Overview of DNORS Survey Design
The DNORS sample frame covers the City of New Orleans, Louisiana, which coincides precisely with Orleans Parish. Data collection was carried out in 149 area segments (Census block clusters formed to have a minimum number of housing units). Due to the hurricane damage inflicted in this county, substantial effort was mounted to account for and identify housing units that existed prior to Hurricane Katrina but were destroyed in its aftermath. The interviewers were trained to do a complete enumeration of housing units, beginning from a listing of all addresses supplied by the United States Postal Service for the segments. A total of 7,728 sample addresses were randomly selected for the DNORS sample. A detailed discussion of the DNORS survey design can be found in the DNORS Overview, User's Guide, and Codebook [PDF].
The study population consisted of families living in City of New Orleans/Orleans Parish, Louisiana in August 2005, immediately prior to any residential dislocation caused by Hurricane Katrina, and who were still alive and could be located either by telephone anywhere in the United States or by face-to-face interviewing, as necessary, in Orleans Parish.
DNORS collected information on pre-Katrina household members through a household roster module and additional detailed information from selected individual adult respondents. Interviews were conducted by telephone, with a small percentage of those still living in Orleans Parish interviewed face-to-face if a telephone interview could not be done. The survey was fielded in the fifth year following Hurricane Katrina, between mid-2009 and mid-2010.
Within each of the selected households, interviewers would first collect a household listing of people who were living at the sampled address in August 2005, at the time of Hurricane Katrina. From this list, the screener interview determined who was eligible to be interviewed for the main interview. The screener collected location and contact information for the selected respondents.
Sample Design and Characteristics
DNORS had two types of respondents: Adult Respondents and Household Respondents. These respondents were drawn from the list of household members who were living at the sampled address in August 2005, at the time of Hurricane Katrina, and who were aged 18 years or older at the time of the first successful contact with the household (i.e., the screener interview).
The Adult Respondent was randomly selected from the list of those living adult pre-Katrina household members. The Adult Respondent was administered the individual interview section of the main interview but was not asked to provide detailed information on individuals in the roster of pre-Katrina household members. Female Adult Respondents with children under the age of 18 years were additionally administered the Child Module which collected information on up to three of her children under the age of 18 years.
The Household Respondent was randomly selected from the list of those living adult pre-Katrina household members who were identified as knowing the most about the status of pre-Katrina household members. Because the Adult Respondent was selected first, the Household Respondent could be the same person selected as the Adult Respondent. The Household Respondent was administered the household roster section that collected detailed information on each pre-Katrina household member. The Household Respondent was also given the individual interview section.
Household Survey Content
The DNORS questionnaires included a screener and a main interview that produced two types of interviews: an Adult interview (given to the Adult Respondent and the Household Respondent), and a Household interview with a person identified as knowledgeable regarding family and housing (the Household Respondent). In addition, the main interview contained a special child module with questions asked of female Adult Respondents with children under the age of 18 years.
Pre-Katrina Household Roster
The pre-Katrina household roster, based on the listing obtained in the Screener Interview, was completed by the Household Respondent. The roster (Section C of the Main interview) asked detailed questions about each pre-Katrina household member, ascertaining current status and location as well as residential changes immediately following Katrina plus some basic demographic information. Some detailed information, such as job-related questions, was only asked of the respondent.
When the Adult Respondent was not also the Household Respondent, the Adult Respondent was asked the same set of detailed questions but only about himself/herself and was not asked about any other pre-Katrina household members.
The individual interview (Sections D-M of the Main interview) was administered to the Adult Respondent and the Household Respondent, if they were different individuals. These questions covered traumatic experiences, health, mental illness, PTSD, depression, and perceived stress, as well as household income.
Female Adult Respondents with children under the age of 18 years were administered a series of questions for up to three children (randomly selected if she had more than three children under age 18 years) covering the child's father's involvement, and the child's background characteristics, health, and behavior problems. For selected children who were in the pre-Katrina household roster, questions in the Child Module that were the same as those asked in the roster section were not re-asked.