THE PRAGMATIC AMERICAN: EMPIRICAL REALITY OR METHODOLOGICAL ARTIFACT?

• Date: 01 May 2014
• DOI: http://doi.org/10.1111/1745-9125.12035
• Published date: 01 May 2014

THE PRAGMATIC AMERICAN: EMPIRICAL REALITY

OR METHODOLOGICAL ARTIFACT?∗

JUSTIN T. PICKETT1and THOMAS BAKER2

1School of Criminal Justice, University at Albany, SUNY

2L. Douglas Wilder School of Government and Public Affairs, Virginia

Commonwealth University

KEYWORDS: public opinion, response acquiescence, crime attributions, punitiveness,

rehabilitation

Scholars widely agree that the public is pragmatic about criminal justice. The empir-

ical basis for this conclusion is the failure in several previous studies to find a sizable

negative relationship between dispositional and situational crime attributions, or be-

tween support for punitive and rehabilitative crime policies. We suggest, however, that

public pragmatism may be an artifact of the use of unidirectional question batteries in

prior research to measure attribution styles and policy support. When such questions

are used, acquiescent responding can introduce systematic error that is positively cor-

related across items and scales. Drawing on data from an experiment with a national

sample (N =826) of Internet panelists, we examine how this methodological approach

impacts the bivariate correlations and multivariate relationships between attribution

styles and between support for punitive and rehabilitative crime policies. The findings

reveal that using unidirectional sets of questions to measure these concepts likely re-

sults in 1) inflated alpha reliability coefficients, 2) an underestimation of the magnitude

of the negative relationships between attribution styles and between punitiveness and

support for rehabilitation, and 3) an underestimation of the extent to which punitive-

ness and support for rehabilitation are driven by the same factors, working in opposite

directions.

A substantial and growing body of research has investigated public opinion about

the causes of crime and the appropriate criminal justice response to offending (Ap-

plegate, Cullen, and Fisher, 1997; Applegate et al., 2000; Cochran, Boots, and Heide,

2003; Gabbidon and Boisvert, 2012; Grasmick and McGill, 1994; Pickett and Chiri-

cos, 2012; Ramirez, 2013; Thompson and Bobo, 2011; Tyler and Boeckmann, 1997;

Unnever and Cullen, 2010). This line of inquiry has been particularly salient be-

cause of the documented influence of popular attitudes on criminal justice policy and

court decisions (Baumer and Martin, 2013; Brace and Boyea, 2008; Enns, in press;

∗Additional supporting information can be found in the listing for this article in the Wiley Online

Library at http://onlinelibrary.wiley.com/doi/10.1111/crim.2011.52.issue-2/issuetoc.

Direct correspondence to Justin T. Pickett, School of Criminal Justice, University at Albany,

SUNY, 135 Western Avenue, Albany, NY 12222 (e-mail: jpickett@albany.edu). This work was

funded by the University at Albany Faculty Research Awards Program (FRAP)-Category A. The

authors are grateful to Shawn Bushway for his assistance with collecting the data, and to Christina

Mancini and Christi Metcalfe for their helpful comments on an earlier draft of the manuscript.

C2014 American Society of Criminology doi: 10.1111/1745-9125.12035

CRIMINOLOGY Volume 52 Number 2 195–222 2014 195

196 PICKETT & BAKER

Nicholson-Crotty, Peterson, and Ramirez, 2009). An important finding has been that

members of the public “are profoundly pragmatic when they consider the issue of crime

and its control” (Unnever et al., 2010: 453); they attribute crime to both individual failings

(dispositional attributions) and environmental influences (situational attributions) and,

as a result, simultaneously support punishment and rehabilitation (McCorkle, 1993;

Pickett, Chiricos, and Gertz, 2014). Thus, in their seminal review of the research, Cullen,

Fisher, and Applegate (2002: 60) concluded that “the central tendency in public opinion

is to be punitive and progressive.” Likewise, Mascini and Houtman (2006: 832) have

argued that “the habit of conceiving of support for repression and for rehabilitation as

diametrically opposed options should have been abandoned long ago.”

The existence of public pragmatism has strong policy implications. Political respon-

siveness to the interests of a pragmatic public would involve bipartisan efforts to develop

a multifaceted and balanced approach to crime control that ensures tough punishments

for offenders and provides them with rehabilitative programming. The theoretical impli-

cations of public pragmatism are also significant. Evidence of pragmatism suggests that

in the popular view, punitiveness and support for rehabilitation are distinct attitudinal

phenomena, rather than opposing sides of a single construct (Unnever et al., 2010). If

this is accurate, and punitiveness and support for rehabilitation are not opposites, then

they should not simply be related in opposite ways to the same key factors (Mascini and

Houtman, 2006; McCorkle, 1993). A central tenet of the pragmatism thesis, then, is that

the two types of policy support are best explained by separate theoretical models (Mascini

and Houtman, 2006; Pickett, Mancini, and Mears, 2013).

The concern raised in this article, however, is that the pragmatism identified in prior

studies may be an artifact of the methodology typically employed in public opinion re-

search in the field of criminology and criminal justice. Despite the overwhelming ev-

idence of acquiescence in survey responding (Krosnick, 1999; Saris et al., 2010; Schu-

man and Presser, 1981), researchers often have constructed measures of key concepts—

crime attributions, punitiveness, and support for rehabilitation—from batteries of ques-

tions that are unilaterally positive, where a response of “agree,” “support,” or “yes” in-

dicates the presence of the concept of interest.1In these instances, acquiescence—“the

tendency to answer items in a positive way regardless of their content” (Billiet and

Davidov, 2008: 543)—can result in systematic measurement error that biases observed

scores and relationships by introducing nonsubstantive variation in observations that is

positively correlated across items and scales (Baumgartner and Steenkamp, 2001). This

finding is important because the strongest evidence of pragmatism has been the absence

of a hydraulic relation—a sizeable negative correlation—between either dispositional and

situational attributions, or punitiveness and support for rehabilitation (Mascini and Hout-

man, 2006; Unnever et al., 2010). Indeed, the absence of this hypothesized negative cor-

relation is the criterion for pragmatism that Unnever et al. (2010: 433) identified at the

outset of their analysis: “[T]he failure to detect a hydraulic relation would suggest that

Americans are more centrist and pragmatic in their views on crime and its control.”

In the current study, we evaluate how acquiescence bias may have influenced both the

quality of the measures used in prior research and the resultant findings about public

1. The problems associated with response acquiescence are not limited to agree/disagree items; they

emerge across many different types of questions when unidirectional sets of statements are used

(Blamey, Bennet, and Morrison, 1999; Krosnick, 1999; Schuman and Presser, 1981).