Crime and Safety Surveys: School Crime Supplement (SCS)

SURVEY DESIGN

4. SAMPLE DESIGN

Households are selected into the sample using a stratified, multistage cluster design. In the first stage, the primary sampling units (PSUs), consisting of counties or groups of counties, are selected and smaller areas, called Enumeration Districts (EDs), are selected within each sampled PSU. Large PSUs are included in the sample automatically and are considered to be self-representing strata since all of them are selected. The remaining PSUs (called non-self representing because only a subset of PSU is selected) are combined into strata by grouping PSUs with similar geographic and demographic characteristics, as determined by the decennial census. Within each ED, clusters of four households, called segments, are selected. Across all EDs, sampled households are then divided into discrete groups (rotations), and all age-eligible individuals in the households become part of the panel. Such a design ensures a self-weighting probability sample of housing units and group-quarter dwellings within each of the selected areas. "Self-weighting" means that prior to any weighting adjustments, each sample housing unit had the same overall probability of being selected.

Each month the U.S. Census Bureau selects respondents for the NCVS using a "rotating panel" design. Households are randomly selected and all age-eligible individuals become part of the panel. The sample of households is divided into groups or rotations. Once in the sample, respondents are interviewed every six months for a total of seven interviews over a three-year period. The first interview is considered the incoming rotation. The second through the seventh interview are in the continuing rotations. The first interview is face-to-face; the rest are by telephone unless the circumstances call for an in-person interview. After the seventh interview the household leaves the panel and a new household is rotated into the sample. The rotation scheme is used to reduce respondent burden that may result if they were to remain permanently in the sample.

Once in the panel, NCVS interviews are conducted with all household members age 12 or older. After completion of the NCVS interview, an SCS interview is given to eligible household members. In order to be eligible for the SCS, students must be 12 through 18 years old, have attended school in grades 6 through 12 at some point during the school year, and not have been homeschooled during the school year. Persons who have dropped out of school, have been expelled or suspended from school, or are temporarily absent from school for any other reason, such as illness or vacation, are eligible as long as they attended school at any time during the school year. For the 1989 and 1995 SCS, 19-year-old household members were considered eligible for the SCS interview. Prior to the 2007 SCS, household members who were enrolled in school sometime during the previous 6 months prior to the interview were eligible.

Top

Data Collection and Processing

In all SCS survey years, the SCS was conducted for a 6-month period from January through June in all households selected for the NCVS. Eligible respondents were asked the supplemental questions in the SCS only after completing their entire NCVS interview

The 2007 SCS was fully automated; all interviews were conducted through computer-assisted personal interviewing (CAPI), where field representatives used questionnaires loaded into laptop computers to conduct interviews, which could be completed either in person (for the first and subsequent interviews, as circumstances called for) or by telephone. Two modes of data collection were used through the 2005 collection: (1) paper-and-pencil interviewing, which was conducted in person for the first NCVS/SCS interview; and (2) computer-assisted telephone interviewing (CATI), unless circumstances called for an in-person interview. There were approximately 7,146 students who participated in the SCS in 2017; 4,770 in 2015; 5,700 in 2013; 6,550 in 2011; 5,020 in 2009; 6,500 in 2007; 7,110 in 2005; 8,470 in 2003; 9,650 in 2001; 8,400 in 1999; 9,950 in 1995; and 10,450 in 1989.

Interviewers are instructed to conduct interviews in privacy unless respondents specifically agree to permit others to be present. Most interviews are conducted over the telephone, and most questions require "yes" or "no" answers, thereby affording respondents a further measure of privacy. While efforts are made to assure that interviews about student experiences at school are conducted with the students themselves, interviews with proxy respondents are accepted under certain circumstances. These include interviews scheduled with a child between the ages of 12 and 13 where parents refuse to allow an interview with the child; interviews where the subject child is unavailable during the period of data collection; and interviews where the child is physically or emotionally unable to answer for him- or herself.

Top

Estimation Methods

Weighting. The purpose of the SCS is to be able to make inferences about criminal victimization in the 12– to 18–year–old student population in the United States. Before such inferences can be drawn, it is important to adjust, or “weight,” the sample of students to ensure it is similar to the entire population in this age group. The SCS weights are a combination of household–level and person–level adjustment factors. In the NCVS, adjustments are made to account for both household– and person–level non-interviews. Additional factors are then applied to reduce the variance of the estimate by correcting for the differences between the sample distributions of age, race, and sex and the known population distributions of these characteristics. The resulting weights are assigned to all interviewed households and persons in the file.

A special weighting adjustment is then made for the SCS respondents, and non-interview adjustment factors are computed to adjust for SCS interview nonresponse. This non-interview factor is applied to the NCVS person–level weight for each SCS respondent. Through 2005, there was one SCS weight for producing estimates for the NCVS variables and another SCS weight for producing estimates from the SCS variables. Due to the inclusion of the incoming interview variable in the NCVS estimates, the same weight now applies to both.

Imputation. Item response rates are generally high. Most items are answered by over 95 percent of all eligible respondents. No explicit imputation procedure is used to correct for item nonresponse.

Future Plans

Plans for the future of the SCS include a 2022 administration. NCES and Census plan to use findings from the 2019 split-half experiment that tested bullying items to inform the method of collection for the 2022 administration.

 

Top