Local Level Victimization Survey, Saginaw, Michigan, 2014-2015 (ICPSR 39079)

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

Hide

Study Purpose View help for Study Purpose

The goal of the Local Level Victimization Survey (LLVS) is to gain an understanding of the experiences and perspectives of local residents regarding crime and justice. To this end, the Michigan Justice Statistics Center located in the School of Criminal Justice at Michigan State University was awarded a grant by the U.S. Bureau of Justice Statistics (BJS) to conduct a survey in Saginaw, Michigan.

In addition to understanding victimization and perceptions of residents in high crime areas, the survey was designed to test different methods for conducting surveys in these areas. The researchers argued that given the decline in landline phones and expenses associated with mailing paper surveys, the feasibility of using web-based surveys needed to be tested. Thus, the goal of the survey was twofold. First, to test the feasibility of a city-level victimization survey in high crime cities. Second, to see whether or not web-based surveys offer a viable alternative to landline phones and mail as mediums for collecting survey data.

Following Saginaw, the Michigan Justice Statistics Center conducted similar surveys in Detroit, Michigan and Battle Creek, Michigan. These cites were chosen for several reasons. First, the research was conducted in cities that have experienced economic decline, loss of police resources, and high rates of violent crime. Second, two of the three cities Detroit and Saginaw, along with Flint, Michigan, represented three of the original State of Michigan's "Secure Cities". Secure Cities was a partnership between the Governor's Office, Michigan State Police, and localities to address safety in cities with high levels of crime. The surveys were intended to help inform the State's crime reduction strategies. Third, Battle Creek was a non-Secure Cities location chosen to see if there were systematic differences in response rates and patterns in a city with less crime compared to Detroit and Saginaw.

Study Design View help for Study Design

The surveys administered were a modified version of the American Crime Survey Incident Level Questionnaire. The decision to use this survey was suggested by BJS as an efficient way to gather data, since it could be compared with other locations who used the instrument. The survey was modified to include measures of procedural justice, police-community relations, and fear and perceived risk of victimization. The survey was administered to the eastern majority Black and western majority White sections of Saginaw as described in the Sampling section below. Please refer to appendix B of the report for the survey instrument.

Sample View help for Sample

A multi-mode, address-based sampling approach was used for the survey. The Saginaw River bisects the city, with a predominately White and more affluent population on the west side, and a majority Black and relatively disadvantaged population to the east. The goal was to obtain 450 surveys from each side of the city. The eastern side was slightly oversampled due to prior research indicating that survey completion rates tend to be lower in economically disadvantaged neighborhoods. A total of 2,525 letters were distributed to the east yielding 373 completed surveys, while 2,000 letters were sent to the west side yielding 456 completing surveys.

Time Method View help for Time Method

Universe View help for Universe

Adults living in Saginaw, Michigan during 2014 and 2015.

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 Weighted Dataset includes variables capturing the mode of data collection. Measures of community satisfaction, perceived risk, procedural justice, fear of victimization, police legitimacy, collective efficacy, and experiences of fear are captured. Modules documenting multiple violent crime and property crime incidents are included as well. Numerous weight variables relating to homeownership, age, race, area, education, and income were calculated by the PI. Demographic variables for residency, age, gender, race, and education are included.

The Unweighted Dataset is similar to the Weighted Dataset except for two notable differences. First, the weight variables aren't included. Second, open-ended string variables are included which allow respondents to elaborate on their answers to select variables. There's also a variable COMMENTS which captures all additional comments from respondents that they filled out on the back of their survey.

Response Rates View help for Response Rates

The minimum response rate was 22%.

Presence of Common Scales View help for Presence of Common Scales

American Crime Survey Incident Level Questionnaire

Hide

Original Release Date View help for Original Release Date

2025-09-30

Version History View help for Version History

Hide

Weight View help for Weight

The goal of weighting the dataset was to attempt to correct for aspects of the survey design that could have biased the unweight data. The weight dataset includes variables to conduct person and household level weighting listed at the end of the dataset. These variables include WTQ27, WTAGE4, BLACKWHITE, WTRACE, WTAREA, WTEDUC, WTINCOME, AREAWT, Q27WT, TEMPWT1, INCOMEWT, TEMPWT2, and FINALHHWT. Population parameters for household-level weighting were obtained using the American Community Survey (ACS) gathered by the U.S. Bureau of the Census. Three characteristics were used to for household weighting: geographic area based on census block group especially as it relates to the east and west sides of the Saginaw River, household occupancy status related to owner versus renter occupied, and household income data showing the number of households falling into 16 annual income categories between less than $10,000 and $200,000 or more. Person-level weighting was based on ACS population parameters for five characteristics: race, education, age, sex, and Hispanic. For additional information on the weighting strategy, please refer to Weighting Report within Appendix C of the Technical Report.

Hide