Housing and Children's Healthy Development Study (HCHD) Wave 1, Cleveland, Ohio, and Dallas, Texas Metropolitan Areas, 2017-2018 (ICPSR 39274)

Time Period(s) View help for Time Period(s)

Date of Collection View help for Date of Collection

Data Collection Notes View help for Data Collection Notes

Study Purpose View help for Study Purpose

The goal of the Housing and Children's Healthy Development (HCHD) Study was to better understand how children's residential context--home, neighborhood, school, and family--affects their healthy development. The study was designed to address these policy research questions:

1. What are the effects of housing on children net of the other important influences on children's lives, including their families, neighborhoods, and schools?
2. What features of housing matter most?
3. For whom and in what circumstances does housing matter?
4. How do families with young children make housing, neighborhood, and school choices; what are the effects of these choices; and how would these effects change if their choices changed?

Study Design View help for Study Design

The Housing and Children's Healthy Development (HCHD) used a longitudinal, randomized control trial design. To select study sites, the study team reviewed characteristics of all U.S. public housing authorities (PHA) where voucher offers relied on randomization. Additional selection criteria were variation in geographic location, housing market characteristics, and racial and ethnic mix of the metropolitan area population; high performance based on Housing and Urban Development's (HUD) assessment of PHA management and reputation in the field; and commitment to participating in the study.

Three separate samples were drawn from each area:

1. Applicants for housing vouchers likely to be selected to receive housing voucher (the treatment group);
2. Applicants unlikely to receive housing voucher (the control group); and
3. General population.

Once the voucher and general population samples were selected, they were contacted by the Survey Research Center (SRC) at the University of Michigan to determine eligibility in the study based on the presence of a child within the age range and current residence in the greater Cleveland (Cuyahoga County) or Dallas areas. SRC interviewers attempted to screen and interview all households prior to any notification from the public housing authorities as to whether or not respondents had received a housing voucher.

Wave 1 interviews were conducted in-person in 2017-2018 before families in the voucher treatment group received a voucher. SRC deployed trained interviewers to conduct interviews, typically in the primary caregiver's home. Interviewers led personal interviews with primary caregivers using computer-assisted personal interviewing (CAPI) and collected physical measures from both primary caregivers and children, including dried blood spots for the voucher sample only. Primary caregivers completed a self-administered questionnaire. Children were administered standardized tests of reading and math achievement (Woodcock-Johnson) and a computerized task evaluating executive functioning (Hearts and Flowers). Interviewers also systematically observed the home environment, the neighborhood environment, and parent-child interactions. Interviewers also provided participants with diaries to complete with information about the focal child's day and mailed back to the study team.

Sample View help for Sample

Wave 1 data were collected from 1,801 primary caregivers and 2,396 children. The overall sample was designed to be a representative sample of the greater Cleveland (Cuyahoga County) and greater Dallas areas. The Cuyahoga Metropolitan Housing Authority and Dallas Housing Authority each drew a random sample from the treatment and control groups. Both the treatment and control households were expected to meet the voucher applicant criteria.

The Cuyahoga Metropolitan Housing Authority used its 2015 waiting list to draw a voucher experimental sample. The initial announcement generated more than 50,000 applicants. After removing duplicate cases, the list was randomized and the top 10,000 applicants were placed on an active voucher waiting list. Eligible applicants were offered the option of study participation. Of the 9,429 remaining waitlist cases in March 2017, 2,894 met the criteria. After opt-outs (29 percent of applicants), the treatment group was selected from 2,041 cases. The control group was selected from cases not put on the active waiting list and was also randomized. 8,138 applicants met the child criterion and did not opt-out.

The Dallas Housing Authority established a new waiting list for the study. The list was advertised between April 6-18, 2017 and received roughly 14,000 applications during open enrollment period (April 18-24, 2017). After removing duplicate and ineligible cases, the waiting list contained 8,436 cases. These cases were randomized and the first 5,000 applicants placed onto an active waiting list. With 35 percent of the waiting list opting out of the study, the treatment group was selected from the remaining 3,263. Applicants not selected for the waiting list were randomized into the control group.

For the general population sample, the Survey Research Center (SRC) at the University of Michigan randomly selected a stratified population sample from the Cleveland (Cuyahoga County) and Dallas metropolitan areas designed to geographically overlap with the voucher applicant population. Using commercially available data, households were stratified based on expected household income, oversampling lower incomes. As was the case for the voucher population, eligibility was based on the presence of a child aged 3-10 (both years inclusive) and a primary caregiver living in the household.

For further details on sample design, please refer to the sampling and weighting methodology documentation provided by the study team.

Time Method View help for Time Method

Universe View help for Universe

Primary caregivers and children aged 3-10 years living in the Cleveland (Cuyahoga County) and Dallas metropolitan areas.

Unit(s) of Observation View help for Unit(s) of Observation

Data Type(s) View help for Data Type(s)

Mode of Data Collection View help for Mode of Data Collection

Description of Variables View help for Description of Variables

The variable CASEID can be used to link all datasets.

Primary caregiver interview. The main interview covered the following topics:

• Demographics (e.g., race, ethnicity, educational background, employment)
• Household roster
• Housing characteristics and quality (e.g., number of rooms, spacing, reliability of utilities, security, hazards)
• Housing costs
• Considerations in choosing a residence
• Household income, assets, and debts
• Child/Primary caregiver physical and mental health, including physical measures (height, weight, blood pressure, blood spot collection)
• Child's behavior
• Family environment and routines

Self-administered questionnaires. Primary caregiver participants answered questions about what they liked/disliked about their current housing, characteristics of their home (e.g., noise level, reliability of utilities), neighborhood safety, housing preferences, and satisfaction with parenting. Voucher applicants were asked additional items about their voucher status. Primary caregivers also answered a questionnaire about their child, providing information on their child's room in the home, places their child has lived, and disciplinary measures used against their child.

Post-interview observations. Interviewer completed these observations about the home: housing type and structure, physical environment (e.g., crowding, number of rooms, noise levels, available space for play), home ownership, signs of alcohol/drug use, overall organization. Interviewer also rated the focal child's behavior during the cognitive assessments (e.g., was the child on-task, distracted, destructive, patient, cooperative, engaged, etc.) and parent-child interactions during the research visit.

Laser tape measurements. If allowed access, interviewer took length and width measurements (feet/inches) of every room in the house, noting if a child respondent slept in the room.

Neighborhood observations. Interviewer recorded information about traffic volume, street condition, building conditions, primary land use, and if any of the following characteristics were present: window bars/gratings on doors or windows, abandoned buildings, vacant lots, bars/liquor stores, graffiti, garbage/little, drug/alcohol paraphernalia, abandoned cars, trees, or neighborhood crime watch signs.

Open-ended responses. The research team used a Blaise-to-SAS program to separate close-ended responses from open-ended responses. A "1 (Yes, there is an answer)" will appear in the main data files (DS1, 5, 8, 11, 14, 17, 20, 23, 26) if there is a response provided for that variable in the "Open-Ended Responses" data files (DS2, 6, 9, 12, 15, 18, 21, 24, 27).

Interviewer remarks. The "Interviewer Remarks" data files (DS3, 7, 10, 13, 16, 19, 22, 25, 28) contain supplemental remarks that are entered by the interviewer. They correspond to the variable indicated from the main data files, but they do not replace data in those variables. Rather, they are contextual notes that the interviewer entered about the response provided by the respondent, especially if the response could not be coded.

Response Rates View help for Response Rates

• Overall: 79 percent
• Voucher sample: 82 percent
• Population sample: 72 percent
• Biomarker collection: 99.4 percent for caregivers, 83 percent for children
• Laser tape measurements: 88 percent

Presence of Common Scales View help for Presence of Common Scales

• Woodcock-Johnson Letter-Word
• Woodcock-Johnson Applied Problems
• Caldwell-Bradley Home Observation for Measurement of the Environment (HOME)
• Maternal Cognitive Sensitivity
• Patient Reported Outcomes Measurement Information System (PROMIS)
• Parent-Child Conflict Tactics Scale
• Behavior Problems Index

Original Release Date View help for Original Release Date

2026-05-04

Version History View help for Version History

Weight View help for Weight

Dataset 29 in this collection contains a final weight variable, FINAL_WGT. Please refer to the sampling and weighting methodology documentation provided by the study team for further details on calculation and use.