There are varying degrees of control over outcomes. The judgment comes from the correlation between an unobserved variable (effort at controlling or preventing obesity) with the observed variable (actual obesity).