Archive for January 20th, 2011

20th January
2011
written by Pete Witte

Since the release of The State of Homelessness in America – and frankly, long before that – we’ve gotten questions about homelessness data.

It was my job at the Alliance to conduct data acquisition and analysis for The State of Homelessness report – and we thought it might be nice to shed some light on the process:

First, we needed to acquire all of the data. Now I’d love to say that this was as easy as going to websites of public-data sources and clicking a link that read Click here to quickly download all the data you need for your report-making-fun, but as anyone who has worked with data knows, data acquisition is much more complicated.

It was further complicated by the fact that – as report readers know – we gathered a lot of data. If you check out the Appendix of the report, you can see that we acquired data from the Census Bureau, the Bureau of Labor Statistics, the Bureau of Justice Statistics, the Administration for Children and Families, and Realty Trac – just to name a handful.

Getting our hands on the data, as anyone who’s tried to extract data from a federal agency can tell you, was no small feat. We made Freedom of Information Act (FOIA) requests, sent request letters by email, downloaded huge micro-data files that took hours, and so on.

Once a data sources were acquired and once I was familiar with the variables, definitions, and limitations of the data, it was time to run analyses, organize and interpret the data, and put the data into tables for further interpretation. What we receive are huge files – often spreadsheets – of raw data. We then had to filter through the data, find the specific pieces of information we were looking for, and interpret those into data that we could use for our report.

For example, we knew we wanted to measure income vulnerability, but we needed to figure out how: did we want to include employment income or all sources of income? What about the average number of weeks worked? Average income among poor workers or all workers or both? These are the types of questions we asked as we put the project through quality assurance crash tests.

From the beginning, we planned to assess changes in homeless counts, as these changes have been tracked by the Alliance for a number of years. Most of the economic and demographic factors were also planned but some of the measures were given tune-ups as we went along. And that meant going back into the fray to get more information.

I know this doesn’t answer every question but I’m hoping that this helps readers understand what kind of crunching the data went through before it went into our report! We look forward to hearing from you about The State of Homelessness and we hope that the data in the report proves useful as we move closer to ending homelessness.