Supporting Statement for OMB Clearance Request
Part B
Moving to Work, Landlord Incentives Cohort Evaluation
Contract Number: GS00F252CA
October 3, 2022
Prepared for:
Elizabeth Rudd
Paul Joice
Office of Policy Development and Research
U.S. Department of Housing and Urban Development
Submitted by:
Abt Associates Inc.
6130 Executive Boulevard
Rockville, MD 20852
University of Hawai’i
2424 Maile Way
Honolulu, HI 96822
Table of Contents
Part B: Collection of Information Involving Statistical Methods 1
B.1: Respondent Universe and Sampling Methods 1
B.2: Procedures for Collection of Information 5
B.3: Methods to Maximize Response Rates and Deal with Non-response 9
B.5: Individuals Consulted on Statistical Aspects of the Design 11
Part B of the Supporting Statement for the Moving to Work, Landlord Incentives Evaluation – sponsored by the Office of Policy Development and Research at the U.S. Department of Housing and Urban Development (HUD) – considers the issues pertaining to Collection of Information Employing Statistical Methods. HUD has contracted with Abt Associates Inc. and its subcontractors to conduct the evaluation. The evaluation team will conduct 1) an impact study, 2) a process study, and 3) a cost study.
The Landlord Incentives Evaluation will collect descriptive information on the programs and policies implemented by 28 PHAs selected to join the Landlord Incentives Cohort (“treatment PHAs”). To rigorously evaluate the impact of the landlord incentives, the study will also compare the outcomes achieved by the treatment PHAs to those achieved by a group of 112 similar PHAs that do not have MTW designation (“comparison PHAs”). However, for multiple data collection activities, the study will only investigate the 22 comparison PHAs chosen for further in-depth analysis due to their similarities to their matched treatment PHA. These 22 PHAs are referred to as the “in-depth comparison PHAs.” See section B.2.1 for detailed information on how we will select these 22 in-depth comparison PHAs.
The evaluation relies on multiple data sources, both secondary data (that is, already existing data) and primary data (that is, new data we will collect as part of this evaluation). The secondary data used in the evaluation is primarily HUD administrative data—notably HUD’s Housing Inventory Portal (HIP) database and the MTW Supplement—as well as the American Community Survey (ACS) data.
This submission seeks clearance for these primary data collection activities:
Semi-structured interview guides for site visit and telephone interviews with staff from treatment and a subset of comparison PHAs; and
Semi-structured interview guides for site visits with landlords within a subset of treatment and comparison PHA service areas.
We will collect data from the entire population of treatment group PHAs (n=28). All of the Landlord Incentive Cohort (treatment) PHAs have agreed to data collection as a condition of their participation in MTW.
For the comparison group, the web survey (discussed more below) will target all 112 comparison group PHAs, but the in-depth site visits and interviews in year 5 will occur at only 22 comparison PHAs, which we have characterized as the “in-depth comparison PHAs.” Because Hawai‘i is the only large PHA in the treatment group, we will include its closest comparison PHA in the site visits. This leaves another 21 comparison PHAs to be selected for the in-depth comparison sample. See section B.2.1 for detailed information on how we will select these remaining 21 in-depth comparison PHAs.
All 28 treatment PHAs and all 112 comparison PHAs will receive a baseline web survey in year 2 and a follow-up survey in year 5. The surveys will gather detailed information on incentives and program outcomes that would not otherwise be available from secondary data sources. The web survey will provide valuable descriptive information for the process study and will also provide important context for the impact study. It also will help the study team prepare for site visits. We will incorporate survey responses into the summaries of PHA characteristics to be developed for site visitors prior to the visits.
Among treatment PHAs, we aim to achieve a response rate of 100 percent; we will continue to follow up by phone or email until all 28 treatment PHAs have provided complete responses. Among comparison PHAs, we aim to achieve a response rate of at least 60 percent, including complete responses from all 22 in-depth comparison PHAs. We will continue to follow up by phone and email toward that goal.
Exhibit 1: Sample Sizes and Response Rates by PHA Grouping
|
In Study |
Completed |
Response Rate |
Treatment PHAs |
28 |
28 |
100% |
Comparison PHAs (all) |
112 |
67 |
~60% |
In-depth Comparison PHAs (this is a subset of all Comparison PHAs) |
22 |
22 |
100% |
Total |
140 |
95 |
68% |
Site visits for in-person interviews will take place at all 28 treatment PHAs in year 2 and again in year 5. Site visits will take place at the 22 in-depth comparison PHAs in year 5. In year 3, interviews will be conducted by phone with the 28 treatment PHAs only.
Exhibit 2: Response Rates for PHA In-Person (Site Visit) Interviews
|
In Study |
Participates in PHA Site Visit Year 2 |
Participates in PHA Site Visit Year 5 |
Response Rate for Both Years |
Treatment PHAs |
28 |
28 |
28 |
100% |
In-depth comparison PHAs |
22 |
- |
22 |
100% |
Total |
50 |
28 |
50 |
100% |
Exhibit 3: Response Rates for PHA Phone Interviews
|
In Study |
Participates in PHA Site Visit |
Response Rate |
Treatment PHAs |
28 |
28 |
100% |
Total |
28 |
28 |
100% |
Interviewees will include PHA and HCV program leadership such as the executive director or director of the HCV program; client-facing staff such as caseworkers and housing navigators; the inspection manager or lead inspector; and staff responsible for landlord outreach and recruitment such as a landlord liaison. We may also include PHA partners and other stakeholders such as local realtors who work with voucher households or organizations that provide housing navigation services to them. In our initial communications with PHA sites, we will confirm which staff and stakeholders will be most appropriate to interview.
Selecting PHAs for the Landlord sample
The 20 PHA service areas where we will interview landlords are a subset of the total treatment group and the associated in-depth comparison group PHAs, where our site visits will occur. We will split these 20 PHA jurisdictions between two qualitative analysis technique: semi-structured thematic interviews (TIA) and a structured Qualitative Impact Assessment Protocol (QUIP), described in detail below. We will split these 20 PHA service areas between the TIA and QuIP approaches. We will interview using the TIA with landlords at 7 treatment group and 7 comparison group PHA service areas. We use the QuIP with landlords at 6 treatment group PHA service areas.
We will select the treatment group PHA service areas (7 TIA, 6 QuIP) from among the evaluation’s 28 treatment PHAs. We will select comparison group PHA service areas (7 TIA) from among 22 PHAs where we are conducting site visits. In selecting PHAs for the TIA and QuIP landlord interviews we will aim to maximize heterogeneity of the PHA service area contexts while maintaining largely comparable treatment and comparison groups (Exhibit 4).
Exhibit 4: PHA Sampling Strategy for Landlord Interviews
Size of PHA |
TIA Interviews |
QuIP Interviews |
||
Treatment |
Comparison |
Treatment |
Comparison |
|
Large |
1 (Hawai‘i) |
1 |
0 |
0 |
Medium |
3 |
3 |
3 |
0 |
Small |
3 |
3 |
3 |
0 |
Total |
7 |
7 |
6 |
0 |
Note: Size is based on number of voucher units.
We plan to stratify and sample based on PHA size (dividing our sample into two equal sized groups based on units so that the selection groups for medium and small are equal) and then ensure that we have the greatest possible heterogeneity across the sample in (1) incentives offered by the PHAs and (2) market context (market tightness and census region). For both the TIA and QuIP landlord interviews, ensuring substantively different landlord incentives means that final site selection will occur after the initial round of PHA interviews and the year 3 phone interviews, when we expect to have clarity regarding each PHA’s concrete plans.
TIA PHA Sample
We will select the Hawai’i PHA’s service area and its closest matched comparison PHA in order to understand the experiences of landlords working with a large PHA. We will select PHA service areas for three medium PHAs and their closest matched comparison PHAs, and three small PHAs and their closest matched comparison PHAs. Within the medium and small PHAs, we will purposively select the three sites to ensure heterogeneity as follows:
At least one in a tight rental market and one in a loose market.
Different census regions.
Substantively different landlord incentives.
QuIP PHA Sample
We will select six treatment PHA service areas—three medium and three small—that provide the greatest heterogeneity in market context and that represent a range of different landlord incentives using the same criteria as in the selection of PHAs for the TIA sample.
Sampling Landlords for In-Depth Interviews
Once we have selected the PHA service areas, we will sample landlords within each drawing from a combination of PHA landlord lists, Craigslist1 and other websites most frequently used by landlords to list available rental units in the PHA service area. We will aim to complete 20 landlord interviews from each PHA selected for either the TIA or the QuIP sample, totaling 400 interviews. In treatment group PHAs, we want to make sure that we conduct interviews with different types of landlords: landlords that rented to voucher holders prior to the MTW landlord incentives; landlords that started renting to voucher holders since the MTW landlord incentives; and landlords that have not yet rented to voucher holders. In comparison group PHAs in the TIA sample, we want to conduct interviews with a mix of landlords that have and have not rented to voucher holders. Exhibits 5 and 6 provide detail on the sampling frame, including both recruitment pool and expected interviews completed, for landlord interviews.
Landlords are varied in size, experience and organizational structure. In most markets we expect to see a range of different types of landlords from small “mom and pop” landlords who own one rental building of no more than one to five units, to large national ownership companies who contract out with property management companies. This variety in landlord type will influence whom we talk to. For smaller landlords we anticipate few staff and will be able to easily identify the individual who manages renting out the unit. In cases where a property is professionally managed by multiple people, we will target the staff member with the greatest decision-making authority. In most cases this will be the Property Manager, although the Assistant Property Manager will be interviewed in cases where the alternative is non-response. Staff who exclusively work in tenant placement, marketing, administration, or maintenance will not be interviewed.2
Landlords will likely vary in responsiveness to our recruitment phone calls and emails depending on whether the PHA has an existing relationship with them or not. Given the level of effort budgeted for recruitment, we expect overall to see 25 percent of landlords contacted to respond to our recruitment efforts and complete an interview (Exhibit 5-6).
Exhibit 5: Sampling Frame for Interviews with Various Types of Landlords in Treatment PHAs
Type of Landlord |
TIA Interviews per PHA (n=20) |
QuIP Interviews per PHA (n=20) |
||||
Recruitment |
Interviews |
Response Rate |
Recruitment |
Interviews |
Response Rate |
|
Pre-MTW landlord incentives HCV landlord |
24 |
6 |
25% |
24 |
6 |
25% |
New HCV landlord |
28 |
7 |
25% |
28 |
7 |
25% |
Non-HCV landlord |
28 |
7 |
25% |
28 |
7 |
25% |
Total |
80 |
20 |
25% |
80 |
20 |
25% |
Exhibit 6: Sampling Frame for TIA Interviews with Various Types of Landlords in Comparison PHAs
Type of Landlord |
TIA Interviews per PHA (n=20) |
||
Recruitment |
Interviews |
Response Rate |
|
HCV landlord |
40 |
10 |
25% |
Non-HCV landlord |
40 |
10 |
25% |
Total |
80 |
20 |
25% |
In addition to sampling based on landlord relationship to the HCV program, we will also examine the geographical distribution of properties where relevant. We will look at the baseline geographical distribution of voucher properties or any program goals of increasing landlord participation in under-represented neighborhoods within the voucher program. We will also examine the proportions of landlords of different types and sizes—corporate large landlords or small individual owners—within a PHA service area in order to ensure a range of landlord sizes. Random sampling within these secondary categories will be to ensure that, as best we can, we have a representative sample of landlords in our interviews.
We will collect data from the entire population of treatment group PHAs (n=28). All of the Landlord Incentive Cohort (treatment) PHAs have agreed to data collection as a condition of their participation in MTW.
The comparison group will be made of 112 comparison PHAs, but the in-depth site visits and interviews in year 5 will occur at only 22 comparison PHAs, which we have characterized as the “in-depth comparison PHAs.” Because Hawai‘i is the only large PHA in the treatment group, we will include its closest comparison PHA in the site visits. This leaves another 21 comparison PHAs to be selected for the in-depth comparison sample.
We used an iterative process to select four comparison PHAs out of all of the PHAs in the US for each of the 28 treatment PHAs, totaling 112 comparisons.
Step 1. To maximize internal validity of the impact study, we first identified the PHAs that meet the same eligibility characteristics as the treatment PHAs did. To be eligible to apply for the MTW designation in the Landlord Incentive Cohort, a PHA had to—
Have no more than 27,000 combined public housing and HCV units.
Have an HCV program.
Be a high performer.
Not be troubled, not be participating in HUD’s Community Choice Demonstration,3 and not already a participant in any of the other three MTW expansion cohorts.
Comply with requirements to submit data to HUD, including a reporting rate in HIP of at least 90 percent of households.
We added a sixth requirement, that comparison PHAs must—
Have sufficient and interpretable data on issuances to allow an estimate of voucher success rate in 2018, 2019, or both.4
These five cohort eligibility criteria and the added data availability criterion produced a set of 953 possible comparison PHAs from among the 2,163 PHAs in HUD’s 2020 Picture of Subsidized Households dataset that have voucher programs.5
Step 2. Of all possible pairs of treatment-comparison PHAs, we ranked pairs by their match quality. First, we aimed to match the treatment-comparison PHA pairs on three characteristics that enhance the face validity of the treatment-comparison contrast:
Same geographic region as the treatment PHA;6
Similar laws governing discrimination based on source of income (from Poverty & Race Research Action Council information)
Total number of voucher and public housing units between 0.5 and 2 times the number of voucher and public housing units of the treatment PHA.
Imposing these three criteria alone provided many more possible treatment-comparison pairs than required. Therefore, we next selected from within the refined set of treatment-comparison pairs to PHA pairs that match as closely as possible on the key characteristics explained in the text box.7 Using the principals of rigorous evaluation, we selected these characteristics to improve internal validity because they (1) control for selection bias, (2) control for baseline values of outcome measures, and (3) control for characteristics hypothesized to strongly influence outcomes.
Instead of treating all key characteristics as equally important, we used coarsened exact matching (CEM) to examine their influence, even if limited, on PHAs’ decision to apply to participate in the Landlord Incentives Demonstration. This step further enhances the internal validity of the treatment-comparison contrast. CEM is a proven method that yields high-quality matches with respect to selected covariates and often succeeds where other methods, such as propensity score and nearest neighbor, do not (Iacus et al., 2012). To match potential comparison PHAs to the treatment PHAs, CEM considers a list of characteristics (provided by the researcher) and then finds the optimal “coarsening” for each baseline variable in the model (and combinations of those variables). For example, CEM might “coarsen” the baseline success rate by identifying the cut-offs for a low, medium, and high voucher success rate category. CEM optimizes the cut-offs so that they are maximally informative about the decision to apply to participate in the Landlord Incentives Demonstration.
Using the CEM model, we still found more than four possible comparison PHAs for each treatment PHA. Therefore, we imposed additional tie-resolution criteria to improve the similarity of treatment and comparison groups. For this tie-resolution criteria, we use a distance metric that includes all of the key characteristics. The distance metric gives equal weight to all six key characteristics.
Step 3: We selected the top four comparison matches for each treatment PHA, based on (1) the number of strict criteria on which the comparison and treatment PHA matched (a whole number between 0 and 3) (2) the number of key characteristics on which the comparison and treatment PHA matched within CEM category (a whole number between 0 and 6), (3) a continuous distance measure of the comparison and treatment PHAs’ key characteristics.
Step 4: Of the 112 comparison PHAs selected in Step 3, 98 were unique. That is, 18 comparison PHAs were one of the four best matches to more than one treatment PHA. To avoid duplicates, we assigned the duplicate comparison PHA to the treatment PHA that most needed it (that is, for the treatment PHA with fewer good-quality matches), forcing the other treatment PHAs with the same comparison PHA to select the next-best comparison PHA. These 112 comparison PHAs will be reached out to for the PHA web surveys.
The balance test finds only two statistically significant differences in the baseline characteristics—average months families spend on the waitlist (32 for comparison PHAs, 25 for treatment PHAs- a difference of 28 percent of a standard deviation in months on the waitlist) and homeowner vacancy rate (1.36 percent for comparison PHAs, 1.12 percent for treatment PHAs- a difference of 34 percent of a standard deviation in vacancy rates).
The final step to determine the 21 in-depth comparison PHAs (not including Hawai'i's comparison PHA) from the 112 comparison PHAs will be to use voucher success rate data availability to narrow down the best treatment-comparison PHA matches, then selecting the final 21 comparison PHAs to reflect a group that maintains diversity in PHA voucher program size, source of income laws, rental market tightness, and census region. If one of our 22 selected in-depth comparison PHAs (including Hawai'i's comparison PHA) does not agree to a site visit, we will replace it with the corresponding treatment PHA’s next best match.
PHA Web Survey and Interviews
This submission does not require a statistical methodology plan, as the sampling of PHA participants is purposive. The study team will work with each PHA’s Executive Director to identify the most relevant staff to answer the study’s research questions. The respondents are limited in number and are not intended to constitute a representative sample of all staff in the study PHAs. In addition, we will not seek to draw statistical inferences from the study data covered in this submission.
Most respondents will only be involved in annual data collection. Some PHA staff may be involved in a limited amount of ad-hoc data collection from the research team in the form of short telephone calls or email exchanges, but these are not formal reporting mechanisms, will not follow a standardized script, and will not be burdensome.
Landlord Interviews
The sampling of PHAs service areas for the landlord sample is purposive and requires no statistical methodology plan. We will randomly select the sample of landlords within these service areas stratified based on landlord relationship to the HCV program (in the program or no), type and size of landlords (corporate large landlords or small individual owners), and geographical distribution of properties to ensure that, as best we can, we have a representative sample of landlords in our interviews.
Primary Data Collection Activities (Web Survey, PHA interviews, and Landlord Interviews)
The site visits and interviews are designed to provide in-depth qualitative information; no estimation procedures will be used. The data analysis will be descriptive.
PHA Web Surveys
All 28 treatment PHAs and all 112 comparison PHAs will receive a baseline web survey in year 2 and a follow-up survey in year 5. We will send the link to the survey to director of the HCV program, and we will ask them to identify the appropriate person or people to complete the survey. We will program surveys using Confirmit, using a mix of open-ended and close-ended items. Surveys will take approximately 30 minutes to complete.
PHA Phone Interviews
We will conduct phone interviews with all treatment PHAs in year 3. As with site visits, we will use existing data to compile a summary of key PHA characteristics prior to the interviews. We will schedule up to three interviews at each PHA. Each interview will be attended by two members of the study team, a lead interviewer and a notetaker, ideally the same people who participated in site visits at the PHA in year 2. Interviews will last approximately 30 minutes and will be recorded and transcribed with participants’ permission. If participants are unwilling to be recorded, Abt staff will take detailed notes.
In-Person Interviews with PHAs
Two members of the study team will attend interviews, a lead interviewer and a notetaker. Each interview will take 30-60 minutes depending on the topic. With participants’ permission, interviews will be recorded and transcribed.
We do not believe we will be communicating with people from special populations.
We have minimized burden on treatment PHAs by having telephone interviews in year 3 and just two site visits, one at baseline (year 2) and one after multiple years of implementation (year 5).
Before starting recruitment, we will contact the directors of each HUD Regional Office to notify them of the study and request their support. For treatment PHAs, we will coordinate with HUD to ensure that PHA leadership and staff understand the timeline and expectations for PHA data collection from the outset.
For the 112 comparison PHAs, we will ask HUD to send a letter from the Assistant Secretary for Public and Indian Housing a few weeks before the web survey will be fielded in year 2. The letter should explain the MTW landlord incentive evaluation, notify the PHA that it has been selected as part of the comparison group, and request the PHA’s cooperation in the web survey. For the 22 in-depth comparison PHAs, the letter should also ask them to agree to a site visit and explain its timing. In year 5, we will ask the Assistant Secretary to send another letter to all 112 comparison PHAs about the second web survey, and a follow-up letter with more specifics about the commitment to the site visit to the 22 in-depth comparison PHAs.
To prepare for our site visits, we will reach out to the directors of both treatment and comparison PHAs, asking them to identify the appropriate key contact at the PHA. We will email that person to coordinate site visit logistics and schedule.
Treatment PHA Landlord Recruitment
We will recruit landlords differently according to their relationship to the HCV program. We will identify the HCV-participating landlords in treatment PHAs using PHA records such as HUD Form 50058 or the addresses to which the PHAs send the rent subsidy checks. We may also collect landlord contact information from websites such as www.affordablehousing.com, a website on which PHAs will often request HCV landlords list their units. We will identify non-HCV landlords in treatment PHAs from Craigslist and other apartment listing sites. We will screen to ensure we include only listings that contain valid contact information and have rents between 50 and the larger of 120 percent of the FMR or 110 percent of SAFMR. We will screen listings in two ways. First, we will screen to assess whether or not they mention Section 8 or HCV being welcomed by the landlord. Second, for the TIA study we will ask a landlord during the recruitment conversation whether they accept Section 8 or HCV. We are interested in landlords who are currently renting to and welcome HCV, as well as landlords who are not currently participating in the program. We will not be able to assess whether a landlord would not accept HCV unless they reveal this in the interview, since we are not conducting correspondence testing prior to recruiting for the landlord interviews.
We will conduct recruitment in the service areas of treatment group PHAs slightly differently according to whether we are doing the TIA interviews or the QuIP interviews. Exhibit XX provides details about the kinds of outreach and recruitment we will conduct in treatment group PHAs.
Exhibit 7: Details of Recruitment Materials for TIA and QuIP Landlord Data Collection, Treatment Group PHAs Only
|
TIA Interview Landlords |
QuIP Interview Landlords |
PHA outreach to HCV landlords |
We will ask PHAs to send an outreach letter to all sampled HCV landlords to alert them to the interviews and encourage participation |
None; to minimize awareness of interviews being part of research |
Abt study team letter |
We will send a letter to the selected sample of HCV and non-HCV landlords directly |
We will send a letter to the selected sample of HCV and non-HCV landlords directly |
Abt study team phone calls/emails |
We will make up to 3 phone call and email follow-ups after the initial letter. When we do not have sufficient response, we will make an additional 3 phone calls, text messages, or emails |
We will make up to 3 phone call and email follow-ups after the initial letter. When we do not have sufficient response, we will make an additional 3 phone calls, text messages, or emails |
Comparison PHA Landlord Recruitment
In the service areas of the comparison group PHAs, we will recruit landlords only for the TIA interviews. (The QuIP data collection is designed to find how effective the MTW incentives are, and thus is not relevant to the comparison PHAs.) The comparison landlord sample will be selected from (1) landlords actively leasing to voucher tenants and (2) those who either do not mention voucher status or state that they do not accept vouchers.
As in the treatment PHA service areas, the HCV participating landlord sample will come from information supplied by the PHA, and non-HCV landlords will come from Craigslist and other local real estate listing sources. Ideally, recruitment will follow the same process as in the treatment sites. Exhibit 8 provides details about the kinds of outreach and recruitment we will conduct in comparison group PHAs. Should any comparison PHAs not consent to assist in recruitment, we will recruit HCV landlords from public sources such as www.affordablehousing.com or from Craigslist, including both advertisements explicitly soliciting voucher holders and those that make no mention of vouchers.
Exhibit 8: Details of Recruitment Materials for TIA Landlord Data Collection, Comparison Group PHAs Only
|
TIA Interview Landlords |
PHA outreach to HCV landlords |
We will ask PHAs to send an outreach letter to all sampled HCV landlords to alert them to the interviews and encourage participation |
Abt study team letter |
We will send a letter to the selected sample of HCV and non-HCV landlords directly |
Abt study team phone calls/emails |
We will make up to 3 phone call and email follow-ups after the initial letter. When we do not have sufficient response, we will make an additional 3 phone calls, text messages, or emails |
Early drafts of the interview protocols have been reviewed by HUD personnel, Abt Associates staff, and our Resident Expert Panel in order to ensure that the instruments are clear, flow well, and are as concise as possible.
The individuals listed in Exhibit 9 below made a contribution to the design of the evaluation. Baseline data collection will be administered by MTW4 staff, under the direction of Abt Associates (and overseen by Meryl Finkel as Project Director).
Exhibit 9: Individuals Consulted
Name |
Telephone Number |
Role in Study |
Dr. Philip M. E. Garboden |
808-956-7800 |
co-Qualitative Data Lead |
Hannah Thomas |
617-520-2632 |
Director of Qualitative Analysis |
|
|
|
Inquiries regarding the study’s planned analysis should be directed to:
Meryl Finkel |
Abt Associates, Project Director |
617-349-2380 |
Elizabeth Rudd |
HUD, Contracting Officer’s Representative |
202-402-7607 |
1 While Craigslist is not fully representative of the rental market across all cities, it is used consistently by landlords across most markets. As a result it will offer the research team the best option for building a consistent recruitment approach for non-HCV landlords across the sampled PHA service areas. If we find that craigslist is under-utilized in a site, we will pursue alternative websites guided by conversations with PHA staff.
2 Although it would be ideal to complement these data with ownership interviews, such triangulation is not practical given site visit scheduling.
3 HUD is operating the Community Choice Demonstration to learn whether and to what extent mobility-related services facilitate families with children in moving to and remaining in opportunity areas. The Demonstration is being implemented by nine sites, consisting of 13 PHAs.
4 Rates were computed by New York University’s Furman Center using HIP data from HUD. Some PHAs lack interpretable data because of unexpected discrepancies or omissions in their submitted HIP data.
5 This is the number of PHAs in HUD’s 2020 Picture of Subsidized Households dataset.
6 We used the same five regions as in HUD’s selection notice for the first cohort of the MTW Expansion: Northeast, Southeast, Midwest, Southwest, and West. For the states in each region, see PIH Notice 2018-17, available at https://www.hud.gov/sites/dfiles/PIH/documents/PIH-2018-17MTWDemonstrationProgram.pdf
7 The key characteristics were selected from among 90 possible characteristics. We examined (a) 31 characteristics about the housing market including vacancy rates, median rent, percent of households with own children, rent as a percent of income, and various categories/universes for these variables (all renters, all homeowners, both), (b) 56 variables from the Picture of Subsidized Households about the voucher program, (c) and three baseline measures of outcome variables including success rate, HUD expenditure, and median search duration.
| File Type | application/vnd.openxmlformats-officedocument.wordprocessingml.document |
| File Title | Abt Single-Sided Body Template |
| Author | Katheleen Linton |
| File Modified | 0000-00-00 |
| File Created | 2024-12-13 |