Methodology

CURA’s Community-Engaged Research Model

The Center for Urban and Regional Affairs (CURA) believes in the production of community-engaged research, which means that we value the meaningful involvement of our community-based partners and their clients throughout the research process, from the identification of research question(s) to the dissemination of results. CURA’s community-based research model aims to invert the traditional academic research model of entering a community as the expert, extracting data, and returning to the academy. There is power in defining research questions and in controlling the production of knowledge. When research is conducted in communities of color and low-wealth communities, a power imbalance often exists between researchers and community-based organizations. CURA's community-based action research model aims to reorder that power relationship.

We believe that such engagement can influence research to be more community centered, useful, and trustworthy—and ultimately lead to greater use and uptake of research by practitioners, clients, and policymakers. In short, we embrace a collaborative model of engaged research in which it is critical to consult our partners on the development of the research design/questions and collaborate with our partners and relevant stakeholders as we conduct the research and disseminate research findings. We want to ensure that we can troubleshoot any unforeseen challenges/barriers together and encourage our partners to utilize CURA’s research and recommendations to strengthen their impact on low-income communities and their investors. When we use a community-based research model, community members are not the subjects of research—they are the co-producers of knowledge.

Early Engagement Partnerships

The first step in elevating the expertise and power in our communities was to connect with community partners who are currently working in the area of evictions. In the fall of 2017, Dr. Brittany Lewis set out to listen to and engage with community partners to ensure that:

  • the project was useful and important to our community-based housing partners and local government;
  • we defined the characteristics of our study participants to help minimize risk and any unnecessary disruption to their lives;
  • those people most impacted by evictions in North Minneapolis are at the center of the data collection pool;
  • meaningful and direct connections were made with end users of our research findings.

To meet these objectives, Dr. Lewis conducted one-on-one interviews with local housing practitioners and those most affected by housing instability in North Minneapolis. These partners have critical insight into the realities of evictions and housing stability that helped to inform CURA and its partners’ knowledge of the work. Additionally, these insights helped to inform the semi-structured interview tools used for both tenant and landlord interviews.

Early Engagement Partners Included:

InquilinXs UnidXs por Justicia

Neighborhood Hub

HOMELine

Northpoint

Mid-Minnesota Legal Aid

People Serving People

Minneapolis Promise Zone

Community Action Partnership

Hennepin County Courts

City of Minneapolis, Regulatory Services

Minnesota Multi Housing Association

Stairstep Foundation

Urban Homeworks

1 Family, 1 Community

Project for Pride in Living

Minneapolis Public Schools

AEON

Faith leaders across 55411 & 55412

Northside Achievement Zone

St. Stephens Human Services

Minneapolis Public Housing Authority

 

 

CURA Evictions Research Project Advisory Council

            The second step in engaging the community in the project was to convene an Advisory Council comprised of tenants, landlords, community organizers, community-based staff members, and staff members from the city of Minneapolis as well as Hennepin County (see acknowledgements and copyright information). The Advisory Council has had four primary roles and has aimed to collaboratively:

  1. develop interview questions and participant engagement protocols;
  2. help design recruitment strategies and actively recruit study participants;
  3. review de-identified interview transcripts, major themes, and common narrative frames found in qualitative data; 
  4. disseminate research findings to appropriate stakeholders along with CURA’s independently developed policy recommendations based on well-informed research, including an analysis of best practices in other cities.

In the fall of 2017 and early 2018, the CURA Evictions Research team worked with the Advisory Council to finalize the recruitment strategy for the project as well as the interview protocol for both landlords and tenants. In the spring of 2018, Dr. Lewis and the research team began outreach, recruitment, and data collection.

Research Design       

The research design for this project is a convergent parallel mixed methods design. The goal is to collect qualitative and quantitative data to provide a stronger and more comprehensive picture of the issue of evictions in North Minneapolis (Creswell, 2014). This project draws on both in-depth interviews and critical ethnographic observations, as well as Hennepin County Housing Court records and rental license records. Qualitative and quantitative data were collected for both landlords and tenants, with all four categories being analyzed separately. The analyses were then integrated and explicated in the project’s findings.

Research on evictions can be challenging due to a lack of recordkeeping, the vast array of experiences of both tenants and landlords, and the lack of importance placed on low-income mothers of color (Hartman and Robinson, 2003). The strength of a mixed methods approach is that each method can provide different types of information and can minimize the limitations of the other method (Creswell, 2014), which is critical in the study of an elusive and complex process such as evictions.

Setting and Participants

The Minneapolis Innovation Team’s Evictions in Minneapolis report found that of the over 3,000 evictions filed each year (2015/2016), 45% to 48% of renter households that experienced evictions took place in two zip codes: 55411 and 55412. As noted previously, North Minneapolis residents have experienced historic disinvestment and intentionally structured racial segregation and discrimination. Additionally, these two zip codes host a high percentage of low-income Black mothers as well as a high rate of individuals who receive county and government assistance. For the purpose of CURA’s research and capacity, we strategically focused on these two zip codes when identifying both tenant and landlord interview participants. In total, 68 tenants and 32 landlords were interviewed. Specific demographics of participants are outlined in the following relevant sections.

Landlords

Landlord Overview

In the 2018 CURA evictions study, we conducted interviews with landlords who have filed eviction actions on tenants in the 55411 and 55412 zip codes in the last 3 years. The landlord interviews were conducted to learn more about:

  • what policies and procedures they have in place to determine that filing an eviction is the best course of action for dealing with a tenant;
  • how they decide when to evict a tenant and then determine both the cost benefit of eviction and owning rental property in the two zip codes;
  • what practices or tactics they employ once the decision is made to evict.

Landlord Sampling Strategy

Landlords were chosen for interviews based on a controlled random sampling using the contact information found from the Active Rental License dataset, which is regulated by the city of Minneapolis and shared through the Minneapolis Open Data portal. Data was pulled on January 24, 2018.

Records were separated into two lists based on zip code (55411 or 55412) and then further sorted by tier classification (Tier 1, Tier 2, Tier 3) within each zip code, resulting in six separate listings of properties. For each tier, a list of unique property owner names was generated. Some property owner names appeared more than once in these lists if they used multiple spellings or name formats to apply for licenses (different spellings, use or omission of a middle initial, etc.). Using the Excel random number function, a random number was generated and assigned to each name in each unique-name list. The names in each list were then ordered from smallest random number value to largest random number value. The top 10 property owner names were selected from each list in each tier, resulting in a sample of 60 unique names.

Second and third samples were pulled using the same procedure as the first sample. The second sample was pulled from the complete list of property owner names—names from the first sample were not removed. Consequently, some of the same names appear in both the first and second samples.

Response rate of landlord recruitment strategy:

Sample 1: 60 names

Intakes completed

12

Interviews completed

12

Unable to contact

10

Sample 2: 60 names

Intakes completed

10

Interviews completed

1

Unable to contact

12

Sample 3: 60 names

Intakes completed

20

Interviews completed

12

Unable to contact

5

 

Participants

 

Landlord Profile

Landlord Profile

Source: The Illusion of Choice interviews and intake data, CURA 2018

A total of 32 landlords were interviewed including 72% (23) males and 28% (9) females. The sample primarily self-identified as White male 53% (17), 19% (6) identified as White female, with 13% (4) identifying as Black or African American female, 6% (2) as East Asian male, and 3% (1) as South Asian female, Latino male, and Native American female, respectively. Of the landlords interviewed, 19% (6) either worked for organizations or were personally listed on the Minneapolis Innovation Team’s report as frequent filers. Finally, 16% (5) of the landlords interviewed reside in 55411/55412, with 84% (27) residing elsewhere.

Frequent Filers of Eviction Actions Representation in Focus Zipcodes

Source: Minneapolis Innovation Team, 2016 and city rental license data

 

Procedures[1]

Data for the landlord qualitative portion of this project were collected through semi-structured interviews at a place of convenience for the landlord.[2] An initial introductory email was sent to all owners who had entered an email address on the rental license application. Each owner in the sample then received a minimum of two follow-up calls from graduate research assistants who explained the project and invited participation. Additionally, any landlord who appeared in the sample and was also noted by the Minneapolis Innovation Team’s report as a frequent filer was personally contacted by Dr. Lewis to ensure representation from these individuals in the sample.

All interested landlords were asked to complete an intake form to self-identify demographic information as well as easily quantifiable data such as the number of properties owned and/or managed in 55411/55412 and rental term lengths of properties. Upon recruitment into the study, landlords were walked through a consent form that described the project, the voluntary nature of participation, and contact information for both Dr. Lewis and the University of Minnesota.

To begin the interview, landlords were asked to identify themselves and describe how long and why they chose this type of work. Participants were asked the same structured interview questions, and answers were probed when appropriate. Interviews lasted approximately 45 to 60 minutes, and each participant was given a $50 Visa gift card in appreciation for their time and expertise.

Quantitative Data Analysis

Our analysis examines ownership trends and dynamics within four scales: (1) the specific properties of the 32 interviewed landlords, (2) the properties and landlords in the study area of 55411 and 55412, (3) the city of Minneapolis with the study area of 55411 and 55412 excluded, (4) and the entire city of Minneapolis. The four scales were chosen to illuminate not only how the interviewed landlord sample differed from trends and practices in rental property quality, ownership, and management citywide but also how the study area differed from the city as a whole.

In addition to the self-reported data from the intake forms and qualitative interviews with the 32 selected landlords, we utilized:

  • Minneapolis active rental licensing data, extracted from the city’s open data portal in January 2018, to identify the number of landlords and units at our specified geographic scales, the type or structure of ownership of those units, the quality or tier of those units, and to compare the nature of this data, where fields are self-reported by property owners, to data collected in a third-party nature by the county (see next).
  • Hennepin County parcel data, downloaded in January 2018, to identify homesteading status, taxpayer information, and property type for the 32 interviewed landlords.

Tenants

Tenant Overview

In 2018, we conducted interviews with 68 tenants who had experienced an eviction filing in either of the zip codes (55411 or 55412) since 2015. The tenant interviews were conducted to:

  • identify the conditions that often lead to housing instability and eviction;
  • gain a clearer understanding of these tenants’ housing composition and stability overtime;
  • understand the various income streams and the networks of support that tenants rely on for survival.

Tenant Sampling Strategy

In the spring of 2018, the Evictions Research team began recruitment for participation in this project. A purposeful, homogeneous sampling strategy was utilized to engage participants with similar personal experiences of an eviction filing in the same two zip codes, while each participant provides a unique voice and insight for a deep understanding of the evictions experience (Patton, 2015). Participants were recruited through outreach with Hennepin County Housing Court, community-based partnerships and events, and word of mouth. In total, 34% (23) were recruited from Housing Court, 19% from a partnership with HOMELine, 16% (11) from a community-based organization, 13% (9) from Legal Aid, 10% (7) from a personal referral or flyer, and 7% (5) from a partnership with People Serving People. Tenants were recruited for participation in the project until August 31, 2018. They were eligible for the project if they had received an eviction filing within the last 3 years in either the 55411 or 55412 zip codes. Eviction actions were confirmed by the research team through Hennepin County case records.[3]

Referral source for tenant interviews

Referral source for tenant interviews

Source: The Illusion of Choice interviews and intake data, CURA 2018

 

Participants

In total, 68 individuals participated in the tenant interviews. The majority of tenant participants identified as female, 78% (53), with 22% (15) identifying as male. The largest percentage of participants self-identified as Black or African American females at 62% (42), with 18% (12) identifying as Black or African American males, 6% (4) as biracial or multiracial females, 6% (4) Native American females, 3% (2) White females, and 1% (1) East Asian female, East Asian male, Native American male, and White male, respectively. The average age of respondents was 44.2 years old, with a range of 22 to 70 years old. In total, 13% (9) reported that they did not complete high school, 37% (25) reported a high school diploma or GED, 44% (30) reported some college, 4% (3) had a bachelor’s degree, and 1% (1) reported a doctorate. Finally, 28% (19) reported receiving either a public housing or Section 8 subsidy. 

Tenant Profile

Tenant Profile

Source: The Illusion of Choice interviews and intake data, CURA 2018

 

Procedures

Data for the tenant qualitative portion of this project were collected through semi-structured interviews at a place of convenience for the tenant. Through community-based partnerships and attendance at Hennepin County Housing Court, the CURA Evictions team advertised the opportunity to participate in the project and interested tenants opted into the program. Eligible tenants, those who had experienced an eviction in the 55411/55412 zip codes within the last 3 years were enrolled in the study through an intake process that included questions regarding relevant demographic, income, and eviction experience information. 

Upon recruitment into the study, participants were invited for a one-to-one interview at a time and place of their convenience. At times, these interviews were conducted at Housing Court, immediately after an eviction hearing, and others were scheduled within the following week. At the beginning of the interview, Dr. Lewis reviewed a consent form that described the project, the voluntary and confidential nature of participation, and contact information for both Dr. Lewis and the University of Minnesota.

To begin the interview, tenants were asked to state their names, how they located the property in question, and whether or not it was their first choice. All tenants were asked questions from an interview protocol, and answers were probed when appropriate. Additionally, as with semi-structured interviews, participants had the opportunity to elaborate on responses when appropriate. Each interview lasted approximately 45 to 60 minutes, and each participant was given a $50 Visa gift card in appreciation for their time and expertise.

Quantitative Data Analysis

The quantitative data analysis for tenant data consisted of examining court documents related to each participant’s unlawful detainer (UD) filings (when available). Records were accessed using public workstations made available for that purpose in the 4th District Court Records Center in the basement of the Hennepin County Government Center. Court records were accessed by entering the first and last name of the interviewee. Records consisted of scanned court documents, organized by case number. Each document was reviewed for key information including:

  • address for UD filing;
  • date of filing (when the landlord submitted the eviction filing to the court);
  • name of landlord;
  • reason stated for filing a UD;
  • whether or not the plaintiff/defendant appeared in court and had legal representation;
  • name of the referee;
  • date and outcome of the hearing;
  • sums or actions agreed to in settlement agreements;
  • any additional actions resulting from noncompliance with the settlement agreement;
  • notation of any additional documents;
  • notation of when/whether a Writ of Recovery of Premises was issued/executed
  • whether or not any rent escrow or conciliation filings appeared to be associated with the UD filing.

Notes were compiled on any attachments, and any extensive handwritten notes that appeared on documents were copied verbatim or noted in a detailed manner. Data were then examined on a case-by-case basis as well as in the aggregate to provide a more comprehensive picture of the nature of eviction action filings among the tenant participants.

Qualitative Data Analysis Process: Landlords and Tenants

The process for analyzing the interview data from the landlord and tenant interviews follows a similar, multistep process, though each group was analyzed separately.

To begin, each interview was audio-recorded. Dr. Lewis also took extensive notes and immediately following the interview, she noted important aspects and reflections. This process was documented and critical to eliminating recall bias. Additionally, each audio recording was transcribed verbatim through Rev.com. A member of the research team then reviewed each transcription to ensure its quality, as well as clarify any points in the interview that were noted as inaudible. A second review of each transcription was completed at that time, and the transcriptions were de-identified to protect the identity of the participant.

Data analysis for the interviews utilized both inductive and deductive processes. First, inductive grounded theory techniques of open coding and constant comparison were used to evaluate emerging themes in the data. Open coding allows the research team to inductively look for patterns in the data, whereas constant comparison is a process of evaluating where emerging themes were similar and different across and between interviews (Corbin and Strauss, 2015; Patton, 2015). Throughout data collection and analysis, the evictions team collaborated on compiling emergent themes and eventually created a code book with concepts from the data. Additionally, previous research and case studies provided a deductive framework for where the emergent themes were or were not congruent. Any disagreements on themes were resolved through a team discussion. Finally, the interviewer’s notes and reflections were integrated with the emerging concepts and themes, as well as the actual statements from interviewees, to make meaning of the similarities and differences across eviction filing experiences.

Evictions Research Project Advisory Council

Evictions Research Project Advisory Council

Photo by Jonathan Miller

The mixed methods approach is intentionally employed in this project to minimize the limitations of each method. However, to ensure the rigor of qualitative data, the research team engaged in several strategies to ensure quality assessment. Two meaning-making sessions were held with the Evictions Research Project Advisory Council across the project timeline, with a landlord session in the fall of 2018 and a tenant session in the winter of 2019. Participants were given copies of de-identified transcripts and asked to generate their own themes and profiles of landlords and tenants, respectively. This form of data triangulation increases the credibility of data due to the use of multiple perspectives in the findings (Lincoln and Guba, 1985). By engaging in data analysis strategies collectively with individuals who are deeply ingrained in the work, as well as others who can approach the work from the outside, the findings are more transferable and provide a stronger illustration of the impact of evictions, an underanalyzed phenomenon yet a social crisis, particularly for low-income communities of color.


[1] All procedures, as well as consent protocols and measurement tools, were approved by the Institutional Review Board (IRB) of the University of Minnesota.

[2] In the rental license records, each property owner name is connected to an “applicant”; in some cases the owner is the applicant, and in other cases, the applicant is an individual or company hired to manage the property. Owners for each property were contacted.

[3] Exceptions included cases that had been expunged from a tenant’s record and/or three cases of informal evictions, which are noted in the report.