I bet you could model the likelihood of this, seeing if the respondents are enriched for families who are likely to be in communication with each other to some

I bet you could model the likelihood of this, seeing if the respondents are enriched for families who are likely to be in communication with each other to some degree (family, friends, neighbors, coworkers, in the same school, etc) or if its not significantly enriched compared to random chance.