Project Summary

Kent County Homeless Study



Kent County, Michigan is located in South Western Michigan. Population estimates for 1999 were 550,388 people. 1998 estimates show 28.2% of the population are less than 18, 10.8% is over 65, approximately 85.8% of the population is white non-Hispanic, 9.0% is black, 3.7% is Hispanic, and 2.2% is other. Approximately 80% of the population over 25 have graduated from high school and 20.7% of the same population is a college graduate. Nearly 70% of the population owns their own home. In 1995, the median household income was $39,240 with 9.6% of the population below the poverty level, and 14.1% of all children living in poverty. This is significantly less than that for the state as a whole (20.1%). The main metropolitan area is Grand Rapids, the second largest city in Michigan, with a population of 191,000 (U.S. Bureau of the Census 2000).

Figure 1. Map of Kent County

In the spring of 2000, the Kent County Health Department, Department of Development requested the support of the Michigan State University's, Center for Urban Affairs on investigating the of factors affecting homelessness in the Kent County area. This was to be accomplished by analyzing the characteristics of homeless shelter occupants and trends in shelter use based on shelter intake forms gathered from 1994-1999 from selected shelter in Kent County.

The Michigan State University, Homeless Study Research team entered the data from 11,132 intake forms, which was used to create a database. The description of the data was based on the standard intake form (Appendix A). The intake form completed by shelter workers provided the demographics for each client. There were approximately 13 participating shelters from the area in the data set. The analysis examined overall incidents, length of stay, and recidivism. This report provides a summary of these findings as well as suggested recommendations for the community.

The summary of the findings from the analysis area as follows:

The research team in consultation with the research advisors and in consideration of the data analyzed suggests the following recommendations be considered in the future efforts of combating homeless in Kent County:


Kent County shelter providers have demonstrated considerable insight and leadership over the past years in responding to the needs of homeless persons throughout their community. It is hoped that this analysis will further strengthen their efforts to assist this vulnerable population in their time of need.

Research Method

The period of data collection was from October 1, 1994 through September 30, 1999, a total of five years. Each year begins October 1st and ends September 30th. During that time, a total of 11,132 intake forms were collected from the participating Kent County shelters listed on page. Shelter workers familiar with the form completed the intake forms. Most of the information was gathered at the time of arrival from the individual seeking shelter, or in the case of multiple clients, an adult/parents of the homeless family. Occasionally, intake forms were completed at a later time. All data was self-reported and voluntarily provided by the clients.

Table 2 lists participating shelters in the study. While this list represents almost all of the shelters in the county it is significant to note that at least one large "faith-based" shelter that serves men only did not participate in the project. Their choice not to participate suggests an undercount of the total homelessness of the county and specifically an undercount of men in the study population.

Participating Shelters

Table 2.


Section 1.

Initial Shelter Study Findings

The following pages represent the data collected and analyzed in the study. The presentation of this information is similar to the format of the questions asked on the shelter intake form (Appendix A).

Chart 1

Prior Shelter History

Chart 1 indicates that of the number of valid intake forms analyzed 70 % of those seeking shelter indicated no previous shelter use. 30 % had previously been in a shelter.

Chart 2

Recent Residence

Chart 2 indicates the reported residences of shelter users. The chart indicates that 76 % of people listed Grand Rapids as their most recent residence and only 13 % reported coming from outside of Kent County.

Chart 3

High School/GED

Chart 3 indicates that 54 % of shelter users reported having a high school diploma or a GED while 46 % did not. The U.S. Census estimates of 1990 indicate a high school graduate rate of 80% in Kent County.

Chart 4

Attended College


Chart 4 reports that 20 % of shelter users attended college. This is similar to the census estimate of 20% who graduated from college in Kent County in 1990.

Chart 5

Family Composition

Chart 5 indicates that 78 % of shelter incidences reported involved females either as "single" females (35 %) or females with children (43 %).

The other category (9%) represented couples or with couples with children, males with children or extended family members.

Chart 6

Family Size

Chart 6 indicates that almost half of shelter users came alone. However, 20 % of the individuals involved a family of four or more people. It is unclear from the intake forms whether the responses to these questions represent all family members of those seeking shelter or the homeless person's family size including those not seeking shelter. If all members are not seeking shelter, for example, a single runaway youth might report a family size of 5 even though only one individual is seeking shelter.

Chart 7

Family Race

Chart 7 reports that 49 % of shelter users were black, 40 % were white, 6 % were identified as Hispanic and 5 % indicated "other" race. The 1990 census estimates for race in Kent County are 9% black, 4% Hispanic and 2% other. The intake forms indicate that minorities are disproportionately represented in the homeless shelter population.

Chart 8

Sources of Family Income

Chart 8 indicates that the single most frequently sited source of family income for shelter users was employment at 28 %, followed closely by Aid to Families with Dependent Children (AFDC) at 24 %. It is important to note that 23 % indicated no sources of family income.

We can observe that 43 % of the shelter users rely on some form of government transfer payment for their family income, (SSDI, SSI, Social Security, AFDC). Respondents identified other sources of money in 16 % of the cases. This suggests a population that is particularly vulnerable to public policy changes that support needy families.

Chart 9

Reason For Homelessness

Chart 9 indicates that the majority of shelter users reported either financial problems or domestics violence as their reason for being homeless. While drug and alcohol abuse was identified in 5.4 % of the cases, in discussions with the shelter providers of Kent County it was their general opinion that this self reported data was a significant undercount. Shelter users are hesitant to report illegal or inappropriate use of substances. Twenty-four percent of the intake forms indicated that there were other reasons not listed on the intake form.

Chart 10

Move From Shelter to Other types of Overnight Stays

Chart 10 indicates where shelter users reported moving when leaving a given shelter. Twenty-one percent of the respondents moved to another shelter, 11 % to friends of family and 15 % to their last residence. However, it is unknown where 28 % of shelter users moved to, while the remaining 12 % reported moving to another place not listed in the intake form.

Section II.

A Second Look at Homelessness in Kent County, Michigan

Various ways of examining homelessness are presented in this section. This section will explore overall incidents, length of stay, and recidivism in the homeless population. The data analysis is completed with a set of comparative analysis of variables suggested by the data or requested by the project advisors.

The first method to quantify homelessness was to count the actual intake forms – 11,132 over the period of time studied. Each form could represent and individual or a family, and it could represent a first time incident or a repeat incident. In an effort to provide policymakers and the shelter providers of Kent County with a more accurate assessment of the nature of homelessness in the greater Grand Rapids community, a unique identifier for each case was developed. This unique code was based on the client’s initials followed by six digits representing their date of birth. This identifier allowed the researchers to estimate recidivism rates and other variables such as the average length of stay or "bed-nights", which is family size x length of stay. The calculation of "bed-nights" was useful because it allows the shelter provider to estimate total shelter needs for homeless persons in the target area.

Of the 11,132 incidents in the study, 5,104 were unique incidents and over half (6,028) involved recidivism. An estimated 48% of the incidents involved single persons. For families, the average family size was 2.2 persons. With an average length of stay of 9.5 days, the average family homeless incident (intake form) used almost 21 bed-nights. Over a quarter million "bed-nights" were used during the five- year period by all shelter users.

About two-thirds (66 %) of the shelter intake forms involved children, defined as those under 18 years of age. Fifteen % of all these cases were single children under 18 years. In total there were over 14,000 children involved in homelessness. Children used over half the bed-nights during the period of the study. Thirty-seven percent of the incidents of homelessness involved children under the age of 6 years and 32 % involved children ages 6-17 years of age. The number of children experiencing homelessness in Kent County is worthy of additional attention by shelter providers and public officials.

Chart 11

Incidents of Homelessness By Year

Chart 11 shows the distribution of incidents over the five-year period of the study. The numbers remained stable for the first two years, dropped steadily for two years, and then rebounded last year. The fluctuation in 1997-98, in part, may represent the choice of specific providers to not submit reports for that year.

Chart 12

Bed-Nights By Year

Chart 12 reports the "bed-nights" per year. Not surprising, this follows the pattern of usage presented in Chart 11.

Chart 13

Length of Stay

Although the most frequent length of stay was one day, as indicated by Chart 13, many stayed for a much longer time, raising the average length of stay to 9.5 days. Most stays were less than a week, but 8 % stayed more than a month.

Chart 14

Length of Stay (Days) By Reason for Homelessness

Chart 14 indicates the relationship between the length of stay and the reported reason for stay. Substance abuse and medical recuperation report the longest lengths of stay.

Chart 15

Length of Stay By Shelter Group

Among the people examined in this study, as indicated in Chart 15, men had a much longer average stay than women or families. However, most of the incidents involving single men were reported from only one shelter so care should be taken in drawing significant conclusions from this analysis.

Chart 16

Incidents of Homelessness By Shelter Group

Chart 16 indicates that shelters serving women and children reported the most incidents of homelessness. They were followed by intake/assessment shelters and shelters serving special populations. The lack of participants by "faith-based" in male shelters in this data collection is a serious limitation of this particular analysis.

Chart 17

Repeated Incidents Of Homelessness

Slightly over one half of the incidents of homelessness involved recidivists or persons having more than one homeless episode as shown in Chart 17. 15 % of the incidents involved individuals who had at least four other homeless episodes during the five year time period of the study.

Chart 18

Use of Homeless Shelter In Bed-Nights By Recidivism

Chart 18 shows that almost 60 % of the shelter bed usage were by recidivists. The chronically homeless (having 5 or more incidents) consumed 15 % of the total "bed-nights" in the study population.

Chart 19

Percent of Recidivism by Reason for Homelessness

Chart 19 indicates the group most likely to experience more than one homeless incident were those homeless due to financial conditions. More than 60 % of the intake forms indicating financial as the reason for homelessness had more than one episode. Persons homeless because of mental health problems or domestic violence were likely to return somewhat over half the time (approximately 50-60 %). Those least likely to have more than one homeless episode were persons in recuperation and runaways.

Chart 20

Recidivism By Gender

From Chart 20 we can observe that men included in this study were more likely than women to have had only a single episode of homelessness. However, women were twice as likely as men to have experienced five or more homeless incidents. Again, this data should be interpreted cautiously since only one shelter reported most of the incidents involving single men.

Chart 21

Recidivism By Race

Chart 21 shows a relationship was found between race and recidivism. African- American homeless persons were more likely to have more that one incident of homelessness.

The data suggested there was a slight relationship between where the client moved after leaving the shelter and recidivism. Clients who moved to another shelter or to a substance abuse treatment shelter were more likely to repeat their homeless episode. Less likely to repeat were those who moved out of town, returned to their previous residence, or moved to a VA facility.

Chart 22

Numbers of Homeless People By Race and Gender

In this study, interaction was found between race, gender, and homelessness. As seen in Chart 24, in each of the racial groups, there were more homeless females, but this difference was most extreme for blacks where there were five times as many females as males.

Chart 23

Recidivism By Race and Gender

The group most likely to recidivate consisted of African-American females. Chart 23 shows that almost two out of three African-American females experience recidivism. About half of the white and other female recidivate, as does 40 % of all the males.

The next section will outline a set of recommendations for consideration by Shelter Providers and community leaders in Kent County.

Section III.

Recommendations and Conclusion

In the United States, there are an estimated 700,000 people who are homeless on any given night, and up to 2 million people experience homelessness during one year (NCH 1999). Kent County, Michigan like many other communities throughout the United States has, in part, responded to the needs of its homeless citizens. In examining the intake forms collected by shelter providers over a 5-year period, a number of potential areas of emphasis emerge.

Clearly one of the most significant findings is the prevalence of children in the shelters of Kent County. Either as runaways or as members of a homeless families 2/3 of the shelter incidences reported children present. Children experiencing homelessness have the potential to experience a myriad of other negative social, psychology and economic consequences. Addressing the unique challenges of youth homelessness presents and immediate and significant challenge to concerned citizens and community leaders.

The data collected by the shelters of Kent County also reveals that homelessness occurs for a variety of reasons in the community. Lack of financial resources is a significant contributor to shelter use. This "cause of homelessness" is compounded when the source of family income is examined. The dependence of shelter users on public assistance programs (TANF, SSI, VA) suggests that public policy initiatives that effect the level and availability of these resources will have a direct effect on the need for shelter beds. It is highly likely that with the recent changes in the states public welfare programs that as families reach the end of their eligible years of assistance that the need for shelter beds may increase.

In analyzing the reasons for homelessness and the nature of recidivist in the shelter users, it is clear that shelter providers must continue to strengthen their functional partnerships with agencies addressing substance abuse, domestic violence, the school district and the producers of affordable housing. Homelessness can be viewed as a symptom of other social, economic and emotional challenges. To reduce the number of homeless in a community, the public and private institutions and leaders of the business community will need to build effective and focused efforts that addresses the factors that precipitate the homeless experience. Though Kent County’s Continuum of Care program is fairly comprehensive, additional partnerships with supporting agencies could enhance the community’s goal of reducing homelessness.

Of less immediate significance, but of potential future importance, is the need to provide consistent training to shelter intake workers on the use of the intake form. Minor modifications in the form are recommended such as clarifying family size questions, reasons for homelessness, education levels and where shelter users are going when they leave the shelter. Improvements in these questions will allow future analyst to refine and build upon the findings and recommendations presented in this study.

In conclusion, Kent County shelter providers have demonstrated exceptional insight and leadership over the past years in responding to the needs of homeless persons. It is hoped that this analysis will further strengthen their efforts to assist this vulnerable population in their time of need.


National Coalition for the Homeless. 1999. "How many people experience homelessness?" NCH fact sheet #2. Washington, DC.

PowerPoint Presentation


CEDP Home Page    Back