The public salience of crime, 1960–2014: Age–period–cohort and time–series analyses†

• Author: Luzi Shi,Yunmei Lu,Justin T. Pickett
• Date: 01 August 2020
• Published date: 01 August 2020
• DOI: http://doi.org/10.1111/1745-9125.12248

Received: 11 January 2019 Revised: 30 January 2020 Accepted: 9 April 2020

DOI: 10.1111/1745-9125.12248

ARTICLE

The public salience of crime, 1960–2014:

Age–period–cohort and time–series analyses*

Luzi Shi1Yunmei Lu2Justin T. Pickett3

1Department of Criminal Justice, Bridgewater

State University

2Department of Sociology, University at

Buffalo, SUNY

3School of Criminal Justice, University at

Albany, SUNY

Correspondence

LuziShi, Depar tment of Criminal Justice,

BridgewaterState University, 131 Summer

Street,Bridgewater, MA 02325

Email:lshi@albany.edu

∗Additionalsupporting information can

befound in the full text tab for this arti-

clein t he WileyOnline Library at http:

//onlinelibrary.wiley.com/doi/10.1111/

crim.2020.58.issue-3/issuetoc.

Theaut hors wouldlike to thank Dr. Peter K.

Ennsfor sharing his data on newspaper cover-

ageof crime and Dr.Mark Ramirez for sharing

hisdat a on presidential statements about pun-

ishment.The aut hors are also grateful to the

editorsand t he anonymousreviewers for their

helpfulfeedback and comments.

Abstract

The public salience of crime has wide-ranging political and

social implications; it influences public trust in the govern-

ment and citizens’ everyday routines and interactions, and

it may affect policy responsiveness to punitive attitudes.

Identifying the sources of crime salience is thus important.

Two competing theoretical models exist: the objectivist

model and the social constructionist model. According

to the first, crime salience is a function of the crime rate.

According to the second, crime salience is a function of

media coverage and political rhetoric, and trends in crime

salience differ across population subgroups as a result of

differences in their responsiveness to elite initiatives. In

both theories, period-level effects predominate. Variation

in crime salience, however, mayalso reflect age and cohort

effects. Using data from 422,504 respondents interviewed

between 1960 and 2014, we first examine the nature

of crime salience using hierarchical age–period–cohort

(HAPC) models and then analyze period-level predictors

using first differences. We find that 1) crime salience

varies mostly at the period level; 2) crime salience trends

are parallel (cointegrated) across demographic, socioeco-

nomic, and partisan groups; and 3) crime salience trends

within every population subgroup are most consistent with

the constructionist model. The crime rate does not exert a

significant effect in any subgroup.

KEYWORDS

crime, media, most important problem, parallel publics, public opinion

568 © 2020 American Society of Criminology wileyonlinelibrary.com/journal/crim Criminology.2020;58:568–593.

SHI ET AL.569

Understanding how the salience of crime to the public varies over time—that is, when and why mem-

bers of the public come to view crime as a pressing social problem—is important for at least three

reasons. First, shifts in the percentage of the public identifying crime as the country’s most impor-

tant problem (MIP) contribute to subsequent changes in aggregate support for punitive crime policies

(Ramirez, 2013), which in turn exert large effects on incarceration rates and criminal justice expen-

ditures, at both the state and federal levels (Enns, 2014, 2016). Put simply, crime salience indirectly

affects crime policy by shaping policy attitudes. Second, because issue salience interacts with aggregate

policy preferences to affect policy outcomes (Burstein, 2006; Lax & Phillips, 2012; Monroe, 1998),

punitive attitudes are likely to have larger effects on policymaking when the salience of crime to the

public is high (Zimring & Johnson, 2006). Third, beyond its policy implications, crime salience has

significant social, cultural, and political repercussions (Garland, 2001; Simon, 2007). Elevated crime

salience undermines trust in the government (Chanley, Rudolph, & Rahn, 2000). It also motivates

various forms of adaptive reactions—avoidance, precautionary, and defensive behaviors—by private

citizens to enhance their personal safety and protect familymembers, which in turn affect their everyday

interactions and social relationships (Warr, 2009; Warr & Ellison, 2000).

Two theoretical perspectives explain changes in the public salience of crime over time. According

to the first, the objectivist model, trends in crime salience reflect changes in actual rates of crim-

inal offending (Pickett, 2019). According to the second, the constructionist model, crime salience

varies because of changes in elite initiatives—specifically, changes in media crime coverage and polit-

ical rhetoric about crime (Beckett, 1997; Beckett & Sasson, 2004).1Both models indicate that crime

salience mainly varies at the period level, but their similarities stop there. Adjudicating between the

models is important because trends in crime salience play a central role in many theories of culture,

social structure, and crime control in late modern society (Beckett & Sasson, 2004; Garland, 2001;

Savelsberg, 1994; Simon, 2007).

Unfortunately, the few empirical studies that included aggregate data to analyze changes over time

in the public salience of crime had notable limitations. Either the researchers only reported bivariate

findings (Miller, 2013, 2016), evaluated short time periods using data from only one or two polling

organizations (Beckett, 1997; Lowry,Nio, & Leitner, 2003), or overlooked the implications of nonsta-

tionarity for time-series analysis (Hill, Oliver, & Marion, 2010; Oliver, 1998, 2002). None, as far as

we know, used multilevel data to examine simultaneously the independent effects of age, birth cohort,

and time period on perceived crime salience in the United States. This is important because “individ-

ual aging, historical contexts, and generational membership are simultaneously related to the passage

of time but have separate effects” (Gray, Grasso, Farrall, Jennings, & Hay, 2018, p. 2). As Anderson,

Lytle, and Schwadel (2017, p. 835) noted, “[S]ome social attitudes change across generations or birth

cohorts rather than across time periods.”

Additionally, according to the constructionist approach, elite initiatives are race coded and should

have larger effects among individuals who harbor anti-minority sentiments that increase their sensi-

tivity to racial “dog whistles” (Drakulich, 2015); crime salience trends should thus differ across racial

and political groups (Beckett & Sasson, 2004; Tonry,2011). Cr ime salience should also diverge along

socioeconomic cleavages. According to Garland (2001, p. 152), forexample, the “professional middle

classes” were “the dog that did not bark” during the U.S. imprisonment binge because they experienced

especially large increases in crime salience. According to the objectivist approach, on the other hand,

crime salience trends are parallel across different racial, political, and socioeconomic groups because

they all respond to the crime rate (Page & Shapiro, 1992; Pickett, 2019).

1As used herein, the term “elite initiatives” refers to agenda-setting activities by either media or political elites (Beckett &

Sasson, 2004; Chomsky & Herman, 1988).