![Sample our Law journals, sign in here to start your FREE access for 14 days]({"type":"image" , "src" : "/sda/5942/TOC-Subject-Sample-2021-Law.jpg"})
Abstract
Recent media coverage in Baltimore, MD and Ferguson, MO on tensions between the police and public highlight the importance of understanding how criminal justice practices affect public opinions. High crime neighborhoods are central to this research. However, residents of these contexts are typically minority and low-income residents, who have a reduced propensity to participate in surveys. Additionally, high crime areas pose a challenge for interviewer safety and retention. Past research on in-person surveys in high crime contexts is limited and spans multiple social science fields, making identification of best practices more challenging for criminologists. We review best practices and theories, identify issues when surveying in high crime contexts, and describe our challenges and outcomes conducting face-to-face surveys in 71 hot spots in St. Louis County, MO. We highlight our use of student interviewers to increase the response rate and reduce costs, while adding value to students’ educational experiences and knowledge.
Disclosure statement
No potential conflict of interest was reported by the authors.
Notes
1 While our primary research project applied an experimental approach, we did not apply an experimental design to understanding the best methods for door-to-door surveying. Rather, we seek to grow the knowledge base about conducting face-to-face surveys in high crime areas by being transparent with our own experiences—challenges we faced, what we did, and the consequences. This is not unlike scholars in anthropology and sociology who have written about strategies that they have used in ethnographic research to be safe in dangerous contexts (e.g. Sluka, Citation1990; Williams, Dunlap, Johnson, & Hamid, Citation1992).
2 Legleye and colleagues (Citation2013) also suggest that persistence may improve responses and reduce socio-demographic bias sometimes produced as a result of non-response in difficult to reach populations.
3 Past research supports the capacity of advance letters to increase response or cooperation rates by as much as 8 (Leeuw, Callegaro, Hox, Korendijk, & Lensvelt-Mulders, Citation2007) to 27% (Iredell, Shaw, Howatt, James, & Granich, Citation2004).
4 Tatoos and features that may not work well in affluent or more conventional contexts can be assets in these contexts.
5 One reason for this dearth of scholarship may be that active researchers are protecting proprietary strategies that make them more competitive in funding environments.
6 The authors are grateful to Drs Justin Ready, Chris Koper, Craig Uchida, Josh Hinkle, and David Weisburd for sharing their advice and experiences with us. In the absence of much published research to guide our efforts, we found personal suggestions from those who have attempted interviews in high crime areas invaluable.
7 As vacancies were challenging to determine, especially in the apartment complexes, which composed two-thirds of the sample, the number of valid and eligible addresses is likely lower. Although the post office records for vacancies were helpful, as described below, we undoubtedly underestimate the vacancy rates. Thus, the response rates are conservative.
8 Ferguson and Mindel (Citation2007), who studied fear of crime, had a 33% response rate. In crime hot spots, Hinkle et al. (Citation2013) had a 46.1% cooperation rate (excluding contacts with language/cognitive barriers).
9 While we struggled with extreme weather at the second wave, we saw considerable improvement in the average time spent in the field per successful completion at the third wave compared to wave one. While we spent an average of 84 min in the field per completed survey at time one, we had reduced that time spent to 72 min in wave three.
10 Much like Legleye and colleagues’ (Citation2013) persistent approach to telephone survey non-contacts and refusals, we adopted an approach of callback attempts when residents were not home and even after receiving non-violent refusals to participate from at least one household member.
11 The postcard displayed the university logo and name. It stated the survey purpose, explaining the importance of residents providing their opinions about police, crime, and their neighborhoods. The postcard mentioned that the survey was federally funded by the National Institute of Justice, and provided the principal investigator’s name, university affiliation, contact number, and e-mail. Potential respondents were advised to expect surveyors to wear badges identifying them as members of the research team. Residents also were given a timeframe within which to expect to be approached by SIUC interviewers.
12 Identified vacancy rates across the 71 hot spots ranged from zero to 55%, with an average of 17%. Vacancies were more easily identified and coded by interviewers in single family areas, where boarded up windows and doors, notices on doors, and empty rooms were readily visible to interviewers. The postal notification about vacancies in apartment complexes was particularly valuable. Apartments that are vacant do not offer the same visual cues.
13 This is supported by past research (indeed Gwiasda and colleagues (Citation1997) used more than three times as many female interviewers as males).
14 The principal investigator and a fellow professor who served as a consultant to the project trained interviewers and the field research coordinator at wave one.
15 The St. Louis area has a number of colleges and universities, and we were able to solicit help from a network of faculty members to identify promising interviewers.
16 Average cost per survey conducted by paid interviewers was about $22.
