Strange problem.
I have a dtype == object dataframe column with string values and NaNs. Looks like this:
df Response 0 Email 1 NaN 2