In July of 2004, HUD issued regulations that require the following information to be collected by all homeless services agencies receiving HUD funding:

• Name
• Social Security Number
• Date of Birth
• Ethnicity and Race
• Gender
• Veteran Status
• Disabling Condition
• Residence Prior to Program Entry
• Zip Code of Last Permanent Address
• Program Entry Date
• Program Exit Date
• Unique Person Identification Number
• Program Identification Number
• Household Identification Number

With the personal identifying data collected within an HMIS, unduplicated counts of homeless persons using services can be obtained for any specified period of time. The Household Identification Number also enables counts of homeless families. The demographic and circumstances of homelessness information collected can be used to produce profiles of homeless persons using shelter. Program Entry and Exit dates allow one to calculate the number of days spent in shelter.

This workshop will demonstrate the types of research that can be conducted with HMIS data. Examples will be provided from research projects undertaken by the workshop presenter and his colleagues at the University of Pennsylvania. The findings of these projects include the following:

• The annual prevalence of the sheltered homeless in New York City and Philadelphia were found to be between 4 to 6% of persons in poverty
• Patterns of episodic and chronic users of shelter services have been identified among the shelter populations of New York City and Philadelphia
• A time-series analysis of Philadelphia data found that shelter use did not change significantly after the implementation of welfare reform in Pennsylvania
• By matching Philadelphia's HMIS with the City's AIDS Registry, shelter users were found to be 10 times more at risk of AIDS than the general public
• Event history analyses have revealed the major influences on shelter exits in New York City and Philadelphia

This workshop will also review the types of statistical analysis that can be conducted with HMIS data. Drawing on his 8 years of experience as a trainer for SPSS, Inc., the workshop presenter will use simulated HMIS data to demonstrate the following statistical techniques available in the SPSS statistical software program:

• Linear, binary logistic, and multinomial logistic regression
• Cox regression (an event history technique)
• Time-series analysis
• ANOVA

The steps to conducting these statistical techniques in SPSS will be presented, and the output of the techniques will be explained for workshop attendees. The methods of producing unduplicated counts and matching HMIS data to other databases will also be demonstrated.