My current method for identifying outliers only works for highest and lowest value by the use of command: sum age height weight
– what about the second highest value which might be an outlier as well for these three variables? Could I loop across all variables I could pull out the top/bottom five values for review?- what would be command for this?
I heard of 'foreach' command that can sort this. Any help please?
Try this link: https://www.google.com/search?ei=Ij9UXurlFs2gtQX046SYDw&q=how+to+detect+outliers+in+stata&oq=Outlier+detection+in+Stata&gs_l=psy-ab.1.2.0j0i22i30j0i333l3.470089.486375..492613...0.2..3.220.4343.41j5j1......0....1..gws-wiz.....6..0i71j0i22i10i30j0i362i308i154i357j0i273j0i131j0i67j0i131i67.A30tuY7KM_s
Best, D. Booth BTW my fav approach is in the attached. some of these ideas are implemented in Stata, see: https://www.google.com/search?ei=EEFUXtfeGZTPtQa96YuIBg&q=doing+robust+regression+in+stata&oq=Robust+Regression+in+Stat&gs_l=psy-ab.1.9.0j0i22i30l9.311310.325099..333816...0.2..3.158.4131.42j6......0....1..gws-wiz.....6..0i71j0i67j0i333j0i362i308i154i357j0i273j0i131j0i22i10i30.CVS_NMwwhPE
try sorting the variable. This code gets you to smallest and biggest six. The trick is that _N is a system variable which is the number of observations in the dataset (or by-group if you combine it with by) and _n is the number of the current observation.
sysuse auto
sort mpg
gen extreme_mpg = abs((_N-_n)) < 6 | _n < 7
- Badges
- Science topic