Monday, May 4, 2009


The last few days I was researching shelter studies. I was in need of a number and learned some useful stuff about shelters one has to consider when talking about shelters in the context of domestic violence.

First of all something about the shelter service. Most DV-shelters offer emergency shelter (the average stay here is 36 days) and transitional housing. Transitional housing means after the stay in a shelter one can stay up to 24 months (average stay is 300 days) until permanent living arrangements are made. That is usually the place (including hotel rooms) where an american shelter puts men who need shelter. The percentage of men who actually stay in an emergency shelter is below 1%. A real surprise to me was the sheer number of children who stay at a shelter. Arround 51% of shelter inhabitants are children, for transitional houses it are arround 60%.

Another important thing to keep in mind is the organisation of the 2,000 DV shelters in the USA. Most of them are small shelters and it seems to me there isn´t a huge network and there is not much information on the number of sheltered women (the number I was searching for). What I did find though was snapshot data from the National Census of Domestic Violence Services (Census). This survey calls every DV-shelter available and asks about the inhabitants for that day. With this number I was able to calculate a number which is of course not accurate. It should however be good enough to give us a rough overview.

To clarify the table below. During the Census not every shelter gave out information (or couldn´t be reached). The return rate tells us how many shelters answered and again how accurate this information is. The number I used for this table is the number of sheltered adults (no children) as well as requests for emergency shelter that couldn´t be met. I didn´t include transitional housings as this would include previously sheltered women as well and screw up my calculation. The estimated total should give us information on sheltered adults if the respons rate was 100%. This guess shouldn´t be that inaccurate. I included all 3 snapshots there are to see if there is a huge fluctuation in that number. It is not. (Information taken from the Census Data of the National Network against DV found here)

Estimate for the daily snapshot:


I really hope blogger doesn´t screw my table up. Anyhow my second estimate the projected number for a year will be a more inaccurate guess. I am using the average shelter stay (36 Days) with the number of days in a year to create a factor I can multiplicate my Estimated Total with. The factor is 10.1. The closer the shelters were to an average day the better my guess is.

The last column takes the following information from a different study into account:

Study is based on a survey of 3,410 people served by DV shelters in eight states during a six-month period in 2007/2008.
One in four (24%) had stayed in a shelter before that visit during which they took that survey [...] the mean length of time participants had stayed in shelter was
33 days, and the median was 22 days (27 days for mothers)
from here

I just subtracted the quarter to avoid double counts.

Year___Estimated / Year___Without multiple visits

2006 from here
2007 from here
2008 from here

The intersting question now is, how inaccurate is my guess. I found some information for the state of Texas for certain years which I can compare with the detailed state information from the 2008 census.

Validation of Data from Texas:
Services Received In Texas by Family Violence Victims
2001 - 12,581 Adults served

Compared to 1179 * 10.1 = 11,907

(I did not use the unmet request number here and had to calculate the number from the study as the responded rate was not 100% ( 71% - 1675 -> 100% = 2359) I took into account that 50% of those should be children)
from here

The difference is not that big. It wouldn´t surprise me though if my guess is in the +/- 50% range.

Why is this number important? I have seen a number used to give an "at least" rate of horrible DV cases, badly battered women. But can we asume that all those women stand for the worst DV cases there are? We can´t and sadly there is not much information I could find for the USA. Canada has a similar Census report with more interesting information:

Shelters for abused women

Nearly one-third of all women who had sought temporary accommodation in a shelter for abused women on April 14, 2004 had stayed there at some time during the past, according to a new report.


Of all women who had stayed in shelters previously, 40% had been there once in the previous year, 38% had been there two to four times, and about 1 in 10 had been to the facility five times or more during the previous year.

The largest proportion of women staying in shelters, just over one-third, were between 25 and 34 years old.

The report is based on data from the Transition Home Survey, a biennial census of residential facilities that provide shelter to female victims of domestic violence. The survey provided information on the facilities and their services for the full year between April 1, 2003 and March 31, 2004, as well as a one-day snapshot of activities and clients served on April 14, 2004.


According to the one-day survey snapshot, taken on April 14, 2004, of the 6,109 women and children residing in shelters, about 5,000 had escaped an abusive situation representing about three-quarters of all women residing in shelters. Women also turned to shelters for other reasons such as problems with housing, addictions and mental health.

The vast majority of the women staying in shelters to escape abuse were fleeing psychological or emotional abuse. Almost 7 out of 10 reported physical abuse, 50% threats, 46% financial abuse, 31% harassment and 27% sexual abuse.


Women using shelters for reasons other than to escape abuse constituted about one-quarter (24%) of shelter residents. Over two-thirds of these women sought shelter because they were unable to find affordable housing.

Seven out of 10 women with parenting responsibilities brought their children with them to shelters. Children accounted for about half of shelter residents and two-thirds of these children were under the age of 10.
from here

Looking at the above for the situation in Canada, about 1/4 are not abused (most of them just unable to find a house and using the shelter as a social service) and about 1/3 are not physically abused. I do not want to marginalise emotional abused women here, but this just shows us however that it is not so simple to say something about the situation of women in shelters so it is only faulty to equal shelter inhabitants with severly beaten women.

By the way the number that was given to me by a feminist was 380% higher than the number I calculated.

No comments:

Post a Comment