Transformation from 3-digit ISCO68 to 4-digit ISCO88 or directly to EGP scale

[Arthur Eberle]

Hello everybody,

I am currently writing my master thesis and I am stuck with a small but necessary detail before I can (finally!) run my analysis and complete my work. Unfortunately also the Stata experts and tutors at my university never have been confronted with this kind of problem so I really hope you guys can help me! =)

Do you know how I can transform a 3-digit ICSO68 variable into a 4-digit ISCO88 variable or if I can directly transform from 3-digit ISCO68 into EGP scale?
Or more basic: I have a 3-digit ISCO68 variable and I must have in the end an EGP scale.

Some of you might give me now one of the following links (http://www.harryganzeboom.nl/isco68/index.htm or http://www.people.fas.harvard.edu/~i...ion_tables.htm) but they don't help as they only transform from 3-digit ISCO68 to 3-digit ISCO88. Unfortunately, I only know how to transform a 4-digit ISCO88 variable into EGP scale, not how to do it with 3 digits.
Maybe some of you already dealt with this kind of problem and already know a good and practicable solution. I would be very thankful for every advice, suggestion or hint!

Have a good start into the new week!
Arthur

[Dave Airey]

There are two packages in Stata, isko and isco. I assume you found those inadequate?

[Arthur Eberle]

Hi Dave,
I installed and used the isko-package (is the isco-package different to it?) but it transforms only 4-digits ISCO88.
So theoretically, if there's not direct transformation possible from 3-digits ISCO68 to EGP, it would already help to find a way to transform it to 4-digits ISCO88.

[Dave Airey]

I think you should look at the isco package. It transforms ISCO68 to EGP.

[Arthur Eberle]

Thanks a lot!
I just tried and unfortunately it requires a 4-digit ISCO68 variable. When I used 3-digit ISCO60 variable instead, two-third of my sample was missing. So I assume, it's not really working.

Do you know a possibility how to transform 3-digit ISCO68 into a 4-digit version? Having done that, I should be able to use then the isco package (again, big thanks!).

[Dave Airey]

No, I don't, sorry, it's not my area. Just an idea, what happens if you append a zero to your three digit code?

[Arthur Eberle]

Never mind, you already helped with the isko/isco hint!

I haven't tried yet to add a zero to my three digit code because I read in some papers and FAQs, that the outcome is not reliable and that there are many changed categories between 68 and 88.
But maybe I will investigate a little bit further and maybe there's a way.