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: A Statalist glossary

From   Nick Cox <[email protected]>
To   [email protected]
Subject   st: A Statalist glossary
Date   Sat, 1 Jan 2011 10:56:04 +0000

Best wishes to Statalist members for 2011 and for your uses of Stata
to do good statistical science.

As new people are always joining the list, and most old and new
posters appear to be unwilling to read so stiff and structured a
document as the FAQ, I thought I would reissue this glossary from 2010
(with some revisions) as a more informal guide to what we (should) do
on this list.

Nick Cox

Re-reading the entire set of manuals for fun, as one does, it struck me
that StataCorp has been including more glossary sections over recent
versions. That seems like a good idea for this list, so here goes:

A is for advice. All advice, here and elsewhere, is based on one
over-arching principle: Whatever promotes clear, complete and correct
answers delivered quickly is good. (There's a converse....)

A is also for amusement. Don't assume from the stern and stiff opening
of this glossary that it is entirely composed of straight-faced

A is also for archives, as in "use the archives". But first read the
help, the manual and the FAQs.

A is also for ASCII, or plain text, which is expected on Statalist.

A is also for attachments, which should not be sent to Statalist.

B is for basics, which means CDE:

C is for code or commands used. Show us exact code. Do not merely say
that you used some command, or worse, not even specify what command
you used.

D is for data. Show us examples of your problem with datasets everyone
can use (see -help dta_contents-) or with small fake datasets. If you
can't do that, give us an example of your data.

E is for examples. A concrete example of your problem is worth a
thousand words of arm-waving explanation or speculation, and worth much,
much more than a report that something "didn't work", which could mean
about twenty different things.

F is for FAQ. You should read
<> before posting.
See also StataCorp's own FAQs: <>

f is also for -findit-. -findit- is your friend. (Don't know about
-findit- yet? Then do read the help.)

G is for Gould. See William.

G is also for Gutierrez. Bobby Gutierrez is Director of Statistics at
StataCorp. When he writes, pay attention!

g is also for -gllamm-. In Klingon that means "quick".

H is for help. Did you read the help carefully?

H is also for hyphens, as in references to some Stata -commandname-. The
hyphen convention is a Gould idea.

IVs is for instrumental variables, or sometimes independent variables.
If you didn't know that, this won't help (bad news), or you're not an
economist (good news).

J is for jokes. There are plenty of in-jokes on Statalist, but you
shouldn't expect me to explain them here. If someone's sense of humour
appears a bit weird to you, you are almost certainly correct. See also

K is for Kit Baum, as in "Thanks, as usual, to Kit Baum". He does much
of the real work around here, like maintaining SSC.

K is also for Kolenikov. Not to be confused with Kalashnikov. See also

L is for lousy subjects for your postings, such as "help" or "problem"
or something only a little more informative than that. Be specific!

L is also for lousy or lazy literature references, as in "I want to use
the test of Sue, Grabbit and Runne (1989)". Full details please!

M is for moderator, meaning Marcello Pagano, sine quo non and primus
inter pares. If he gets annoyed, you've been really bad. He also does
much of the real work around here.

M is also for Maarten Buis, who has been dispensing excellent advice
on this list for some years.
It came as a small shock to some in 2010 that all that time he had
been a graduate student yet
to finish his doctorate. Now he has it!

M is also for Manual, as in Read The Fine.

N is for nice, as in "be nice", and as in "even if people appear to be
nasty when you are a bit naughty, they are really nice underneath, and
are doing it in everyone's best interests".

N is also for Nick, which is probably a coincidence. See also J.

O is for "official", which means "whatever code StataCorp admit
responsibility for". (This is not a joke, unless it is.)

O is also for operating system, which you should make explicit if it's
relevant. Don't assume that the whole world uses Windows!

P is for pedantry. As Bertrand Russell almost said, a pedant is a person
who prefers to be correct. Not a dirty word on this list.

p is for pweights, problem, plague, pestilence, pain, and, more
fortunately, Pitblado.

Q is for questions. See BCDE again, or for the first time.

R is for R. No one's agin it (really!). Its value far exceeds its price.

R is also for re-posting your question. Disapproved.

S is for Stata, silly. By the way, some of us get a bit irritated if you
spell it STATA, which is wrong, or at the very least a couple of decades
out-of-date. See also P.

S also signals Some Alternative Software (originally Some Athenian

S is also for SMCL, which means SMCL Makes Cooler Logs.

S is also for survey statistics and Steve Samuels. Fancy that.

T is for technicalities. We love them.

U is for -update-, which keeps your Stata up-to-date (surprise). By
extension, see also -adoupdate-.

U is also for "user-written", as in "do explain where user-written software
you refer to comes from".

V is for version. If you are using an out-of-date version (especially 10.1 or
earlier), then say so. Being out-of-date is not a sin; but declaring it
increases your chance of being told of a solution you can use.

V is for Vice-Presidents. Alan Riley and Vince Wiggins are
Vice-Presidents of StataCorp. When they write, pay attention!

W is for William Gould, President of StataCorp. When he writes, pay
double attention! He probably throws away more good programs in a
typical week than anyone else writes in a good year.

x is for predictors or covariates. (Are you still saying independent

y is for response or outcome. (Are you still saying dependent variable?)

Z is the end of the English, meaning British, alphabet.
*   For searches and help try:

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