Different number of observations. Can I use column percentages?

[Guest]

Hey,

I am new here and this is my first question... pls be gentle with me haha

So, I am currently working on an assignment regarding the topic of electoral research.

I will come straight to the point: I want to analyze a specific topic using a dummy-variable as my starting point. This variable is labeled "residence urban or rural". My goal is to analyze a bunch of crosstabs and present my results in mentioned assignment.

Problem is, the dataset I use has a relatively huge difference in observations between "rural" and "urban" (rural = 1639; urban 1751) so, I already realized that using row-percentages is not a valid option because the difference in observations would distort my results. (residence urban or rural is in the colums and variable XY is in the rows) First question: Am I right?

Second question: I can use colum-percentages right? Because even tho I have ~ 100 observation less on "rural" if I use colum-percentages this would not matter because I am now using the percentages independently from the entire number of observations.

Example using , colum :

example using , row :

[Carlo Lazzaro]

Guest:
it is not a matter of being rude... please see Help - Statalist #4.
I would recommend you to take a look at any decent textbook on categorical variables (see, for example Categorical Data Analysis | Wiley Series in Probability and Statistics) and discuss with your colleagues what percentage makes more sense reporting, as you clearly have different sample sizes for rural/urban.