Bookmark and Share

Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

st: AW: RE: AW: AW: Categorize variable

From   "Martin Weiss" <>
To   <>
Subject   st: AW: RE: AW: AW: Categorize variable
Date   Mon, 12 Apr 2010 19:53:53 +0200


Re -irecode()-, you may also like


-----Ursprüngliche Nachricht-----
[] Im Auftrag von Nick Cox
Gesendet: Montag, 12. April 2010 19:51
Betreff: st: RE: AW: AW: Categorize variable

gen mygroups = 10 * ceil(weight/10) 

would get you most of the way. This has two small but notable

1. The values are transparent as bin limits, 170 meaning (160, 170] etc.

2. Very simple code, including no need for value labels. 

I realise that this will produce more classes than intended if there are
values below 140 or above 190, but there are solutions for that if it's
a real nuisance (e.g. -recode-). 

The larger question is why you want to do this, as it is throwing away
detail that might be useful. 

-irecode()- is another way to do it. 


Martin Weiss

BTW, my solution is a bit more complicated than initially imagined

set seed 123456
set obs 1000
gen weight=rnormal(165,6)
su weight, mean
egen float mygroups = cut(weight), at(`=r(min)-0.01' 150(10)180
la def mylab 0 "<150" 1 "150-160" 2 "160-170" 3 "170-180" 4 ">180"
la val mygroups mylab
table mygroups, contents(freq min weight max weight) mis

Martin Weiss

Look at -help egen, cut()-


I have a continuous variabile called weight. How can I categorize this 
variable into groups: <=150lb (150-160], (150-160], (170-180] and >180?

*   For searches and help try:

*   For searches and help try:

© Copyright 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index