How to generate missing data under a proportion

[Chang Xu]

I want to simulate missing value based on my own dataset, but I don't know how to randomly 'replace' the data by missing value for a variable under a certain proportion. Is there any functions or commands for it? Thanks

[Joseph Coveney]

Assuming that "under a certain proportion" means the rate at which you want the variable to be replaced with missing, then

quietly replace variable = . if runiform() < 0.25

where the "certain proportion" is 25% in this example.

Make sure that your original dataset has been saved somewhere (or else work with a frame copy) and (set the seed beforehand if you're interested in reproducibility.

[Chang Xu]

Assuming that "under a certain proportion" means the rate at which you want the variable to be replaced with missing, then

quietly replace variable = . if runiform() < 0.25

where the "certain proportion" is 25% in this example.

Make sure that your original dataset has been saved somewhere (or else work with a frame copy) and (set the seed beforehand if you're interested in reproducibility.

Thanks Joseph. It works well !

Best regards
Chang