We demonstrate that the observed (in-)stability of n-doped carbon nanotube films in air not only depends on the employed dopant but is also strongly affected by sample-specific factors, such as the film thickness and density. We show this for two typical dopants, polyethylenimine and a potassium crown ether complex, by preparing films of increasing thickness. We argue that reports on dopant stability cannot be properly assessed without knowledge of these sample-specific parameters, which explains some of the conflicting results in the literature.