Good catch. Of course, we'd have to know the base rate of dysfucntional families to know if it's significant. We'd also need better data, as his sample is small and non-random.