Difference of the commands "mean" and "center"

[Silvia Haertl]

Dear all,

could someone maybe help me with this:

What is the difference between the command

mean varlist [if] [in] [weight] [, options]

and the command

center varlist [weight] [if exp] [in range]

?

[Rich Goldstein]

-mean- is an estimation command while the user-written -center- (as the FAQ explains, you should tell us where you got this), is for generating new variables that are the centered (or standardized, optionally) version of old variables - since the help files would tell you this, I think I may be missing something in your question; if I did, please clarify

added: forgot mention that I am referring to Ben Jann's -center- which is at SSC

[Silvia Haertl]

Dear Rich:

Thank you for your answer! Sorry, I haven´t clarified where I got the -center- command from.
I was also referring to Ben Jann´s -center-

Jann, B. (2004). center: Stata module to center (or standardize)
variables. Available from
http://ideas.repec.org/c/boc/bocode/s444102.html.

I think I just didn´t get it right.

If I have for example the following data:
'lrgdp' is the log of real GDP per capita
and I want to center the variable 'lrgdp', would that be the same as to sum up all the levels of 'lrgdp' and then divide them by the number of the variables (here: 20) ?

* Example generated by -dataex-. To install: ssc install dataex
clear
input int year str11 country float lrgdp
1870 "Australia"  2.627285
1871 "Australia" 2.6345375
1872 "Australia"  2.710996
1873 "Australia"  2.786211
1874 "Australia"  2.789214
1875 "Australia"   2.86745
1876 "Australia"  2.834332
1877 "Australia"  2.841745
1878 "Australia"  2.901711
1879 "Australia"  2.884088
1880 "Australia"  2.903649
1881 "Australia"  2.943671
1882 "Australia"  2.848141
1883 "Australia" 2.9487655
1884 "Australia" 2.9078684
1885 "Australia"  2.936056
1886 "Australia"  2.914104
1887 "Australia" 2.9838395
1888 "Australia" 2.9548414
1889 "Australia" 3.0089836
end
format %ty year

Thank you very much in advance!

[Rich Goldstein]

if I understand your question correctly (note that there are 20 observations not 20 variables), then no, it would not be the same; loosely speaking, -center- first does what you say and then subtracts that mean from each observation so that the mean of the centered variable is now 0

[Silvia Haertl]

Dear Rich:

Thanks, that was exactly what I didn´t understand. Thank you very much!

[Rich Goldstein]

you're welcome