Formula to standardize a variable without "gen" and "egen"

[Maelle Banuls]

Hello,

I am a French student who is starting to work with Stata 16, and I need to find a formulation to answer the question:
"Create a new variable z standardizing variable x from a command other than gen and egen. The operation must be performed on a single line of code".

I transform the real variable into z and x for a better understanding, I don't know if that's how you do it, I'm new.

I specify that I spent time looking for this everywhere, without concrete results.


Thank you for your help !

[Nick Cox]

I have no idea how this could be done. I think you should bounce this back to your teachers for help. Pas possible, je crois.

-- unless you're expected to use Mata.

[Paul Dickman]

What is meant by standardize? There are many ways to standardize a variable. Can you give an example (e.g., using gen or egen)?

[daniel klein]

Perhaps you are supposed to find and use community-contributed commands that do this? Or you are supposed to write such a command and use it?

[Maelle Banuls]

All right, I will share with you the question seen in class and its answer.

We had to create the variable "farm_area_norm" which normalizes the variable "ha_landused_hh".
Here is the answer seen in class :

egen Mean_farm= mean(ha_landused_hh) if ha_landused_hh !=.
egen Sd_farm=sd(ha_landused_hh) if ha_landused_hh !=.
gen farm_area_norm = (ha_landused_hh - Mean_farm)/Sd_farm if ha_landused_hh !=.
sum ha_landused_hh Mean_farm Sd_farm
sum farm_area_norm
drop Mean_farm Sd_farm

Now it would be a question of being able to answer this question via another command that would allow this to be encoded in a single line.

Thank you for taking the time to answer me.

[Attaullah Shah]

Your instructor might be asking you to write a Stata program and use all the above lines in that program, thereby, reducing all these liens to a single line of code.

[Paul Dickman]

Your instructor might be asking you to write a Stata program and use all the above lines in that program, thereby, reducing all these liens to a single line of code.

That's the only way I can think of doing it in one line. There's already a user-written command for doing this (type "findit zscore" to download it). There are other similar commands.

The values you want should be the same as the standardised residuals from a null regression model, but I'm not aware of a way of implementing that in one line of code (other than writing a program).

[Nick Cox]

Meanwhile #5 is pretty poor code. It's inefficient and indirect to fire up egen to produce new variables holding constants -- only to drop them again immediately once you've extracted results.


The code is careful about avoiding missing values, which is better style, but the missing values will be ignored any way.


This does most of the work directly:

su ha_landused_hh 
gen farm_area_norm = (ha_landused_hh - r(mean))/ r(sd)