egen help: quintiles

[Alan Jeddi]

Hello, I am trying to generate a new variable, alcohol use (in quintiles). I would like to take my variable and cut it in to 5 equal parts.
I have demonstrated an association with another variable when being measured continuously, but the magnitude is very small. Thus I would like to rescale.

I have tried various permutations of the egen command, but I cannot quite find the right syntax to divide it up into 5ths.

Any help would be greatly appreciated.

[Carole J. Wilson]

See

help pctile

[Nick Cox]

https://bmcmedresmethodol.biomedcent...471-2288-12-21 and https://stats.stackexchange.com/ques...ictor-variable are helpful in this territory. Throwing away information will not in general clarify an association.