Stata The Stata listserver
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

RE: st: RE: mode


From   "Nick Cox" <n.j.cox@durham.ac.uk>
To   <statalist@hsphsun2.harvard.edu>
Subject   RE: st: RE: mode
Date   Thu, 15 Apr 2004 18:47:02 +0100

I'm not clear on why you want the mode 
or what you want to do with it, so 
advice is difficult. Let's look 
at data accessible to all: 

. sysuse auto, clear 

Suppose you notice that -rep78- 
contains missing values and, faute de mieux, 
you decide to impute those by the mode. 
Then 

. egen mode = mode(rep78) 

. tab mode 

shows that the variable is a constant, 
the mode for all the values, so that if you were to 

. replace rep78 = mode if missing(rep78) 

that would copy the mode into missing values. 

But hang on: this is very crude. For example, 
a tabulation of -foreign- and -rep78- makes 
it clear that knowing -foreign- makes a difference
to the distribution of -rep78-. So a better guess
would be the mode of -rep78- given -foreign-. That 
is 

. egen mode2 = mode(rep78), by(foreign) 

except that, as trying it will show, you now have
to start thinking about appropriate choices. 

-mode()- was in the first instance written because
people asked for it, and -- just as saner people
practise the piano, or play card games, or make millions -- 
the author likes little programming problems. But I've never
found the mode enormously useful in practice... 

Nick 
n.j.cox@durham.ac.uk 

peter harper
 
> Thank you Nick for your help. I am a little uncertain
> as to what I do next. My variable is increase in wage
> settlements occuring each year. I have done 
> . egen modeinc = mode (incres)
> The variable modeinc appears on my variable list. I am
> not sure what I do next. I have tried various things
> but I do not appear to get the mode of this variable.
> Doing it manually is out of the question as I have an
> extremely large dataset.
> I would be be grateful, if you would let me know,
> references on how I obtain mode.
> Thanks in advance
> 
>  --- Nick Cox <n.j.cox@durham.ac.uk> wrote: > This
> isn't a simple question at all. If your 
> > variable is discrete, two or more values may tie
> > for mode. If your variable is continuous, how
> > to define the mode is an open issue. Nevertheless, 
> > -search mode- points to an -egen- function =mode()- 
> > and a user-written command -modes-. From your
> > question
> > the first may be closer to what you want. 
> > 
> > The Statalist FAQ contains advice on using -search- 
> > and other facilities. 
> > 
> > Nick 
> > n.j.cox@durham.ac.uk 
> > 
> > peter harper
> > 
> > > I wondered
> > > if anyone can tell me how do I obtain 'mode'(of
> > > course, it is term like mean, median etc.) of a
> > > variable for individual years. I have a dataset
> > for 21
> > > years.
> > 

*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/



© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index