The Homeless Chronicles. Part 10

This is the tenth in a multi-part series about homelessness. Previously we noted that

• Homelessness changes depending on your definition, from very narrow to very broad.
• There are three main groups of homeless people – (1) people who chose to be homeless, (2) people whose disabilities create/contribute to homelessness, and (3) people who experience a major negative event which propels them into homelessness.
• The life style homeless basically want to be left alone. The temporary homeless are looking for help to return them to their normal life. The chronic homeless have such disabling conditions they are the most difficult to reach and to help.
• In recent years, a greater percent of the homeless are coming from the temporary group, fueled largely by the high cost of housing in relationship to their income.
• The commonly used federal survey under –estimates the number of homeless. In reality, there are over a million homeless people nationwide, 250,000 in California and between 10,000 and 15,000 in Orange County.
• There is wide diversity among California counties in the percent of the population who are homeless. Orange County at 0.22% (22 homeless people for every 10,000 residents) is mid low range with an average of 0.39% for the State.
• There is wide diversity among California counties in the percent of the homeless population who are unsheltered. Orange County at 58% is mid low range with an average of 65.3% for the State.
• There is wide diversity among California counties in the percent of the homeless who self-report being mentally ill. Orange County at 28% is mid range with an average of 29.5% for the State.
• The percent of homeless in Orange County who self-reported being mentally ill increased by 133% from 2017 to 2019, an increase unparalleled in the County’s history and in any other California county.

Today we’ll get some perspective on why the number and percent of homeless people who report being mentally ill has increased so dramatically in Orange County. We’re going to look at 5 factors:

• Changes in language used in the surveys
• Changes in methodology used in the surveys
• Commonalities among counties experiencing an increase
• Economic support for mental health programs
• Policies by the Board of Supervisors

Today we’ll look at the first three. Next time we’ll look at the remaining two.

Was the dramatic increase caused by a change in language?

The OC 2019 survey used the following question – “Do you have a serious mental illness or emotional impairment that seriously limits your ability to live independently?”

The 2017 survey asked – “Which of these situations keep you from holding a job or living in stable housing?” and then offered 7 responses, one of which was “Psychiatric/emotional condition”.

So there was a change in what the volunteers were asking. Can any or all of these changes account for the dramatic increase?

OC is not the only county to change the language. Other counties did too (e.g., Contra Costa, Sacramento) but the changes were relatively small compared to the changes in OC, which went from 12% to 28%.

Another way of assessing the change is to determine whether or not “serious mental illness” is more potent than “psychiatric/emotional condition”. In a review of 18 surveys from 15 different counties, the average prevalence of MI was computed for the different language used. The average prevalence for “psychiatric or emotional conditions” and “serious mental illness” were virtually indistinguishable (22.5% vs. 25.6%). There is some suggestion that “mental health condition” or “issue” generates higher positive responses (37%) and “serious mental health problem” generates fewer positive responses (19.2%) but neither of these was used in OC.

Can the differences be accounted for by the fact that some counties use a single question and some counties included mental illness as part of a list of conditions? In 2019 OC used a single question and in 2017 they used a multiple choice question. The average for single question was 26.9% and for multiple choice it was 32.3%. There is a difference between single question and multiple list responses, however, the use of the single question should have predisposed OC to having a lower incidence, but in 2019 it doesn’t. It’s higher.

The changes in language between the two surveys can’t account for the dramatic increase.

Was the dramatic increase caused by a change in methodology?

Almost every time the PIT surveys are conducted the methodology is changed to account for deficiencies encountered in the previous count. OC CEO Frank Kim recently commented that the 43% overall increase in the number of homeless from 4,792 (2017) to 6,860 (2019) was due to “much more thorough and sophisticated” approach, emphasizing the increase in volunteers, the increase in people surveyed, and the use of better technology. This “upgrade” occurs in OC and in all the other counties, so the fact that with changed methodology the numbers of homeless who self-report SMI decrease in more than half of the counties but increase dramatically in OC cannot be used as a justification for the increase. Here are some illustrative quotes about methodology from counties that found a decrease in the percent of homeless mentally ill –

Volunteer Turnout

• “…the largest turnout out of volunteers to date in Sacramento.”
• “…near 49% increase in the number of volunteers participating in the count, from about 150 volunteers in 2018 to 300 volunteers in 2019.” Kern
• “…an increased number of volunteers (more than 700), and increased participation from city leadership.” Riverside
• “…an over 1,000% increase in the number of community members willing and able to volunteer to count the homeless.” San Joaquin

New Use of IT

• “The use of a mobile application for data collection in the unsheltered count helped immensely to improve the quality and completeness of count and survey data.” Sacramento
• “Enumerators had the option of using paper surveys or electronic surveys on a tablet to conduct the count.” Yolo
• “…increased survey efficiency through use of a mobile web based technology instead of paper surveys…” Butte
• “…increased efficiency through use of a mobile web-based technology instead of paper surveys…” Riverside

Overall

• “…in 2019 the shift in methodology was much more significant in order to improve the accuracy of the unsheltered count.” Sacramento
• “The thorough group review of the survey resulted in the elimination and/or simplification of questions. Also resulted in the collection of more relevant and meaningful data.” Yolo
• “…an enhanced count methodology that covered a wider area within the allotted count time.” Riverside

OC, like virtually every other County in California, improved the methodology in 2019, so the increase in homeless mentally ill can’t be ascribed to the change in methodology especially since more than half the counties showed a decrease, not an increase. Here is the data for the counties listed above, all of whom implemented new technology, used more volunteers, and increased the number of people surveyed –

• Sacramento – 61.2%
• Yolo – 26.9%
• Butte -26.7%
• Kern – 14.7%
• Riverside – 14.5%
• San Joaquin -13.9%

Moreover, the changes in the percent of homeless mentally ill people are relatively small from year to year in these counties, regardless of the constant improvements in methodology over time.

Are there “Commonalities” at work?

Most of the hypotheses generated from a hypothetico-deductive perspective have been ruled out. In such a case, it’s not uncommon to resort to an analysis of commonalities. Looking at the data there were 10 counties (out of 30) that had an increase in the percent of homeless self-reporting SMI, and of these 10, 5 had an increase of 15% or more (San Bernardino, San Joaquin, Tulare/King, SLO and OC). Do these counties have anything in common that might account for their increase?

The counties range widely in size from 3,000,000+ (OC) to 284,000+ (SLO) and in density from 4,000 (OC) to about 100 (San Bernardino, Tulare/King) people per square mile. The number of homeless ranges from 6,860 (OC) to 1,069 (Tulare/King) and the percent of homeless people ranges from 0.12 (San Bernardino) to 0.40 (SLO). The percent homeless who are unsheltered averages 68.5% with a range from 57.7% (OC) to 79% (SLO). So there is nothing in the obvious demographics grouping these cities together.

Looking at the starting point in terms of what percent of the population self-reported SMI prior to the increase, the starting points for the 5 counties showing a 15% or more increase ranged from 12% (OC) to 31% (Tulare/King) which is extremely wide as is the starting points for the rest of the counties – 18% (San Diego) to 53% (Kern).

There are no obvious commonalities that can account for why there is such an increase for self-reported SMI in OC. That being said, it is beyond the scope of this report to delve into all the possible commonalities that might account for the increases. Factors such as rental costs in relation to the pay structure for wage labor jobs, the ethnic/racial composition of the population in relation to the ethnic/racial composition of the homeless population, law enforcement standards, etc. have not been considered.

Summary

Neither the change in language nor the change in methodology can account for the dramatic increase in self-reported SMI for homeless people in OC. Among those few counties like OC that experienced an increase in SMI, no commonalities could be found, including population size, density, nor baseline census.

Next time we’ll look at two causes that may account for the recent dramatic increases in OC in homeless people who self-reported being mentally ill.

About the Author

Dr. Jim Gardner is the former Mayor of Lake Forest. A Clinical Psychologist, he is a former University Professor and Department Head. He authored several reports about homelessness.