Exploring the Social Construction of Crime by Neighborhood: News Coverage of Crime in Boston

ABSTRACT

Information disseminated by the news shapes the way that the public perceives criminal events, often providing a distorted view of crime. Previous research has largely overlooked neighborhoods in discussions of how the news portrays crime. This study examines the ways that the news media report the neighborhoods in which homicides, robberies, and assaults are committed. Multiple theoretical perspectives rooted in the law of opposites and racial typification provide differing explanations for the reporting of crime. Using Boston as a test site, this study employs a content analysis of The Boston Globe crime articles to identify the neighborhoods in which instances of homicide, robbery, and assault receive coverage. A comparison with official crime data from the Boston Police Department suggests differences in neighborhood reporting trends for robbery and assault but not for homicide. Specifically, the news media tend to disproportionately report more robberies and assaults in neighborhoods with lower levels of neighborhood disadvantage. Implications for the social construction of crime and neighborhoods as well as criminal justice response for disadvantaged neighborhoods are discussed.

Disclosure statement

No potential conflict of interest was reported by the author.

Notes

¹ The rarity perspective has also been referred to as statistical deviance (Pritchard and Hughes 1997) and relative frequency (Lundman 2003).

² This trend was crucial to news article selection in order to ensure that articles from a time of rising violent crime, a peak of violent crime, and a decline in violent crime were selected; thus, the articles in this story represent all three conditions in violent crime trends. While yearly trends are not given much consideration in criminological scholarship due to potential fluctuations over shorter time periods, they do matter for media reporting. It is common for news articles to compare current crime counts to their levels a year prior. Reporting crime as rising helps to satisfy the “risk” criteria of newsworthiness (Jewkes 2015), and the media tend to frame crime in this way when possible. This study considers such yearly trends in order to reduce the impact that yearly crime fluctuations have as confounding factors in the media’s decision to report on specific crimes. The 2005 to 2008 time period was the only recent time period in which trends were exhibited for all three crimes examined in this study.

³ These keywords were used because they pertain to the overarching categorization of an incident. For example, a newspaper story that describes a “mugging” would be categorized with the keyword “robbery” in the archive. Similarly, “homicide” falls under the category of “murders and murder attempts.” This was confirmed by using multiple search terms for a single crime and obtaining the same useable stories.

⁴ The term “murder attempts” was included in this search term because the categorical keyword utilized by The Boston Globe for murders is “murders and murder attempts.”

⁵ Due to a very small number of stories reporting rape, an examination of rape is not conducted in the current study.

⁶ Examples of excluded stories include editorials on punishment and legislative action, stories on activism taken against a particular crime, and articles on police activity that do not report a specific crime (crackdowns, community policing efforts, etc.).

⁷ Such follow-up stories potentially play a role in shaping the reputation of a neighborhood, especially if crime stories for a particular neighborhood receive more attention than others. However, as the interest of this paper is the comparison of newspaper crime reports to official crime reports, it was necessary to eliminate follow-up stories to better capture the basic proportion of stories that received any coverage. Police data will only report a given crime once; inclusion of one crime incident multiple times would skew the newspaper data and make it unfit for comparison. Additional analyses that consider follow-up stories are discussed in the Results section.

⁸ In the case that a story contained at least one initial report of a crime along with accounts of previously reported crime, only the initial report was coded in the sample. This occurred rather frequently, as it is common for a news story to make an initial report of a crime then evaluate it based on similar crimes that had occurred recently or in the proximate geographical area.

⁹ The specific address of a crime was rarely reported in the newspaper article. It was far more common for the newspaper to report the street on which the crime occurred along with the nearest intersecting street. From this information, it was possible to pinpoint more exact locations that allowed for determination of the neighborhood. Reported information such as nearby schools or public landmarks was also used to determine or confirm the neighborhood of the crime. In the event of a discrepancy with reported neighborhood, the neighborhood as identified by the objective determinants of address or landmark was used in this analysis.

¹⁰ The cases in which the neighborhood was not explicitly identified by the newspaper were included in the analysis, as specific geographic information relating to the location of the crime was cited in the news articles. It stands to reason that a local news audience would be able to identify the neighborhood in which the crime occurred based off of this information. I ran additional models that excluded the cases in which the newspaper did not explicitly identify the neighborhood; there were no substantive differences in the results of these models compared to the models cited in the current study.

¹¹ A key difference between newspapers and other types of news media is that newspapers can offer much more detailed and in-depth coverage of an incident, especially compared to the TV news format which is limited by time and footage considerations. The greater leeway in a newspaper’s ability to select incidents to cover may result in more muted trends compared to other news formats; as such, the findings of the current study can be interpreted as conservative and possibly indicative of more pronounced trends in other news formats.

¹² In the instances where the maps disagree, neighborhoods were determined by what the majority of the three sources indicated as neighborhood boundaries. In no situation did all three disagree on a particular boundary.

¹³ A limitation of the current study is that it does not examine the nature and placement of crime articles in the newspaper. It is possible that follow-up stories on a given incident are particularly prone to receiving less-detailed coverage or less-prominent placement—factors that may influence the impact of follow-up articles in shaping the social construction of neighborhood crime.

Additional information

Funding

Funding for this project was provided by the Northeastern University School of Criminology and Criminal Justice Summer Research Scholarship.

Notes on contributors

Andrew J. Baranauskas

Andrew J. Baranauskas is an assistant professor in the Department of Criminal Justice at The College of Brockport, State University of New York. His research interests include the social construction of crime by the media, violent crime in urban neighborhoods, perceptions of urban neighborhoods, and public opinion about crime and justice. His work has been published in Criminology. He was named a doctoral fellow of the Academy of Criminal Justice Sciences in 2018.