Methodology

As noted earlier, the runaway/thrownaway estimates in this Bulletin are based on three components of the NISMART–2 studies: the National Household Survey of Adult Caretakers, the National Household Survey of Youth, and the Juvenile Facilities Study.

The National Household Surveys

The National Household Survey of Adult Caretakers and the National Household Survey of Youth were conducted during 1999. Both surveys used computer-assisted telephone interviewing methodology to collect information from a national probability sample of households. A total of 16,111 interviews were completed with an adult primary caretaker, resulting in an 80-percent cooperation rate among eligible households with children and a 61-percent response rate. Each primary caretaker who completed an interview was asked for permission to interview one randomly selected youth in the household ages 10–18. Permission was granted to interview 60 percent of the selected youth, yielding 5,015 youth interviews and a 95-percent response rate for the youth for whom permission was granted.

A set of 17 screening questions was used to determine the eligibility of both the adult caretakers and youth for an indepth followup interview designed to collect detailed information about each type of episode reported in the screening interview. Responses to the following seven episode screening questions steered an adult respondent into a runaway/thrownaway followup interview:²

• In the last year, did [this child/any of these children] leave home without permission and stay away for at least a few hours?
  
  
• Did [this child/any of these children] stay away for at least one night?
  
  
• Did [this child/any of these children] choose not to come home from somewhere when [he/she/they] [was/were] supposed to, and stay away for at least two nights?
  
  
• Did you or any adult member of your household force or tell [this child/any of these children] to leave home, or decide not to allow [him/her/them] back in the home?
  
  
• Did [this child/any of these children] leave for at least one night?
  
  
• Was there any time when having [this child/any of these children] in your home became a lot of trouble and [he/she/they] left?
  
  
• Other than anything you have already told me about, has there been any time, either currently or during the past 12 months, when you did not know where [this child/any of these children] [was/were] living?

The total number of children included in the adult Household Survey sample was 31,787. The total number of youth included in the youth Household Survey sample was 5,015. The adult and youth Household Survey data were weighted to reflect the U.S. Census-based population of children.³

One obvious limitation to the Household Surveys is that they may have undercounted children who experienced episodes but were living in households without telephones or were not living in households during the study period, including street children and homeless families. Although these are not large populations, they may be at risk for episodes.

There is an additional group of children who should be included in the runaway/thrownaway category, but who may not have been well counted by the Household Surveys or Juvenile Facilities Study: permanently abandoned children. Because of the stigma attached to abandonment, it is unlikely that caretakers would disclose such children to interviewers. A special analysis of data from the Third National Incidence Study of Child Abuse and Neglect (NIS–3), obtained through a different methodology, estimated the approximate size of this group to be 56,900.⁴

The Juvenile Facilities Study

Although the majority of runaway/thrownaway children were expected to be living in households at the time of the episode, a considerable number of youth also run away from facilities such as group homes and youth detention centers where they may be living. Therefore, NISMART–2 conducted a national survey of facilities where juveniles reside to obtain information on runaways/thrownaways from such facilities. The information obtained from the Juvenile Facilities Study enabled NISMART–2 to avoid double-counting youth who ran away from both a household and a juvenile facility in the course of the study year.

The respondents in the Juvenile Facilities Study were staff from a nationally representative sample of 74 facilities, including juvenile detention centers, group homes, residential treatment centers, and shelters for runaway and homeless youth. The study used a stratified, two-stage sample. In the first stage, 33 counties were sampled from the universe of U.S. counties. In the second stage, juvenile facilities were sampled within the selected counties.⁵

Respondents at each of the selected facilities were contacted by telephone and interviewed to determine the number of children who ran away from each facility in 1997, and details were obtained about the five most recent runaway episodes. All of the selected facilities that were operational participated in the study, providing facility-level information and yielding a 93-percent response rate for the targeted episode-level interviews. The Juvenile Facilities Study runaways were assigned weights to reflect the probability of having included the facility and episode in the sample and to adjust for nonresponse.

Unified Estimates

The runaway/thrownaway estimates reported in this Bulletin are unified estimates that combine the numbers of countable runaway/thrownaway children reported in each of the component studies discussed above. Any single child was allowed to count in the unified estimate only once, even if a countable runaway/thrownaway episode was reported for the same child in both the adult and youth interviews, or the child ran away from both a household and a facility in the study year. A detailed description of the unified estimate methodology is provided in OJJDP’s forthcoming Unified Estimate Methodology Technical Report.

All estimates in the results reported below have been rounded to the nearest 100. As a result of the rounding, the percentages may not sum to 100.