Local Level Victimization Survey, Battle Creek, Michigan, 2016-2017 (ICPSR 39077)

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 purpose of these surveys, which were previously conducted in Saginaw, MI in 2015 and Detroit, MI in 2016, was two-fold. First, while official data on crime patterns is largely available across the United States, few city level victimization surveys are conducted on a routine basis. This leaves open the possibility that crimes unreported and undocumented by the police are unaccounted for, with potential negative consequences on public responses to safety issues. Second, these studies examined whether or not web-based survey methods could be used to gather victimization data in a manner consistent with what has been accomplished using landline telephones. Understanding the ability of alternative methods of data collection, beyond landline telephone surveys, is important given the decline in landline telephone usage in communities throughout the United States.

Study Design View help for Study Design

The study design for Battle Creek, MI built upon the results of two previous city-level surveys in Saginaw and Detroit, MI, respectively. The current study sampled household addresses in Battle Creek, MI across two income strata, labeled low and high income. Low-income households were those at 200% of the poverty level and below based on United States Census Bureau standards for 2017. Low-income households were oversampled at a rate of 61% with 3,100 prenotification letters sent out to low-income households and 2,000 such letters sent to high-income households. All prenotification letters were sent to households on August 3rd, 2017. Prenotification letters described an incentive associated with completing the survey by August 15, 2017.

To determine whether or not respondents could be influenced to use a web-based modality (i.e., computer, tablet, smartphone) to complete the survey, half the sample in each of the income strata received (at random) one of two prenotification letters. The difference between the two letters is the explicit description of a paper survey as an optional survey mode; while the first condition makes no mention of a paper survey, the second describes it on two occasions. The purpose of this experimental condition was to determine whether the knowledge that a paper survey was possible would influence the respondent's choice on which survey mode they would choose.

Respondents in both conditions were provided a paper and pencil survey, along with a self-addressed stamped envelope. All respondents were provided the same conditions for survey completion following the prenotification letter.

Sample View help for Sample

Address-based random sample of households in Battle Creek, MI. In total, victimization data were gathered about 2,145 adults living in Battle Creek. This is an average of 1.84 adults per household in the survey.

Time Method View help for Time Method

Universe View help for Universe

Adults living in Battle Creek, Michigan

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

This study used a modified version of the American Crime Survey Incident Level Questionnaire. The survey was modified in order to gather information on local community-level concerns with issues such as procedural justice, police-community relations, and fear and perceived risk of victimization. This included items on respondent perceptions of their fellow community members' willingness to call the police to report on an accident or crime, as well as their willingness to provide information to the police. The survey also asked about the prevalence and quality of interactions respondents had with the police in the previous 12 months.

One adult per cooperating household served as the respondent for their household; this person was asked to report data about themselves as well as about as many as four other adults in their household.

Response Rates View help for Response Rates

A response rate of 26.9% yielded N=1167 complete surveys. 375 respondents (32.1%) completed the survey using a web-based platform, 94 respondents (8.1%) used the call-in phone option, and 698 respondents (59.8%) used the paper survey.

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 survey data was to attempt to "correct" for aspects of the survey design and implementation that could have biased the unweighted data. Population parameters for weighting were obtained from American Community Survey (ACS) data gathered by the U.S. Bureau of the Census.

Three characteristics were used for household-level weighting: geographic area (number of adults per each of the two geographic strata), household occupancy status (number of households that are owner- versus renter-occupied), and household income (number of households that fall into each of 10 annual income categories, ranging from "less than $10,000" to "$200,000 or more"). For person-level weighting, five characteristics were used: race, education, age, sex, and Hispanicity.

For additional information on the weighting strategy for this study, please see the Weighting Report included in the study documentation.

Hide