NISMART Fact Sheet: National Incidence Studies of Missing, Abducted, Runaway, and Throwaway Children

October 2002

The Second National Incidence Studies of Missing, Abducted, Runaway, and Thrownaway Children (NISMART–2) is an important resource. The following questions and answers provide a quick overview of NISMART’s purpose, methodology, and findings. For a more detailed discussion of NISMART–2, refer to the corresponding series of Bulletins.

What is NISMART?

NISMART stands for the National Incidence Studies of Missing, Abducted, Runaway, and Thrownaway Children. These studies were undertaken in response to the Missing Children’s Assistance Act (Pub. L. 98–473), which requires that the Office of Juvenile Justice and Delinquency Prevention (OJJDP) conduct periodic studies to determine the number of U.S. children reported missing and the number of children recovered during a given year. NISMART consists of several complementary studies designed to estimate the size and nature of the Nation’s missing children problem. NISMART–2, the second such set of studies (the first, NISMART–1, was conducted in 1988), includes a large national survey (more than 16,000 households) of parents and other primary caretakers who were interviewed about their children’s experiences. The household survey also interviewed a sample of 5,000 youth ages 10–18, an important methodological improvement over the NISMART–1 design. To record the experiences of youth who had run away from residential placements such as group homes, a survey of juvenile facilities was also conducted. And because law enforcement agencies are particularly knowledgeable about the most serious and rarest cases of abduction—stereotypical kidnappings by strangers—a large-scale survey of police departments was conducted to gather detailed information about the characteristics of these crimes.

Have abductions by strangers declined or increased?

Although the number of stereotypical kidnappings by strangers reported by NISMART–1 (200–300) and NISMART–2 (115) appears to reflect a decline in such abductions, these figures are derived from studies that used very different methodologies. For example, NISMART–1 researchers studied police records from a sample of 83 law enforcement agencies.

For NISMART–2, the sample was expanded to more than 4,000 agencies, and data were collected from police personnel who investigated the cases. Because of the different methods used and the rarity of such cases, no scientific basis exists to conclude that there has been a true decline—although it is possible. On the other hand, NISMART–2 results do not indicate an increase in abductions by strangers.

Why can’t I compare NISMART–1 statistics with NISMART–2 statistics?

In planning NISMART–2, OJJDP convened a panel of experts to review NISMART–1 and to suggest design improvements to the studies. As a result, substantial refinements were made to the episode definitions and data collection methods. For example, many of the 354,100 NISMART–1 “broad scope family abductions” were viewed as fairly minor custodial and visitation interference episodes that did not warrant the designation of “abduction.” The NISMART–2 definition of “family abduction” was clarified to make the category more meaningful; as a result, the 203,900 family abductions recorded by NISMART–2 do not correspond to the family abductions captured by NISMART–1. In addition, because distinguishing “runaway” and “thrownaway” categories of children in NISMART–1 was difficult, the two types of episodes were combined into one category in NISMART–2 and, more importantly, youth were interviewed in NISMART–2 but not in NISMART–1. By interviewing youth directly, researchers identified many episodes that were either unknown to, or known but unreported by their caretakers.

Because the NISMART–2 design differs substantially from that of NISMART–1, initial NISMART–2 reports will focus exclusively on findings resulting from improved definitions, methods, and terminology. However, it should be noted that NISMART–2 was also designed to look at historical trends by comparing the two study periods using the most equivalent definitions and methods with the NISMART–2 samples. Those results will be available in a separate publication later this year.

When were the data for NISMART–2 collected?

NISMART–2 studies spanned 1997–99, and all the data for each component study were collected to reflect a 12-month period. Because most of the cases studied were concentrated in 1999, the annual period being referenced in NISMART–2 is 1999.

The reference dates for some of the NISMART–2 component studies vary because of a delay caused by pending Federal legislation that, had it passed, would have made it impossible to conduct the National Household Survey of Youth, a key component of NISMART–2. In anticipation of a quick resolution, OJJDP decided to proceed with the Law Enforcement Study and the Juvenile Facilities Study because neither involved interviewing youth. Had these 1997 studies been postponed until 1999, it is highly unlikely that those estimates would have been statistically different.¹ For the sake of simplicity, all NISMART–2 results refer to the annual period of 1999.

When will the study results from NISMART–2 be released?

OJJDP has published the initial findings in four NISMART Bulletins covering the following topics: unified estimates of missing children, family abductions, nonfamily abductions, and runaway/thrownaway children. Additional findings on the remaining NISMART–2 episode types and on topics such as sexual assault and changes between NISMART–1 and NISMART–2 will be released through early 2003. All NISMART-related documents (e.g., Bulletins, Fact Sheets) will be available on OJJDP’s Web site, ojjdp.ncjrs.org.

1 To illustrate this point, the 95-percent confidence interval for the estimated 115 kidnappings reported in NISMART–2 indicates that if the Law Enforcement Study were to be repeated with the same methodology 100 times, 95 of the replications would produce an estimated 60–170 stereotypical kidnappings. This means that, using a similar methodology to detect a real increase in the number of such cases occurring between 1997 and 1999 or later, the estimated number of stereotypical kidnappings would have to be greater than 210. Such an increase is very unlikely, even in light of the number of high-profile cases that have recently received national attention.