ABS responds to "Census underestimating homeless numbers" Sydney Morning Herald web site (published 11 May)

Dear Sir,

An article on the Sydney Morning Herald web site (11 May) claimed that the Census "appears to be grossly underestimating" Australia's homeless population.

It based this claim on a health service survey of the central business district of one city - Adelaide. This survey, conducted over four months (June-September) in 2005, estimated that between 311 and 455 homeless people were living in the CBD. It compared this outcome to the 2001 Census count. The Census count was four years earlier and conducted on one night in August.

The 2001 Census figure for total homeless in the Adelaide CBD was 712. The count consisted of two groups of homeless - those who, on Census night, had no conventional roof over their heads (104) and those without a usual address who, on Census night, were in such places as refuges and hostels. Census night is a cold time of the year, when many people who may at other times be on the streets are staying in shelters, hostels or with friends/acquaintances in private dwellings, so would not be included in the "primary" homeless count, though they could appear in the overall homeless count.

There are always difficulties in counting a population such as the homeless, but the Australian Bureau of Statistics, which conducts the Census, takes its responsibility seriously. People who work with the homeless - and even homeless people themselves - are recruited and trained by the ABS to be interviewers. Interviewers use a shortened version of the Census form to record details from individuals.

On Census night, 8 August this year, ABS staff will again be doing their best to gain an accurate count of the homeless all across the country.

Dennis Trewin
Australian Statistician
ABS House
45 Benjamin Way
Belconnen ACT 2617

16 May 2006