The LA CoC Homeless Count is conducted annually in an ongoing partnership between the University of Southern California (USC) and the Los Angeles Homeless Services Authority (LAHSA) since 2017. The Homeless Count provides a point-in-time (PIT) estimate of the unhoused population residing within the LA CoC geographic region, which is divided into eight Service Planning Areas (SPAs), including estimates of the sheltered and unsheltered populations. The LA CoC encompasses the whole of LA County, except for the cities of Pasadena, Glendale, and Long Beach who conduct their own respective PIT counts. Estimates of the unhoused population are derived from several data elements, including: 1) an observation-based count by volunteers of unsheltered individuals and dwellings of cars, vans, RVs, tents, and makeshift shelters (CVRTM), 2) a demographic survey of unsheltered adults (age 25 years and older) conducted by a team at the USC Suzanne Dworak-Peck School of Social Work, 3) a youth survey-based count of unsheltered youth (age 24 years and younger) mainly carried out by homelessness youth service provider agencies, and 4) a shelter count based on administrative data from the Homeless Management Information System (HMIS) and reporting from shelters not represented in HMIS. A detailed description of the LA CoC Homeless Count methodology can be found here. Here, we briefly describe each of the data elements and the general methodology used to arrive at the count of sheltered individuals and estimates of unsheltered individuals. 

Point-in-Time Count: LAHSA conducts the PIT Count every January that can be thought of as measuring the visual homelessness in LAC CoC in that volunteers record their observations of the number of people they see living on the streets and dwellings assumed to be housing homeless individuals encountered while canvassing all 2,310 census tracts (CTs) that make up the LA CoC. The counts for individuals encountered on the street are provided by estimated age category (under 18, 18 to 24, and 25 and older), with families counted separately from individuals. 

Demographic Survey: Because the observations made by volunteers do not provide information about the number of people living in the dwellings that are counted or about the demographic characteristics of the unsheltered population, the USC team conducts a separate survey between December and March each year across a sample of CTs in LA CoC, determined by using a two-stage stratified random sample approach. A team of data collectors canvass each sampled CT at least once during the surveying period and conduct surveys among any individual experiencing unsheltered homelessness they locate and who agrees to participate. The respondents are assumed to be selected at random from the homeless population in the CT. The survey collects information on basic demographic characteristics (e.g., age, gender, race/ethnicity), where the respondent slept last night and in the last 30 days, homelessness history, veteran status, health status, employment status, and more. Responses to the Demographic Survey are used to estimate the number of homeless individuals classified within select demographic and subpopulation categories. Proportions of individuals representing specific demographic and subpopulation characteristics are estimated within each household type (i.e., adults with children, adults without children, veterans without children, and veterans with children). These proportions are then applied to their respective PIT population totals at the SPA-level and then summed across SPAs to derive the estimated subpopulation counts for the CoC. Further, weights are used to estimate proportions from the survey data to account for the sampling method used. Sampling weights are applied to each demographic survey respondent, first calculated by taking the inverse probability of CT selection (for the CT where they were surveyed) and then adjusted to account for non-response and individuals who could not be approached (e.g., due to safety concerns) and post-stratified to the household PIT Count. 

Youth Enumeration Survey: Homeless youth are considered a “hidden population” who may not be easily discernible through a visual tally. Therefore, survey-based methods are recommended to separately conduct a homeless youth count. The LA CoC Homeless Count utilizes a youth enumeration survey, collecting information on demographic characteristics and household composition, to estimate the size of the population under 25 years of age experiencing unsheltered homelessness and serve as a PIT count of unsheltered youth. To further describe the unsheltered youth population (e.g., veteran status, chronic homelessness status), data from the Demographic Survey conducted among individuals age less than 25 years are used. The youth enumeration survey is typically conducted during the last 10 days of January across a selected sample of CTs within the LA CoC. Youth count efforts then employ three types of location-based data collection strategies to specifically target homeless youth: 1) street survey teams who canvass selected CTs and survey any identified youth-aged person experiencing homelessness willing to participate; 2) in-person surveys at survey sites (e.g., youth service providers); and 3) phone-based surveys at survey sites. Total, demographic, and subpopulation counts were calculated from the youth surveys by household type (i.e., unaccompanied transition-aged youth and minors and parenting youth) using a similar method to the unsheltered adult count. The main difference in the methodology between the youth and adult unsheltered counts is that the adult unsheltered count is based on observations mostly by volunteers of individuals and CVRTM with complete coverage of the CoC, while the youth count is based on the youth enumeration surveys conducted in a sample of CTs mostly by homeless youth service providers. This means that the sampling weights play a greater role in determining the number of unsheltered homeless youth. Youth sampling weights were calculated for the enumeration surveys as the inverse probability of selection and then adjusted for non-response. The enumeration survey was used to estimate basic demographic characteristics by household type and SPA, and the demographic survey was used to produce subpopulation proportions by household type and SPA.

Shelter Count: Shelter count estimates were generated using PIT shelter count data provided by LAHSA, which is a complete enumeration of all shelters in the LA CoC. The shelter PIT Count provides the raw number of homeless individuals living in emergency shelters, transitional housing, and safe havens, including those receiving vouchers for hotels or motels provided by these shelters. HMIS data is then used to estimate demographic and other subpopulation characteristics of sheltered individuals. The shelter PIT count data are assumed to be a complete enumeration of the sheltered population and, therefore, do not require additional weighting to generate demographic and subpopulation estimates. Demographic and subpopulation characteristics are derived by estimating the proportion within the HMIS data by SPA for each shelter type and household type (i.e., individuals, adult and transition-aged-youth families, and veterans). Proportions are then applied to the analogous PIT shelter counts, by SPA, household, and shelter type, to obtain their characteristics. Subpopulation estimates are applied at the SPA level to SPA-specific counts and then summed to get the total CoC result.