Identifying outliers using as little code

[Greener]

Hi there,

I'm trying to create a dummy variable that takes the value of 1 for outliers in GDP growth from 1990-2000 by using the 1.5*IQR rule in at least code as possible. Other than having a very complex set of code I am also having problems with telling stata to replace 0 with 1 when the value >x And <y. Also I have missing data which is being replaced by 1 when using this type of code.

Thank you in advanced,

Daniel Greener

[Maarten Buis]

There can be a trade-off between the amount of lines and how readable your code is. You can shorten the code below, but that will probably reduce the readability of it.

sysuse nlsw88, clear

// find the bounds
sum hours, detail
local p25 = r(p25)
local p75 = r(p75)
local iqr = `p75'-`p25'
local lb = `p25' - 1.5*`iqr'
local ub = `p75' + 1.5*`iqr'

// make the variable
gen byte outlier = !inrange(hours,`lb',`ub') if !missing(hours)

[Nick Cox]

Daniel does not show code so his difficulties defy detailed comment. But a criterion for outliers cannot here be being above some threshold and also below a lower threshold. Perhaps the code needs | not &.