Originally Posted by Hands Sandon
There's obviously no way to actually count each individual or to measure how long their circumstances persist, so from the data and trends pictures are drawn to as accurately as possible estimate the numbers.
You can sample and then attempt to generalize to the entire population. This would be the same as what is done with polling for a political candidate. Can they really count every voter for a poll? Of course not so they sample a representative number.
You could ask 2,000 people about their housing, their children and their living conditions, give a margin of error and report the findings. This report, you can't call it a study, did nothing like that. I also take issue with doubling up being declared the same as being homeless. A kid sharing a twin bed with their cousin isn't the same as sleeping on a park bench.
Finally this is from your own link.
The report paints a bleaker picture than one by the Department of Housing and Urban Development, which nonetheless reported a 28% increase in homeless families, from 131,000 in 2007 to 168,000 in 2010.
Dennis Culhane, a University of Pennsylvania professor of social policy, says HUD's numbers are much smaller because they count only families living on the street or in emergency shelters.
This is hilarious... their numbers are much smaller because......well because they count the actual homeless as opposed to using made up proxies to give them imaginary estimates of homeless.
Dude, this has got to be a joke.