Abstract
Although dynamic mapping is an increasingly popular tool for law enforcement, this technology is uncommon for reentry practitioners. This article introduces the Community Supervision Mapping System (CSMS), a web-based tool that routinely integrates Rhode Island Department of Corrections data into a user-friendly interface designed for those who supervise and provide service referrals to probationers. Using pre- and post-implementation survey data and actual usage data, we find that probation officers who adopt the new technology find the system useful and easy to use. We also learned that current usage is a better indicator of future use than is a respondent's reported intention to use the system.
Introduction
Computer mapping has the potential to dramatically shift the way in which the criminal justice field approaches topics such as crime control and offender supervision. Indeed, given the valuable role that geospatial applications can play, they are an increasingly important tool for practitioners, policymakers, and researchers in the field of criminal justice. 1 Mapping has been used to identify crime hot spots and trends, assess the allocation of prisoner reentry resources, evaluate potential justice policy options, and explain crime and reentry patterns to government officials and the public. Unfortunately, most justice mapping efforts use static data that can quickly become outdated. In many cases, the mapping must be done by trained analysts and only the final products are available for practitioner users.
This article describes the way in which practitioners accepted and used a new mapping tool that was piloted in the state of Rhode Island in 2008, with the goal of offering regularly updated spatial information within a user-friendly software format for practitioners. The Community Supervision Mapping System (CSMS) was designed primarily to provide probation officers and reentry service providers 2 with information about services, resources, and locations to aid in the supervision and reentry of returning prisoners. 3 Importantly, the introduction of this new web-based tool had varying degrees of user interest and adoption. Employing the technology acceptance model (TAM), a popular and influential theory to explain user willingness (or unwillingness) to accept and use available technology, this article explores the adoption patterns of CSMS by its users. In other words, when a low-cost, user-friendly mapping tool is introduced to a new type of justice practitioner (probation officers, reentry providers) in addition to law enforcement officials (who are typically more exposed to mapping tools), who adopts, why do they adopt it, and how do they use the new technology?
The following sections provide a brief background on previous reentry and justice mapping efforts, an overview of CSMS, and a description of the theories underlying the TAM. A description of the different types of CSMS adopters and the factors that influence their adoption of this new technology is supported by data analyses examining their adoption patterns. This article closes with a discussion of the implications of those analysis results, which offer insights to other jurisdictions seeking to implement new criminal justice technologies.
History of Reentry and Justice Mapping
CSMS builds on previous efforts to use geospatial technology and approaches to understand prisoner reentry at the local level. Over a decade ago, researchers began mapping patterns of incarceration and related issues such as poverty, crime, and public service use across neighborhoods and cities, with much of this work undertaken by the Justice Mapping Center 4 (Clear & Cadora, 2002). Expanding on this approach and other efforts to map justice topics, the Urban Institute (Urban) launched the Reentry Mapping Network (RMN) in 2002. The Network eventually comprised community-based organizations in 14 jurisdictions, including The Providence Plan (the developer of CSMS), which analyzed and mapped reentry data to inform local policy and practice. 5 The work of the RMN sites, the Justice Mapping Center, The Providence Plan, and other reentry mapping projects influenced policy in jurisdictions across the country, from the repeal of offender voter disenfranchisement laws to the reallocation of resources from corrections to reentry support services. 6 CSMS represents an effort to move the field forward beyond static, macro-level maps, creating a new generation of mapping applications that are useful in the day-to-day work of practitioners.
Community supervision agencies, reentry service providers, and other practitioners who work closely with returning prisoners have been somewhat slow to adopt geospatial technology and incorporate it into their work, in part due to the lack of user-friendly software applications tailored to their needs. Some multi-agency criminal justice data systems with mapping components have incorporated community supervision data, but most are complex, proprietary systems focused primarily on law enforcement. 7 Indeed, law enforcement is an entity within the larger criminal justice community that has embraced mapping to the greatest degree. Law enforcement agencies use mapping to identify crime hot spots, examine crime patterns and trends, distribute police resources effectively, and aid investigations. 8 For many law enforcement agencies, real-time, user-driven mapping is now the centerpiece of a strategic, data-guided approach to crime prevention and response. CSMS represents an effort to expand this level of technology adoption to other criminal justice practitioners.
The CSMS
CSMS is an online tool that was designed to be a user-friendly, low-cost software for easy replication in other jurisdictions. The data come from the Rhode Island Department of Corrections and the system is automatically updated each night. The data points available for each probationer include current address, personal characteristics (name, gender, race, date of birth), criminal history (including Department of Corrections identification number, offense history, and prison release date), and probation case information (supervision status, caseload number, probation officer, probation officer contact information, probation start and end dates). Photos are also available for those who were formerly incarcerated in any of Rhode Island’s Adult Correctional Institutions, which amount to approximately half of the probationers. The data can be searched or filtered based on location (distance from a specified address), time frame (such as released in the past month), and/or certain probationer characteristics. 9
The development of CSMS was an ongoing, iterative process that incorporated extensive input from end users; in addition to feedback and guidance from a local advisory group composed of probation officers, probation supervisors, police officers, reentry service providers and a national advisory group of individuals with interest and expertise in justice mapping (which represented nonprofits and criminal justice agencies from across the country), the development team continued to make revisions and expand the application’s functionality in response to user feedback. The Providence Plan also provided training sessions (typically 90 minutes in duration) for new users, provided handouts and tip sheets, and sent “E-blasts”—emails to CSMS users to provide updates and information. Finally, the development team worked to identify “super users” in each office to provide users with peer support and ongoing informal training. Although access to CSMS expanded to include police officers, reentry organizations, and even legal advocates, training and emphasis during the implementation phase were concentrated on probation officers and supervisors.
CSMS seeks to extend utilization of mapping to community supervision officers, reentry case managers, and others who support and/or supervise returning prisoners by enabling practitioners to examine reentry data spatially and dynamically. The hope is that doing so will increase efficiency, improve communication among law enforcement and probation officers, and strengthen the connection between probationers and their probation officers—whether through more accurate or accessible information (i.e., a referral to a service agency) or more contact (such as more frequent home visits). Whereas a technical report details the development, implementation, and evaluation of CSMS in Rhode Island (Lucht, La Vigne, Brazzell, & Denver, 2011), this article explores the use and adoption patterns of CSMS to develop strategies for implementing a dynamic mapping technology for community supervision officers. Specifically, we use the TAM as a framework for understanding CSMS users’ use and acceptance of the system.
The Technology Acceptance Model (TAM)
The TAM originated in the information technology field to develop an understanding of why people were underutilizing technologies designed to improve performance. Thus, early TAM research sought to identify why people intend to use technology so organizations could manipulate certain predictive factors to increase use. TAM is an extension of the Theory of Reasoned Action, or TRA (Ajzen & Fishbein, 1980; Fishbein & Ajzen, 1975), which purports that a person’s attitudes and subjective norms (or the degree to which a person feels they “should” do something) influence behavioral intentions, with the assumption that these intentions are directly related to actual behavior. The original TAM (Davis, Bagozzi, & Warshaw, 1989) proposed that external variables, such as user characteristics and organizational factors, influenced perceived usefulness and perceived ease of use, both of which then impacted the user’s attitude toward using the system. The user’s attitude was hypothesized to affect behavioral intention to use the technology, which in turn would predict actual system use. In addition, perceived usefulness was hypothesized to influence intention to use, and perceived ease of use was hypothesized to influence perceived usefulness. After testing TRA and TAM in a comparative study, Davis et al. (1989) concluded that a revised and simplified TAM would be stronger. The revised model only considers perceived usefulness, perceived ease of use, intentions to use, and actual usage (Figure 1). 10
Revised technology acceptance model, post-implementation (Davis, Bagozzi, & Warshaw, 1989, as depicted in Szajna, 1996).
Although Davis’ first major modification to TAM significantly simplified the model, subsequent versions of TAM have expanded the theory, and researchers from various disciplines have added new constructs to the model to understand technology acceptance in different contexts (for the most notable revised models, see Venkatesh, 2000; Venkatesh & Davis, 2000, for a review of TAM 2; Venkatesh, Morris, Davis, & Davis, 2003, for a test of the Unified Theory of Acceptance and Use of Technology; and Venkatesh & Bala, 2008, for a proposed TAM 3). Across variations of the model, tests of TAM tend to yield findings supportive of its underlying theory. Meta-analyses consisting of 88 studies across fields (King & He, 2006), 26 general TAM studies (Ma & Liu, 2004), and 20 studies in health information technology (Holden & Karsh, 2010) all found that the model generally predicts use and acceptance and suggest that the TAM could be more widely applied, although Holden and Karsh (2010) caution that TAM might need to be adjusted for varying situations and contexts.
In the criminal justice field, studies testing TAM are rare and limited to policing contexts. Although researchers tend to find that TAM is an appropriate model for assessing technology use and adoption for policing technologies, they also often recommend adjustments. Colvin and Goh (2005) validated TAM for patrol officers’ acceptance of a new computer technology, and also suggested modifying the model to account for contextual differences in policing. Specifically, the authors had a stronger model when they incorporated information quality and timeliness in addition to ease of use and usefulness. Gültekin (2011) found that usefulness, ease of use, and subjective norms were all significantly related to behavioral intention for a police computer network and information system for the Turkish National Police, although the author acknowledges the unique circumstances involved with this sample. In reviewing three versions of TAM for a mobile policing technology, Lindsay, Jackson, and Cooke (2011) found that a limitation of these TAMs' theoretical frameworks was the lack of organizational and social context considerations. In addition to these studies, research in the criminal justice field often considers the issues surrounding technology adoption more broadly, suggesting a need for theory-based technology adoption research in the criminal justice field. 11 Finally, although researchers are beginning to test TAM for police technologies, no published studies were found of TAM applications in the area of community supervision of probationers and parolees. This article aims to further our understanding of TAM’s utility in the criminal justice field by applying the model to a new geospatial technology for probation officers. Specifically, we test Davis et al.’s (1989) revised and simplified version of TAM to determine whether the basic model is appropriate for probation officers’ use and adoption of work-related technology.
Method/Data
The data used for this article were originally collected for an evaluation of CSMS. Urban, a nonpartisan research organization in Washington, D.C., worked closely with the Providence Plan as program evaluators during the development, implementation, and evaluation phases. Staff from Urban conducted pre- and post-implementation online surveys with staff from probation offices in Rhode Island in 2008 and 2010 to understand whether and how probation officers were utilizing CSMS for their work. The post-implementation survey items asked about probation officers’ experiences with CSMS, including frequency of use, which features they used, predicted future use, and challenges with the system. In addition to these survey data, Urban researchers obtained actual usage data from CSMS for the 6-month period following the post-implementation survey for respondents with matched (both pre- and post-implementation) survey data, and these cases were used for the final analyses (n = 33). Descriptive statistics for the pre-, post-, and matched samples are reported below in Table 1.
Descriptive Statistics for Sample.
Note. Staff begin in the Probation Officer I position and move up to Level II upon 18 months of experience and by demonstrating successful performance in the Rhode Island Department of Corrections. Some response options were excluded if zero respondents used that option; these questions are noted above by an a. The “most used feature” item and some of the items comprising the usefulness indicator were missing responses; valid percentages are presented for those questions. CSMS = Community Supervision Mapping System.
TAM Indicators
To test TAM, relevant survey items were paired with the theoretical constructs that Davis (1989) identified. First, TAM identifies two specific measures of attitude—perceived usefulness and perceived ease of use—that are theorized to impact intention to use a technology (Davis, 1989). Davis defined perceived usefulness as “the degree to which a person believes that using a particular system would enhance his or her job performance” (p. 320). To measure perceived usefulness, we scaled several items and created one indicator. The items asked whether CSMS: (a) saved time and/or helped the respondent to be more efficient; (b) provided features that were not previously available to the respondent; (c) helped with coordinating services for clients; (d) helped the respondent learn new information about his or her community/clients; (e) increased the accuracy of information the respondent uses; and (f) replaced old methods. 12 These items were summed and categorized as “very useful” (if 4-6 items were useful), “somewhat useful” (if 1-3 items were useful), and “not useful at all” (zero items were useful).
Next, Davis (1989) defined perceived ease of use as “the degree to which a person believes that using a particular system would be free of effort” (p. 320). To capture this component, we asked survey respondents “How user-friendly is the mapping tool?” Respondents had four options: “very,” “mostly,” “somewhat,” and “not at all.” A respondent’s intention to use CSMS was captured by the question “How often will you use this mapping tool in the future?” with five response options: “much more,” “slightly more,” “about the same,” “slightly less,” and “much less.”
Outcome measures of TAM are typically represented by self-reported use and/or actual system use, although studies vary in which they use. As some prior research has done, this article incorporates both to understand differences between the two. Self-reported use was categorized into “rarely/never” (defined as less than once a month), “somewhat frequently” (defined as 1-3 times a month), and “frequently” (defined as 4 or more times a month). Actual system use was extracted for each matched respondent for the 6 months after the post-implementation survey data collection ended (n = 33). Therefore, it is important to note that self-reported use and actual system use did not overlap; actual use was recorded for the 6-month period after the survey took place. These data were then averaged across the 6 months and categorized into the same categories (with the same definitions) as the self-report question.
Results
Figure 2 below displays the correlations between different TAM components. The dashed lines represent significant correlations outside of the TAM, whereas the solid lines represent relationships predicted by the model. The relationship between ease of use and usefulness is significant (at .53), which supports the model. However, the relationships surrounding intention to use are not significant. Specifically, contrary to expectations, perceived usefulness is not significantly related to intention to use and intention to use is not correlated with either self-reported or actual system use. Instead of mediating the effects of perceived usefulness and perceived ease of use, these attitudinal constructs both have direct significant relationships to self-reported use. Finally, self-reported use and actual system use are significantly related at .37. (See Appendix B for the correlation matrix.)
TAM in the CSMS context.
In addition to a direct test of Davis et al.’s (1989) TAM as it applies to the specific geospatial technology studied here, the survey of users identified other factors servings as potential predictors of technology use. An analysis of the ways in which users employed CSMS, for example, sheds light on the various ways in which it was used, with users who predicted extensive use more likely to use a greater number of analysis tools. Among respondents who reported using three or fewer analysis tools within CSMS, for example, perceptions of usefulness were scattered among not useful, somewhat useful, and very useful. However, those using four or more features reported finding CSMS to be somewhat (n = 1) or always (n = 21) useful, and these differences were statistically significant.
Lessons for technology acceptance can also be derived from challenges in technology use reported by survey respondents. Although challenges were not significantly correlated to other indicators in the TAM above, users reported a number of challenges in using CSMS in both surveys and focus groups that were conducted as part of the broader Urban evaluation (see Lucht et al., 2011), along with limitations in the software’s capabilities. Four survey respondents reported technical challenges (such as issues with the CSMS network), eight reported hardware issues (such as slow Internet access or computer-related problems), four reported inaccurate information displayed in CSMS, and seven reported confusion or difficulty using the system. Participants from several focus groups (probation officers, police detectives, and reentry service providers) suggested that although CSMS is user friendly, it takes time to become familiar with the system and build it into their daily routines. Although we might expect users who encounter challenges to report a lower predicted future use of CSMS, all of these respondents actually reported they would use the system the same or more than they do now. This may be related to resolving these challenges; everyone who reported a problem to the developer indicated that the development team was “usually” or “always” helpful in resolving issues.
Discussion
Intention to use is considered to be “the major determinant” of actual usage in the TAM framework, and this proposed link has found empirical support in previous research (Davis et al., 1989). Although most of the previous TAM literature also finds a strong relationship between perceived use and intention to use, the direct relationship between perceived ease of use and intention to use is often variable or declines in significance over time (Davis et al., 1989; Holden & Karsh, 2010; King & He, 2006).
The analysis presented here, by which TAM is tested on a new population of users—probation and parole officers—found that intention to use is not significant in any of the relationships for our sample of practitioners adopting a new geospatial technology. Instead of having a mediating effect, removing intention to use would actually strengthen the model, and current reported use appears to be a better predictor of future use than the user’s reported intention to use the technology in the future. This may be due to the nature of the correctional culture, and our respondents may be more likely to predict socially desirable levels of future use. Respondents were asked in the survey whether they felt pressure to use CSMS, and three respondents reported that they did. Alternatively, they may have genuinely believed they would be utilizing the system more than they did. In general, our respondents tended to overestimate future use, although frequent users were slightly more accurate than those who rarely or never used the system. Due to inaccurate prediction patterns, we replace intention to use with self-reported use to best represent our probation officers’ experiences with CSMS (see Figure 3).
Modified TAM.
Although these findings provide new insights into technology acceptance and use by criminal justice practitioners, they are accompanied by limitations. The most notable limitation in our study is the low sample size—we only had 33 respondents participate in both survey waves (out of 64 invitations in the baseline and 67 in the follow-up; however, not all respondents from these larger pools were employed in the department in both waves). Unlike previous information technology research, neither TAM instruments nor specifically worded TAM questions were employed in this study. 13 However, one of the benefits of using survey questions from an evaluation study is the increased possibility of applying TAM to other fields and in other contexts. For example, our “usefulness” indicator included items such as whether the probation officer learned new information about his or her community and clients due to CSMS—something a more generic instrument might not be able to capture. Another potential limitation is that data on usefulness, ease of use, and intention to use were only collected in the post-implementation survey. Although we were unable to compare the post-survey with the pre-implementation survey model due to differences in survey items, we extended the follow-up time from a shorter post-implementation time frame of 14-15 weeks (Davis et al., 1989; Szajna, 1996) found in previous literature to a 6-month follow-up period. 14 Indeed, Ajzen (1987) hypothesized that the relationship between predicted and actual usage will be stronger after the respondent is exposed to the technology for a period of time, and Davis et al. (1989) and Szajna (1996) both found supporting evidence for this hypothesis.
Future criminal justice TAM research could benefit from building in TAM-specific questions in the pre- and post-stages to determine the best model for the criminal justice context and extending the follow-up period. In addition, larger sample sizes would allow for more statistical analyses than presented here, contributing to the small but important body of research on practitioner acceptance of criminal justice technology.
Implications and Conclusion
As expected, the perceived ease of use and usefulness of CSMS are strongly related for our sample of criminal justice practitioner respondents. Both of these factors are significantly related to self-reported use of the system, and self-reported use is correlated with actual system use in the 6 months (on average) following the post-implementation survey. Our findings suggest that intention to use the system is not an important mediating effect, and considering how respondents describe self-reported system use may be a more reliable indicator of future actual use. As focus group participants in the broader CSMS evaluation study noted, finding the time to become familiar with CSMS and turning it into part of the daily work routine are critical aspects of becoming a frequent user. Therefore, it may be that self-reported use had the strongest relationship to actual use because continuing with a current routine is a more accurate predictor of future use than intentions to use a system. Moreover, the impact of this routine in sustaining use may override other factors that are more important in the earlier stages of adoption, such as having a user-friendly system. Finally, data from focus group sessions suggest that access to ongoing technical support from the development team and continued attempts to transition CSMS from an occasional information source to a routine tool were the most successful strategies for increasing frequent CSMS use.
Footnotes
Appendix
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The data collection for this project was supported by Grant No. 2007-IJ-CX-K021 awarded by the National Institute of Justice, Office of Justice Programs, U.S. Department of Justice. Points of view in this article are those of the authors and do no necessarily represent the official position or policies of the U.S. Department of Justice, The Providence Plan, or the Urban Institute.