(kurtosis)
    sktest `var' if id==`i'
    mat results[`count',5]=r(chi2)
    }    
}
putexcel C2=matrix(results)

This code is just a quick solution and needs modification from your side to give you the desired result.

[roy zhong]

Something along the following lines could work:

putexcel set some-file.xlsx, replace
putexcel A1="Province" B1="Variable" C1="Mean" D1="SD" E1="Skewness" F1="Kurtosis" G1="J-B"
mat results=J(8,5,.)
local count 0
levelsof province, local(prov) clean
forvalues i=1/4{
putexcel A`=`count'+2'="`=word("`prov'",`i')'"
foreach var in netrate urban{
putexcel B`=`count'+2'="`var'"
local ++count
qui sum `var' if id==`i',d
mat results[`count',1]=r(mean)
mat results[`count',2]=r(sd)
mat results[`count',3]=r(skewness)
mat results[`count',4]=r(kurtosis)
sktest `var' if id==`i'
mat results[`count',5]=r(chi2)
}
}
putexcel C2=matrix(results)

This code is just a quick solution and needs modification from your side to give you the desired result.

thanks for your kind help