Hello,
Having a daft moment.

I surveyed 2006 respondents about cats. 1002 were Urban and 1004 rural.
I have two separate variables. The first asks whether, during a 7 year period, (out of all respondents) respondents domestically sourced cats *yes/no. The second asks if (out of all respondents) respondents imported cats *yes/no within the 7 year period.
Unfortunately, I didn't have one variable that asked if they ever domestically sourced or imported a cat.
When I cross tabulate the two variables, there are some respondents who imported and domestically sourced cats. So I want to make sure I don't over count in Q1 below.

Q1: Now, I want to see if there is a difference (using logistic regression) in IMPORTING (only) a cat based on whether a respondent lives in an urban or rural household.
I created a new variable, and to avoid said over-counting, used the following command

generate gotcat=.
replace gotcat=1 if (importcat == 1) & ( sourcecat == 0)
replace gotcat=0 if (importcat== 0) & ( sourcecat == 1)
tab gotcat


Then presumably, it is:
xi: logit gotcat residential

Where "residential" is a categorical variable wherein residing in a rural household = 1 and residing in an urban household =0

Is this the best way to get to evaluate the question of interest?
Thank you in advance