Question about regression and editing values

[Fabian Wettergren]

Hello everyone,

I want to test if people who feel attachment to their region also feel attached to Europe. To test this I want to do a regression analysis. I have so far stumbled onto two problems that I would like to have some input on.

1. A few observations says: "I dont know" or "no answer". How do I remove this?

2. In the answer to the question, very close=1 and not close at all=4. In my head it makes sense to have it the other way around? My statistical knowledge is a bit limited but does this even matter when I do the regression? If so, is there a way to change the values of the answers so very close=4 etc.

Thanks in advance,
Fabian

[Nick Cox]

What you are showing is a tabulation of a variable with value labels. If the label very close corresponds to values of 1 and not close at all to values of 4, then presumably the first two categories have corresponding values less than 1 and the observations you want to ignore can be omitted by insisting that the regression is done only

... if inrange(Q45D, 1, 4)

where I am guessing that the variable is called

Q45D

.

Otherwise, reversing 1 to 4 to 4 to 1 is just a matter of subtraction.

gen Q45Dnew = 5 - Q45D if inrange(Q45D, 1, 4)

You would need to define value labels correspondingly.

That has an intended side-effect of ignoring what you want to ignore. Most experienced researchers would, I guess, treat a variable like this as ordered categorical, not as strongly measured.

[Nick Cox]

Also cross-posted at https://www.reddit.com/r/stata/comme..._of_variables/

As commented there, telling people about cross-posting is a request here and a rule there.