Reception to and Efficacy of a Serious Video Game for Correctional Intervention: Project Choices

Abstract

Project Choices (PC), a newly developed videogame for correctional intervention, consists of realistic decision-making scenarios and cognitive behavioral skills feedback. A pilot study investigated PC engagement and immersion by employing a cross-over design with a sample of 24 men on probation remanded to residential treatment. The study also examined effects of gameplay on criminogenic thinking, self-perceived criminogenic risk, and social problem-solving. As hypothesized, relative to a leisure video game (Tetris), participants generally appeared comparably engaged by and immersed in PC. Most hypothesized effects of PC on treatment-relevant outcomes were not statistically significant; however, PC contributed to moderate to large treatment effects across most outcomes of interest. Although results are promising, PC requires further validation to determine whether it could serve as a useful adjunctive tool for practitioners hoping to further reach and engage corrections clients in criminogenic risk-reduction services. Future research directions for technology like PC are numerous and encouraged.

Keywords

eHealth serious video games correctional intervention rehabilitation

Rearrest for new offenses or community supervision violations is common among persons who are justice involved (Alper et al., 2018), and criminal recidivism entails numerous and substantial costs (Wickramasekera et al., 2015). Fortunately, correctional interventions have been shown to be a cost-effective strategy for reducing recidivism rates when services adhere to the evidence-based principles of the Risk-Needs-Responsivity model (Drake et al., 2009; Wormith & Zidenberg, 2018). However, limited access to and attrition from correctional rehabilitation services are concerns (Olver et al., 2011; Taxman et al., 2014). Thus, developing strategies for expanding access to and increasing engagement with RNR-adherent correctional treatment services is imperative.

A variety of factors may contribute to gaps in access to correctional treatment services, and attrition from such services—including inadequate resources and nonadherence with RNR principles, respectively. For instance, the COVID-19 pandemic presented numerous challenges for the U.S. correctional system (Carson & Nadel, 2022), and while the pandemic resulted in an overall decrease in the number of persons incarcerated, this reduction has since rebounded (Kang-Brown et al., 2021) without a commensurate increase in staff hirings (Felix et al., 2022). Understaffing may include having too few clinical providers to render services to an optimal degree and too few security staff to maximally facilitate movements within facilities to access services that are available (Blakinger, 2022; Carson & Nadel, 2022). In addition, if services are offered that do not adequately adherence to RNR principles such as the responsivity principle, then clients’ motivation for those services and the extent to which they benefit from them can be expected to be less than desired (Andrews et al., 1990; Bonta & Andrews, 2016; Gendreau et al., 1996).

One direction for correctional rehabilitation services that may help to address such challenges is greater incorporation of technology (Kip et al., 2018, 2020). While developments specific to corrections have been slow, technology offers opportunities to further reach and engage individuals in treatment services for a wide range of needs. Technological solutions for health care, often referred to by terms such as eHealth (electronic health) and mHealth (mobile health), can facilitate primary or adjunctive service delivery (Al-Shorbaji, 2013; Marcolino et al., 2018). Examples of potential benefits of technology-facilitated services for persons who are justice involved include standardization, flexibility, portability, and reduced demands on providers (Grove et al., 2021; Kip et al., 2018). Outside of the corrections space, research on eHealth has generally found such technologies to be promising, including in terms of acceptability to clients, symptom reduction, and cost-effectiveness (Bennett et al., 2020; Gentili et al., 2022). This includes a subtype of eHealth commonly termed serious video games (Eichenberg & Schott, 2017; Lau et al., 2017); that is, video games developed specifically for clinical purposes.

Accordingly, the potential for eHealth for correctional populations is being increasingly discussed by professionals (Kip & Bouman, 2021; Kirschstein, Batastini, et al., 2023). For instance, the potential for mobile technology to support community corrections practitioners and clients (Jackson et al., 2015), and positively and enjoyably reinforce appropriate decision-making practice among those who are justice involved (Russo et al., 2017). Persons who are justice involved have also recommended greater incorporation of technology into correctional interventions (King et al., 2017).

Notably, leisure video game playing is highly prevalent in the general population (WePC, n.d.; Yanev, 2021). Considering also the fact that such games can be self-administered and promote learning (e.g., Wilson et al., 2009), serious video games may be especially suited for increasing access to and engagement in, and supporting the desired effects of, correctional intervention services. However, little research has examined the feasibility and impact of such possibilities specifically for persons who are justice-involved. A systematic review of eHealth treatments for mostly adult forensic mental health populations (Kip et al., 2018) identified just six studies of technology that simulated realistic offense-related situations (as relevant to criminogenic risk), but none featured any true gameplay elements. Grove et al. (2021) similarly identified only three reports of simulation games for youth who were justice-involved. The studies typically only evaluated indicators of acceptability to participants, rather than effectiveness. To our knowledge, none of these corrections-relevant gaming technologies have been widely disseminated to practitioners. Unsurprisingly then, an international survey of forensic and correctional mental health professionals found that few (4%) had ever employed video games in practice (Kirschstein, Singh, et al., 2023).

Professionals may be increasingly interested in adopting serious video games for correctional intervention if these technologies were readily available and evidence supported their efficacy as adjunctive service tools. Accordingly, a new serious video game called Project Choices (PC) was developed by the second and third authors. PC is playable on portable devices (Apple iOS) in the interests of accessibility, familiarity, and feasibility for deployment in corrections contexts (see Kruzman, 2018; Ross, 2018). Aspects of the game were informed by a variety of sources, including review of recommendations for psychotherapeutic game design (e.g., Ahmad et al., 2016); related case examples (e.g., Thompson et al., 2010); and the limited number of serious video games that have been evaluated for mental health and forensic populations.¹ Suggestions by the game’s contracted programmers, a company specializing in developing video games in the public interest, were also incorporated. Two-dimensional graphics were used to contain development costs.

The first version of PC,² examined in this study, involves 42 realistic decision-making scenarios either anecdotally reported by men undergoing community reentry to colleagues of the game’s developers or else were informed by studies of women who were justice involved (Cobbina & Bender, 2012; Doherty et al., 2014). Examples of scenarios in the game include feeling overwhelmed with daily responsibilities, encountering other persons with a criminal history, and reflecting upon whether to resume a previously violent intimate relationship. Players can respond to scenarios with various safe, neutral, or risky choices, which influences the probabilities of resultant positive, neutral, and negative outcomes. Outcomes, in turn, influence functional player statistics: criminogenic risk, physical, mental, emotional, financial, and social. Players can also wager on their choices, which further influences the impact of a resultant outcome on their statistics. They are presented with feedback about the positive, neutral, or negative outcome probabilities associated with their choices via a “wheel of fate” feature, and they receive prompts about the desirable or undesirable consequences of changes in their criminogenic risk statistic when reaching thresholds in either direction. Players are further prompted after each scenario about particularly relevant skills from an evidence-based cognitive behavioral correctional intervention program developed by one of the game’s developers (Morgan et al., 2018), brief summaries of which they can review in-game.

The goal of PC is to make it through all scenarios, spread out over two simulated “weeks” within the game, while maximizing one’s functional statistics. Both objectives are most possible when players make safe choices that minimize the criminogenic risk statistic and favor other statistics. To increase realistic simulation and add uncertainty to player-outcome expectancy (i.e., “gamify” application of treatment skills and decision-making relevant to community reentry), high-risk decisions may (infrequently) yield favorable outcomes (i.e., positive impact on statistics). If players make especially risky decisions for specific scenarios, or a series of choices that lead their criminogenic risk statistic to grow too high, they will automatically reoffend within the game and receive an unsuccessful game over.

Current Study

The current study constituted the first pilot of PC, which focused on the game’s engagingness and immersiveness, and effects on several correctional treatment-relevant constructs, relative to a content-neutral control video game (Tetris). We used a crossover study design, in part to maximize the utility of a small sample. We hypothesized that engagement and immersion ratings would not significantly differ between the two study groups (i.e., participants who played PC or Tetris first), given PC’s multiple gameplay mechanics implemented to increase engagement and immersiveness, as informed by previous serious video games (e.g., Reynolds et al., 2017) and also with a general hope of emulating leisure video games. We also hypothesized a pattern of results for other study variables favoring the group that played PC first. Specially, that criminogenic thinking and ineffective social problem-solving orientations and styles would be significantly lower, and self-perceived criminogenic risk and desirable social problem-solving orientation and style would be significantly higher, following PC gameplay.

These hypothesized outcomes stem from PC’s affordance of practice and feedback concerning thoughtful and safe decision-making for solving hypothetical problem scenarios, which would lend to reductions in both criminogenic thinking and generally ineffective problem-solving attitudes and approaches. We also anticipated that PC’s focus on criminogenic risk (e.g., as a functional player statistic) and involvement of uncertainty of in-game simulation outcomes would generally increase player’s perceptions of their own potential for criminogenically risky decision-making in the face of ambiguous situations—including in reference to others.

Method

Participants

All procedures were approved by a university institutional review board, local probation department, and facility administrators. The study was not preregistered, and materials and analysis code for this study are not publicly posted. However, materials will be made available upon reasonable request.

Participants were sampled from a medium-sized (census of approximately 150) residential facility in the Southwestern United States serving moderate to high-risk men (aged 18–90; M^age =33.72; SD^age = 10.34) on probation who had a substance use disorder and their community supervision revoked for noncompliance. All residents were enrolled in a cognitive behavioral program focused on substance use and criminogenic thinking, with program participation averaging 9 months (M^{treatment completion} = 9.12 months; SD^{treatment completion} = 1.71 months; see also Table 1). An a priori power analysis indicated that, for a repeated-measures analysis of variance (ANOVA) with 2 groups, 3 measurements, and the following parameters—1 – β = .80, α = .05, and f² = .25—at least 28 participants were necessary for sufficient power. The target sample size was 40 to account for anticipated attrition, and comparable to most of the few preexisting studies of serious video games in justice settings (e.g., n = 30 in Arborelius et al., 2013, and n = 14 in Reynolds et al., 2017).

Table 1:

Demographics

Demographic Characteristics	Study site		Total sample^a		Project Choices firstb		Tetris first^c
Demographic Characteristics	M (SD)	%	M (SD)	%	M (SD)	%	M (SD)	%
Age	33.7 (10.32)		29.6 (6.72)		27.80 (3.90)		28.4 (5.55)
Years of education	11.21 (1.27)		11.06 (2.12)		11.72 (.65)		11.81 (1.25)
GED		-		35.0		44.4		28.6
Sentence length (months)	61.32 (31.28)		67.73 (39.47)		69.23 (41.03)		48.25 (20.64)
Current sentence served (months)	20.62 (27.27)		18.39 (13.95)		19.75 (17.28)		15.58 (9.87)
Treatment stay (weeks)	-		9.35 (6.35)		10.86 (7.74)		12.36 (9.74)
Race
White		87.0		59.0		55.6		53.3
Black		12.9		23.0		11.1		26.7
American Indian/Asian/Other		-		4.0		11.1		0.0
Multiracial		-		14.0		22.2		20.0
Ethnicity
Non-Hispanic		65		20.9		22.2		43.3
Hispanic		35		41.9		66.7		37.7
Not available		-		37.2		11.1		19.0
Relationship status
Not partnered/single		-		70.0		100.0		46.7
Divorced/separated/widower		-		7.0		0.0		0.0
Partnered/common Law		-		19.0		0.0		46.7
Married		-		4.0		0.0		6.6
Psychotropic medications		35.2		34.9
Index offense
Violent		9.2		9.3
Nonviolent		90.8		90.7

Note. GED = general education diploma; Violent = harmed by or threatened someone with violence (e.g., assault, terroristic threats). “-” indicates data unavailable.

All randomized participants (N = 43). ^bStudy completers (n = 10). ^cStudy completers (n = 14).

Residents within 2 months of anticipated discharge were excluded from participation due to lack of availability connected to community work schedules in later phases of facility treatment. All prospectively eligible individuals at the facility during the study period (June 2021 to January 2022, N = 91) were individually recruited and informed of the general sequence of study participation, as was required by the institutional review board (e.g., that participation would include playing both the experimental and control video game). Those who were capable of and provided consent to participate (N = 43, 47%) were administered baseline measures and randomized to one of the two study groups (see Figure 1). No incentives for participation were provided. Although the target sample size was exceeded for enrollment (N = 43), only 24 participants completed the study (PC-first group n = 10, Tetris-first group n = 14).

Figure 1:

Study Design and Participant Flow Diagram

The 43 enrolled participants had a mean age of 30 years (SD = 7). Most participants identified as White (59%) or Black (23%), with 42% reporting Hispanic ethnicity. Participants reported serving an average probation term of 68 months (SD = 39 months), with an average treatment stay of 9 weeks (SD = 6 weeks) at the time of baseline assessment. Regarding treatment attrition, approximately 44% (n = 19) of residents who consented to participate and were randomized to a study group subsequently withdrew. Discontinuation most often resulted from “boredom” (n = 13; 68% of dropout), though logistical issues also resulted in six additional dropouts (e.g., incurring behavioral infractions that precluded further participation, including via removal from the treatment facility). Noncompletion most frequently occurred during the first week of participation (n = 6; 14% dropout), though participant withdrawal occurred throughout the approximately 8-week study. See Table 1 and Figure 1 for additional descriptive information concerning the sample.

Technology

Two games were administered via Apple iPads (Apple, 2021) with the following specifications: 10.2-inch, A13 Bionic chip, 64 GB, iOS 15.

Project Choices (PC)

As previously described, PC (Morgan & King, n.d.) is an experimental serious video game for correctional rehabilitation. It has not previously been evaluated.

Tetris

Tetris (Tetris Holding, 2021) is a popular leisure puzzle-based video game. It was selected because of its neutral content and popularity.

Measures

Capacity Assessment Record for Research Informed Consent (CAR)

The CAR (Marson et al., 1994) is a brief structured interview that facilitates decision-making about one’s capacity to provide informed consent to participate in research.

Modified Game Engagement Questionnaire (mGEQ)

The mGEQ (Byun & Loh, 2015) is a 10-item self-report measure of engagement, potential engagement, and learning while playing a game. It was adapted from the 19-item Game Engagement Questionnaire (GEQ; Brockmyer et al., 2009), with item responses ranging from 1 (strongly disagree) to 5 (strongly agree). The mGEQ has demonstrated acceptable internal consistency (α = .76; Byun & Loh, 2015) and convergent validity with other game experience questionnaires (rs = .38 to .72; Brockmyer et al., 2009). The mGEQ was specifically chosen to repeatedly assess game engagement as it is a relatively brief measure commonly used in consumer gaming research. The total engagement scale was used in the present study. The αs ranged from .78 to .86 across administrations.

Temple Presence Inventory (TPI)

The TPI (Lombard et al., 2009) is a 38-item self-report measure of several dimensions of telepresence (i.e., immersion in technology). It has demonstrated acceptable internal consistency (αs ranging from .75 to .91; Lombard et al., 2009) and convergent validity with narrative absorption (i.e., total engagement in the experience) and game engagement (r = .55; Martey et al., 2014). The engagement/immersion scale was deemed the most relevant scale for use in the present study because of the aim to assess re-playability and because it was the most comparable to the mGEQ (vs., for example, scales of visual accuracy or realism, intimacy or social richness, and physical presence or transportation). The αs across administrations ranged from .79 to .85.

Psychological Inventory of Criminal Thinking Styles-Short Form (PICTS-SF)

The PICTS-SF (Walters, 2006, 2013) is a 35-item self-report measure of criminogenic thinking that yields an overall general criminal thinking scale, two content scales (current and historical); two factor scales (problem avoidance and self-assertion/deception); and two composite scales (proactive and reactive criminal thinking). It is an abbreviated version of the 80-item PICTS (Walters, 1995). A recent study of the PICTS-SF found it to exhibit strong internal consistency (e.g., α for the General Criminal Thinking [GCT] scale = .96) and adequate concurrent validity with other measures of criminal thinking (rs = .30 to .38; Scanlon et al., 2023). The present study used three PICTS-SF scales (general, current, and reactive criminal thinking). The αs across administrations ranged from .84 to .89 for the GCT scale, .88 to .98 for the current criminal thinking (CUR) scale, and .91 to .99 for the reactive criminal thinking (RCT) scale.

Perceived Risk Inventory (PRI)

The PRI (Kroner, 2014) measures self-perceived criminogenic risk and comprises 35 agree–disagree items. The measure yields four scales: lower than multiple comparisons (e.g., “My risk to offend is similar to those with minor legal violations”); similar to crime occurrence (e.g., “Compared to the average person who has done crime, my risk level is similar”); higher than close normative (e.g., “My risk to offend is higher than people with similar personal characteristics”); and higher than distant comparison (e.g., “My chance of doing crime is higher than the average person”). The PRI has demonstrated acceptable internal reliability in a sample of persons who were justice involved (internal consistency of scales ranged from .69 to .81; Kroner, 2014), and self-perceived risk measured by the PRI was associated with concurrent justice involvement in college samples (rs ranging from .11 to .14; Kroner, 2014). This study used two PRI scales: higher than close normative (HCN) and higher than distant comparison (HDC). The αs across administrations ranged from .66 to .91 for HCN and .71 to .80 for HDC.

Social Problem-Solving Inventory-Revised: Short Form (SPSI-R:S)

The SPSI-R:S (D’Zurilla et al., 2002) is a 25-item self-report measure that yields five scales representing the five-factor theoretical model of social problem-solving (D’Zurilla et al., 2004): positive problem orientation (PPO), negative problem orientation (NPO), rational problem-solving (RPS), impulsivity/careless style (ICS), and avoidance style (AS). Hawkins et al. (2009) observed good model fit for this five-factor model and the SPSI-R:S, as well as adequate convergent validity with other problem-solving measures (e.g., rs = .|52| to .|82|). All five scales were used in this study. The αs across administrations were as follows: PPO (.71–.94), NPO (.82–.85), RPS (.67–.90), ICS (.76–.84), and AS (.85–.88).

Procedures

The study design consisted of a 6-week cross-over trial with randomization (e.g., Senn, 2002; see Figure 1). After completing baseline measurement, participants were randomly assigned to one of the two groups: (a) a course of PC gameplay followed by a course of Tetris gameplay (PC-first group); and (b) the reverse order (Tetris-first group). Conditions were blocked into 3-week administrations of PC or Tetris. For 3 weeks, administration of each video game consisted of 30-minute gameplay sessions 3 times per week in a condition- specific group setting, with a maximum of 10 participants playing at one time. After 3 weeks, participants were readministered the baseline assessment battery. Participants then took a 12-day break before beginning their second phase before re-administration of the other video game at the same dosage. After Week 6, participants were again readministered the assessment battery.

Data Diagnostics and Analytic Strategy

Cases containing substantial missing data for any one variable (>20% of responses missing from a scale) were excluded from subsequent analyses using pairwise deletion to mitigate further loss of data (this approach was more efficient than listwise deletion; Asparouhov & Muthén, 2010). Outliers were determined by converting raw data to z-scores and using the threshold of |3.29|. For the Tetris-first group, there was one outlier case for one of the dependent variables (PICTS-SF), and because it appeared to be due to noncontent-based responding, the case was removed from analyses using this measure. Visual inspection of diagnostic plots (e.g., Q-Q plots) revealed SPSI-R: S and PRI distributions deviated from normality. Examination of skewness and kurtosis statistics indicated that these variables were moderately to highly negatively skewed (>.50). Therefore, PRI and SPSI-R: S variables were logarithmically transformed to resemble normality, which reduced skewness to moderate or less for both measures (<.50). Because F- and t-tests are generally robust to violations of normality, we chose to proceed with these parametric statistical tests.

Data were analyzed using independent-samples and paired-samples t-tests for between- or within-subjects contrasts involving two data points, respectively; and one-way ANOVA and one-way repeated-measures ANOVA for between- or within-subjects comparisons involving more than two data points, respectively. Specifically, the analytic sequence for the hypotheses concerning engagement and immersion consisted of omnibus tests (ANOVAs) to evaluate differences on these variables between groups, followed by paired-sample t-tests to evaluate within-subject differences between games. The analytic sequence for the hypotheses concerning the other study variables entailed a series of repeated-measures ANOVAs to assess pre–post differences within subjects after testing for sequencing effects (detailed below). Greenhouse-Geisser corrections were used when analyses violated the assumption of sphericity (Greenhouse & Geisser, 1959). Effect sizes (Hedge’s g) and 95% confidence intervals were calculated for all post hoc contrasts (and interpreted in line with Cohen, 1988), with the first timepoint for each measure serving as the reference (see Table 2).

Table 2:

Study Variables

Variable (Measure)		Project Choices first			Tetris first
Variable (Measure)	M	SD	g [95% CI]	M	SD	g [95% CI]
Engagement (mGEQ)
Before cross-over (Week 1)	39.50	3.63	—	39.50	14.85	—
Before cross-over (Week 2)	39.00	6.59	−0.09 [−1.08, 0.89]	46.00	12.73	0.47 [−0.42, 1.36]
Before cross-over (Week 3)	36.63	8.63	−0.43 [−1.43, 0.56]	37.50	14.85	−0.14 [−1.01, 0.74]
After cross-over (Week 4)	37.25	12.78	−0.24 [−1.23, 0.74]	39.89	5.75	0.08 [−0.84, 0.91]
After cross-over (Week 5)	37.25	11.65	−0.26 [−1.24, 0.72]	38.90	7.49	−0.18 [−0.93, 0.83]
After cross-over (Week 6)	34.75	11.83	−0.66 [−1.22, –0.08]*	35.45	7.89	−0.29 [−1.22, 0.54]
Immersion (TPI)
Before cross-over (Week 3)	5.33	1.11	—	4.42	2.31	—
After cross-over (Week 6)	5.22	1.72	0.05 [−0.55, 0.64]	4.42	1.62	0.00 [−0.53, 0.53]
Criminal Thinking (PICTS-SF)
GCT
Baseline	55.00	6.76	—	53.57	10.32	—
Before cross-over (Week 3)	52.67	9.39	−0.28 [−1.03, 0.46]	50.64	9.92	−0.29 [−1.17, 0.59]
After cross-over (Week 6)	50.89	9.10	−0.51 [−1.27, 0.24]	50.93	11.34	−0.28 [−1.12, 0.64]
CUR
Baseline	60.89	5.67	—	56.86	8.87	—
Before cross-over (Week 3)	55.67	9.77	−0.65 [−1.60, 0.30]	52.86	8.24	−0.47 [−1.36, 0.42]
After cross-over (Week 6)	54.00	8.23	−0.98 [−1.95, 0.00]	54.00	9.47	−0.31 [−1.19, 0.57]
RCT
Baseline	59.89	5.40	—	57.21	8.09	—
Before cross-over (Week 3)	56.56	8.60	−0.46 [−1.40, 0.47]	54.29	8.12	−0.36 [−1.24, 0.52]
After cross-over (Week 6)	56.56	6.96	−0.53 [−1.48, 0.41]	55.64	8.34	−0.19 [−1.07, 0.69]
Self-Perceived Criminogenic Risk (PRI)
HDC
Baseline	1.11	1.54	—	1.29	0.76	—
Before cross-over (Week 3)	1.44	1.59	0.21 [−0.53, 0.95]	1.14	0.90	−0.18 [−0.87, 0.51]
After cross-over (Week 6)	0.78	1.30	−0.23 [−0.98, 0.51]	1.42	1.27	−0.12 [−0.57, 0.82]
HCN
Baseline	1.33	1.32	—	1.75	1.75	—
Before cross-over (Week 3)	0.67	1.41	−0.48 [−1.24, 0.27]	0.63	1.06	−0.77 [−1.68, 0.14]
After cross-over (Week 6)	0.56	1.67	−0.49 [−1.32, 0.33]	1.50	1.93	−0.14 [−1.01, 0.74]
Social Problem-Solving (SPSI-R: S)
PPO
Baseline	15.33	2.45	—	15.15	3.24	—
Before cross-over (Week 3)	12.89	5.16	−0.60 [−1.36, 0.15]	13.38	2.81	−0.58 [−1.48, 0.31]
After cross-over (Week 6)	11.56	5.81	−0.85 [−1.81, 0.12]	13.15	5.16	−0.46 [−1.17, 0.24]
NPO
Baseline	6.56	3.91	—	4.08	4.82	—
Before cross-over (Week 3)	2.56	2.83	−1.17 [−1.97, –0.37]*	4.62	4.96	−0.11 [−0.77, 0.99]
After cross-over (Week 6)	5.78	4.49	−0.19 [−0.93, 0.55]	4.92	4.31	−0.17 [−0.71, 1.05]
RPS
Baseline	12.67	3.04	—	12.46	3.45	—
Before cross-over (Week 3)	11.11	4.76	−0.39 [−1.14, 0.36]	8.85	4.52	−0.90 [−1.82, –0.17]*
After cross-over (Week 6)	9.56	4.67	−0.79 [−1.75, 0.17]	9.62	5.09	−0.65 [−1.55, 0.25]
ICS
Baseline	9.67	4.69	—	7.08	4.54	—
Before cross-over (Week 3)	7.33	3.81	−0.55 [−1.49, 0.39]	6.46	4.54	−0.14 [−1.01, 0.74]
After cross-over (Week 6)	7.78	4.02	−0.43 [−1.18, 0.32]	7.85	4.47	0.17 [−0.71, 1.05]
AS
Baseline	6.56	5.39	—	5.46	3.73	—
Before cross-over (Week 3)	6.00	4.39	−0.11 [−1.04, 0.81]	5.85	5.10	0.09 [−0.61, 0.78]
After cross-over (Week 6)	6.44	5.36	−0.02 [−0.76, 0.72]	7.08	4.25	0.41 [−0.30, 1.11]

Note. Effect sizes reflect the magnitude of the difference relative to the reference (—) timepoint. mGEQ = Modified Game Engagement Questionnaire; TPI = Temple Presence Inventory; PICTS-SF = Psychological Inventory of Criminal Thinking-Short Form; GCT = General Criminal Thinking; CUR = Current Criminal Thinking; RCT = Reactive Criminal Thinking; PRI = Perceived Risk Inventory; HDC = Higher Perceived Criminal Risk than Distant Comparison;

HCN = Higher Perceived Criminal Risk than Close Normative; SPSI-R:S = Social Problem-Solving Inventory-Revised: Short Form; PPO = Positive Problem Orientation; NPO = Negative Problem Orientation; RPS = Rational Problem-Solving; ICS = Impulsive/Careless Style; AS = Avoidance Style.

p < .05.

Given the potential for carryover (sequencing) effects (i.e., outcomes being influenced by preceding condition exposure), separate mixed (between–within) ANOVAs were used to investigate the effects of carryover effects on desired outcomes and game engagement. We specified group (PC-first × Tetris-first) as a between factor; “time” (baseline × before crossover × after crossover) as a within factor; and engagement (mGEQ), immersion (TPI), criminal thinking (PICTS-SF), self-perceived criminogenic risk (PRI), or problem-solving [SPSI:R-S] as the dependent variable. Significant “group” × “time” interaction effects would indicate significant carryover effects.

Results

Randomization and Sample Characteristics

Descriptives for participant characteristics are presented in Table 1, and statistical information to complement that which is narrated below is provided in Table 2. There were no statistically significant differences across these variables between the two groups (PC or Tetris first). There were, however, statistically significant differences between study completers and noncompleters, with the latter having longer probation sentences, p = .02, g = .73; higher levels of criminal thinking, ps = .04 to .01, gs = 0.63 to 0.77; and higher levels of negative problem orientation, p = .05, g = 0.62.

Engagement and Immersion

Prior to game cross-over, there were no omnibus differences for engagement (mGEQ), F(1, 26) = 1.48, p = .26, η_p² = .05; nor immersion (TPI), F(1, 26) = .99, p = .33, η_p² = .04. There were also no differences for immersion before and after the game cross-over for either group, PC first: t(8) = .16, p = .88, g = .05, 95% CI [−0.55, 0.64]; Tetris first: t(11) = .00, p = 1.00, g = .00, 95% CI [−0.53, 0.53]. There was, however, a significant omnibus within-subjects effect for engagement for the PC-first group, F(2, 24) = 5.33, p = .01, η_p² = .21. Post hoc contrasts indicated that engagement was significantly and moderately lower at Week 6 (i.e., during Tetris gameplay) relative to Week 1 (i.e., during PC gameplay), t(9) = 2.55, p = .03, g = −.66, 95% CI [−1.22, –.08]. Comparatively, there was no evident omnibus change in engagement for the Tetris -first group, F(2, 24) = .45, p = .65, η_p² = .08. No carryover effect was observed for engagement, F(1.61, 12.90) = .72, p = .48, η_p² = .08; nor immersion, F(1, 21) = .06, p = .81, η_p² = .00.

Criminogenic Thinking

Using the PICTS-SF, there were no omnibus within-subject differences in general criminogenic thinking (GCT) for either group, PC first: F(2, 16) = 1.48, p = .26, η_p² = .16; Tetris first: F(2, 26) = 1.01, p = .38, η_p² = .07. There were also no omnibus changes in reactive criminogenic thinking (RCT), PC first: F(2, 16) = 1.32, p = .30, η_p² = .14; Tetris first: F(2, 26) = 1.39, p = .27, η_p² = .10. Results for omnibus changes in current criminogenic thinking (CUR) approached statistical significance for both groups, PC first: F(1.21, 9.70) = 3.75, p = .08, η_p² = .32; Tetris first: F(2, 26) = 2.73, p = .08, η_p² = .17. However, because CUR scores were lower before and after cross-over for both groups relative to baseline, this suggested potential regression to the mean. Effect sizes were indicative of small to large reductions in current criminal thinking (g’s ranging from −.31 to −.98) and reactive criminal thinking (g’s ranging from −.19 to −.53) for treatment completers. There was no carryover effect evident for any of the PICTS-SF scores, GCT: F(2, 42) = 0.19, p = .78, η_p² = .01; CUR: F(2, 42) = 0.93, p = .40, η_p² = .04; RCT: F(2, 42) = 0.20, p = .82, η_p² = .01.

Self-Perceived Criminogenic Risk

Using the PRI, there were no omnibus within-subjects changes in perception of risk relative to an average person (HDC) for either group, PC first: F(2, 16) = 2.40, p = .12, η_p² = .23; Tetris first: F(2, 12) = .39, p = .68, η_p² = .06. In contrast, a significant omnibus change was observed for the PC-first group for perception of risk relative to persons with similar characteristics to oneself (HCN), F(2, 16) = 3.74, p = .05, η_p² = .39; vs. a nonsignificant omnibus change effect for the Tetris-first group, F(1.15, 8.04) = 1.65, p = .24, η_p² = .19.

However, none of the post hoc contrasts for HCN for the PC-first group were significant, and nonsignificant reductions in HCN relative to baseline were also observed for the Tetris-first group. Effect sizes were indicative of small to large reductions in self-perceived criminogenic risk relative to justice-involved peers (g’s ranging from −.14 to −.77), and small reductions or increases in self-perceived criminogenic risk relative to nonjustice-involved persons (g’s ranging from −.23 to .21) for treatment completers. No carryover effect was evident for either of the two perceived criminogenic risk scales, HCN: F(2, 30) = 1.01, p = .38, η_p² = .06; HDC: F(2, 28) = 2.26, p = .13, η_p² = .14.

Social Problem-Solving

Neither group evidenced omnibus within-subjects differences in positive problem orientation (PPO). The effect for the PC-first group did, however, approach significance, F(2, 16) = 3.43, p = .06, η_p² = .30; whereas the effect for the Tetris-first group did not, F(2, 16) = 1.45, p = .25, η_p² = .11. Nonetheless, post hoc contrasts suggested no notable PPO differences relative to baseline. While no difference in negative problem orientation (NPO) was observed for the Tetris-first group, F(2, 24) = .31, p = .74, η_p² = .03, there was a significant difference in NPO for the PC-first group, F(2, 16) = 5.90, p = .01, η_p² = .43. Post hoc contrasts indicated that NPO was significantly lower at Week 3 relative to baseline, t(13) = 2. 70, p = .02, g = −1.17, 95% CI [−1.97, –.37].

There were no omnibus within-subjects changes for either of the groups for two of the three problem-solving styles. The effects for impulsive/careless (ICS)—PC first: F(1.26, 10.09) = 0.88, p = .43, η_p² = .10; Tetris first: F(1.38, 16.56) = 1.07, p = .34, η_p² = .08—and avoidant (AS)—PC first: F(2, 16) = 0.04, p = .96, η_p² = .005; Tetris first: F(2, 24) = 1.65, p = .21, η_p² = .12—were not statistically significant. However, there was a significant omnibus within-subjects effect for rational problem-solving (RPS) for the Tetris-first group, F(2, 24) = 3.31, p = .051, η_p² = .19; vs. a nonsignificant effect for the PC-first group, F(2, 16) = 1.70, p = .21, η_p² = .18. Post hoc contrasts indicated that RPS was significantly lower for the Tetris-first group at Week 3 compared to baseline, t(14) = 2.87, p = .01, g = −0.90, 95% CI [−1.82, –0.17]. Effect sizes were indicative of small to large reductions or increases in problem-solving scales associated with maladaptive functioning (NPO range = −.11 to −1.17; ICS range = .17 to −.55), and small to large reductions on scales associated with adaptive functioning (PPO range = −.45 to −.85; RPS range = −.39 to −.90). There was no significant carryover effect for PPO, F(2, 40) = 0.42, p = .66, η_p² = .02; RPS, F(2, 40) = 0.63, p = .52, η_p² = .03; ICS, F(1.52, 30.34) = 0.99, p = .36, η_p² = .05; nor AS, F(2, 40) = 0.35, p = .71, η_p² = .02. However, a significant carryover effect was observed for NPO, F(2, 40) = 3.01, p = .03, η_p² = .16, suggesting game administration sequence significantly impacted changes in NPO scores. Yet when NPO was examined between subjects at just Week 3, an independent sample t-test did not yield a significant difference in NPO between the PC-first and Tetris-first groups, t(22) = 1.30, p = .21, g = 0.53, 95% CI [−1.34, 0.29].

Discussion

As hypothesized, engagement with and immersion in PC appeared generally comparable to a leisure video game, excepting one result potentially favoring the PC-first group. In contrast, we generally did not observe clear and consistent statistically significant treatment-relevant effects in-line with our hypotheses for PC on outcomes of interest. However, examination of effect sizes indicated PC may contribute to reductions in current and reactive criminal thinking, as well as other improved outcomes for problem-solving.

Engagement and Immersion

The possibility that video game technology may serve as a motivating, experiential treatment adjunct is suggested in part by the popularity of leisure video games. Participants reported generally comparable levels of engagement and immersion for both games, which is promising from a responsivity perspective, considering the clinical nature of PC relative to the leisurely Tetris game. Somewhat notable is that the PC-first group reported higher levels of engagement following the first week of PC gameplay relative to the final week of Tetris gameplay after cross-over. However, the lack of internal replication of this potential PC-favoring effect suggests need of further replication.

Treatment-Relevant Outcomes

There are a variety of possibilities for the general lack of clear and consistent between- group effects (some degree of internal replication was possible due to the cross-over study design we employed) of PC on most treatment-relevant outcomes. Beyond the explanation that analyses were simply underpowered to detect all but substantially sized effects, it is possible that more than three weeks of PC gameplay, or supplementation with facilitation and feedback from a provider (beyond the “bibliotherapy” skills feedback incorporated into the game), is needed for such technology to yield desirable effects on instrumental intervention outcomes (e.g., criminogenic thinking reductions). Regarding the unexpected effect of those in the PC-first group rating themselves as lower in self-perceived risk than persons similar to themselves after baseline, we had anticipated that game elements embedded in PC (e.g., multiple choices and probabilistic outcomes) would result in participants perception of their own risk increasing.

However, it is possible that these participants actually unexpectedly grew more confident in their abilities to make less risky decisions, and in turn, to have decreased their self-reported risk for recidivism relative to their peers. Future work should reevaluate potential decrease in perceived risk more directly to further elucidate the relation between PC exposure and changes in prosocial decision-making confidence, given such effects were only observed in one group. While it is noteworthy that we observed moderate to large effects for reductions in current and reactive criminal thinking—outcomes very relevant to criminogenic risk-reduction interventions—given the unreliability of these effects, further research is needed to determine whether the effects were spurious.

Limitations and Lessons Learned

As the current study was a pilot, there are several limitations of note. Even with the aid of the cross-over study design, we observed an approximately 50% dropout rate. The small final sample size resulted in underpowered analyses and imprecise confidence intervals around effect sizes. Relatedly, the cross-over study design may have contributed to participants growing fatigued over time with gaming in general, impeding potential internal replication of effects observed for one group vs. the other. Also, both the researchers and participants were aware of group assignments (as was required by the institutional review board that approved this study), which leaves open the possibility of socially desirable responding. As for external validity, while substance use problems are very common among persons who are justice-involved (Bronson et al., 2017; Maruschak et al., 2021), the study site was nevertheless highly specific (i.e., serving moderate- to high-risk men on probation with substance use issues who were remanded for treatment), which may limit generalizability. Moreover, considering the study site’s relatively intensive treatment milieu, PC may have only added an overall insignificant amount of programming to participants’ high treatment dosage at the facility. In settings in which lesser treatment programming is offered, or none at all, effects for PC may be different. Other external validity concerns included participation having occurred via self-selection, and several baseline differences having been observed between those who did and did not complete the study.

Further, participants at the study site had access to other leisure videogames beyond the study for playing during recreation time; thus, participants may have been more interested in or familiar with video games than other justice-involved populations.

We highlight next some of our experiences from this pilot that we think may be applicable to other efforts to implement and evaluate serious video game technologies in correctional settings. Regarding buy-in and logistics, early demonstrations to administrators and security staff were necessary to assure them that use of the mobile devices by residents would be controlled and appropriate (e.g., demonstrating the “Guided Access” function that ensured operation of the iPad was limited to the PC application). In addition, working with participants in groups of 10 presented challenges with respect to coordinating the movement of residents.

Administrators frequently had to correspond with their staff to prevent or minimize gameplay interruptions in light of certain pressing facility operations (e.g., attempting to not interrupt gameplay engagement while ensuring accurate headcounts). Of course, such challenges likewise occur in nontechnological treatment programming implementation efforts (see, Scanlon & Morgan, 2024)—with implementation challenges being par for the course in corrections (see Rhine et al., 2006; Salisbury et al., 2019). Technological interventions are unlikely to circumvent these common types of practical challenges. Thus, future serious video game implementations may benefit from broad (e.g., facility-wide) technology demonstration and coordination among stakeholders to ensure all who will be involved have an appropriate understanding of the technology to be deployed, and that planning is in place for how the deployment will proceed considering facility operations schedules and exigent scenarios.

As for the particular stakeholder group of players, some participants in this study mentioned that they found PC’s stats mechanism (i.e., gamified individual performance tracking) to be less appealing than would be competing against other players in their group. These participants voiced a desire for everyone’s performance to be shared among peers. Although this study did not accommodate this preference, future studies of serious video games might consider exploring whether facilitating group-based competition facilitates increased engagement with this type of technology. In addition, players vacillated in their responsiveness to the session dosage of 30 minutes that was preset in the current study. Sometimes, participants requested to leave sessions before 30 minutes had passed, and other times, they requested to continue gaming beyond the half-hour duration of sessions. Future research might thus investigate different gameplay dosages, including both prespecified time intervals and accommodation of individual player preferences (with the latter emulating the pick-up-and-play nature of casual video gaming).

Future Directions

Notwithstanding the assorted limitations of this pilot study, considering a few potentially promising findings, the generally underdeveloped state of the literature in this area, and the seeming need for new partial solutions to long-standing challenges in correctional rehabilitation, further research on PC and similar technologies appears warranted. To our knowledge, no video games designed specifically for persons who are justice involved have yet been validated and made readily available for use by practitioners; thus, further evaluation of PC will be important. This includes different gaming administration dosages, personal characteristics (e.g., demographic effects, familial support) that may influence responsivity to such technology (e.g., willingness to try and other acceptability indicators), and criminogenic risk and needs outcomes. Naturally, strong study designs with larger samples are recommended. Moreover, a decision-making game featuring realistic scenarios faced by persons who are justice involved may have assessment utility (e.g., for criminogenic thinking). This possibility seems as worthy of investigation as intervention-focused evaluations.

Footnotes

We have no known conflict of interest to disclose.

This study was not preregistered. The data that support the findings of this study are available from the corresponding author upon reasonable request. Materials and analysis code for this study are not available. A portion of these results were presented at the 2022 American Psychology-Law Annual Conference. A portion of this research was supported by a student grant awarded to the primary author from the International Association Forensic Mental Health Services.

ORCID iDs

Keegan J. Diehl

Robert D. Morgan

Christopher M. King

Notes

Keegan J. Diehl is a Counseling Psychology doctoral student at Texas Tech University’s Department of Psychological Sciences. His research interests include mental health treatment accessibility in total institutions and psychological assessment.

Robert D. Morgan is the Dean, College of Health and Human Sciences at Southern Illinois University-Carbondale. His research interests are in correctional mental health treatment, forensic mental health assessment, and professional development and training issues.

Christopher M. King is currently an assistant professor in the Psychology Department at Montclair State University. He directs a research laboratory that focuses on forensic psychology, correctional psychology, mental health law, and police and public safety psychology.

Paul B. Ingram is currently an associate professor in the Department of Psychological Sciences at Texas Tech University. He directs a research laboratory that focuses on personality assessment, treatment seeking attitudes and behaviors, and assessment training practices.

Cooper Mitchell is attending Southern Illinois University. Her research interests include rehabilitative services for justice-involved persons with mental illness, justice-involved women issues, and correctional mental health interventions.

References

Ahmad

Ab Rahim

Arshad

N. I.

Mehat

(2016). A multi-domain framework for modelling educational games: Towards the development of effective educational games. In Fook

C. Y.

Sidhu

G. K.

Narasuman

Fong

L. L.

Rahman

S. B. A.

(Eds.), 7th International Conference on University Learning and Teaching (Incult 2014) Proceedings (pp. 423–438). Springer.

Alper

Durose

M. R.

Markman

(2018). 2018 update on prisoner recidivism: A 9-year follow-up period (2005–2014) [NCJ 250975]. U.S. Department of Justice, Office of Justice Programs, Bureau of Justice Statistics. https://bjs.ojp.gov/content/pub/pdf/18upr9yfup0514.pdf

Al-Shorbaji

(2013). The World Health Assembly resolutions on eHealth: eHealth in support of universal health coverage. Methods of Information in Medicine, 52(6), 463–466. https://doi.org/10.1055/s-0038-1627062

Andrews

D. A.

Bonta

Hoge

R. D.

(1990). Classification for effective rehabilitation: Rediscovering psychology. Criminal Justice and Behavior, 17(1), 19–52. https://doi.org/10.1177/0093854890017001004

Apple. (2021). iPad [technology]. Apple Technology.

Arborelius

Fors

Svensson

A. K.

Sygel

Kristiansson

(2013). A new interactive computer simulation system for violence risk assessment of mentally disordered violent offenders. Criminal Behaviour and Mental Health, 23(1), 30–40. https://doi.org/10.1002/cbm.1849

Asparouhov

Muthén

(2010). Weighted least squares estimation with missing data. Mplus. https://www.statmodel.com/download/GstrucMissingRevision.pdf

Bennett

C. B.

Ruggero

C. J.

Sever

A. C.

Yanouri

(2020). eHealth to redress psychotherapy access barriers both new and old: A review of reviews and meta-analyses. Journal of Psychotherapy Integration, 30(2), 188–207. https://doi.org/10.1037/int0000217

Blakinger

(2022, November 4). Why so many jails are in a “state of complete meltdown.” The Marshall Project. https://www.themarshallproject.org/2022/11/04/why-so-many-jails-are-in-a-state-of-complete-meltdown

10.

Bonta

Andrews

D. A.

(2016). The psychology of criminal conduct (6th ed.). Routledge. https://doi.org/10.4324/9781315677187

11.

Brockmyer

J. H.

Fox

C. M.

Curtiss

K. A.

McBroom

Burkhart

K. M.

Pidruzny

J. N.

(2009). The development of the Game Engagement Questionnaire: A measure of engagement in video game-playing. Journal of Experimental Social Psychology, 45(4), 624–634. https://doi.org/10.1016/j.jesp.2009.02.016

12.

Bronson

Stroop

Zimmer

Berzofsky

(2017). Drug use, dependence, and abuse among state prisoners and jail inmates, 2007–2009 [NCJ 250546]. U.S. Department of Justice, Office of Justice Programs, Bureau of Justice Statistics. https://bjs.ojp.gov/content/pub/pdf/dudaspji0709.pdf

13.

Byun

Loh

C. S.

(2015). Modified Game Engagement Questionnaire [Data base record]. PsycTESTS. https://doi.org/10.1037/t39659-000

14.

Carson

E. A.

Nadel

(2022). Impact of COVID—19 on state and federal prisons, March 2020—February 2021 [NCJ 304500]. U.S. Department of Justice, Office of Justice Programs, Bureau of Justice Statistics. https://bjs.ojp.gov/content/pub/pdf/icsfp2021.pdf

15.

Cobbina

J. E.

Bender

K. A.

(2012). Predicting the future: Incarcerated women’s views of reentry success. Journal of Offender Rehabilitation, 51(5), 275–294. https://doi.org/10.1080/10509674.2012.683323

16.

Cohen

(1988). Statistical Power Analysis for the Behavioral Sciences (2nd ed.). Routledge. https://doi.org/10.4324/9780203771587

17.

Doherty

Forrester

Brazil

Matheson

F. I.

(2014). Finding their way: Conditions for successful reintegration among women offenders. Journal of Offender Rehabilitation, 53(7), 562–586. https://doi.org/10.1080/10509674.2014.944740

18.

Drake

E. K.

Aos

Miller

M. G.

(2009). Evidence-based public policy options to reduce crime and criminal justice costs: Implications in Washington state. Victims & Offenders, 4(2), 170–196. https://doi.org/10.1080/15564880802612615

19.

D’Zurilla

T. J.

Chang

E. C.

Sanna

L. J.

(2004). Social problem-solving: Current status and future directions. In Chang

E. C.

D’Zurilla

T. J.

Sanna

L. J.

(Eds.), Social problem-solving: Theory, research, and training (pp. 241–253). American Psychological Association. https://doi.org/10.1037/10805-014

20.

D’Zurilla

T. J.

Nezu

A. M.

Maydeu-Olivares

(2002). Social Problem-Solving Inventory- Revised (SPSI-R): Technical manual. Multi-Health Systems, Inc.

21.

Eichenberg

Schott

(2017). Serious games for psychotherapy: A systematic review. Games for Health Journal, 6(3), 127–135. https://doi.org/10.1089/g4h.2016.0068

22.

Felix

Pyrooz

Novisky

Tostlebe

Dockstader

(2022). Effects of COVID-19 on prison operations. National Institute of Corrections. https://nicic.gov/resources/nic-library/all-library-items/effects-covid-19-prison-operations

23.

Gendreau

Little

Goggin

(1996). A meta-analysis of the predictors of adult offender recidivism: What works! Criminology, 34(4), 575–608. https://doi.org/10.1111/j.1745-9125.1996.tb01220.x

24.

Gentili

Failla

Melnyk

Puleo

Tanna

G. L. D.

Ricciardi

Cascini

(2022). The cost-effectiveness of digital health interventions: A systematic review of the literature. Frontiers in Public Health, 10, Article 787135. https://doi.org/10.3389/fpubh.2022.787135

25.

Greenhouse

S. W.

Geisser

(1959). On methods in the analysis of profile data. Psychometrika, 24(2), 95–112. https://doi.org/10.1007/BF02289823

26.

Grove

King

C. M.

Bomysoad

Vasquez

Kois

L. E.

(2021). Technology for assessment and treatment of justice-involved youth: A systematic literature review. Law and Human Behavior, 45(5), 413–426. https://doi.org/10.1037/lhb0000467

27.

Hawkins

Sofronoff

Sheffield

(2009). Psychometric properties of the Social Problem-solving Inventory-Revised Short-Form: Is the short form a valid and reliable measure for young adults? Cognitive Therapy and Research, 33(5), 462–470. https://doi.org/10.1007/s10608-008-9209-7

28.

Jackson

B. A.

Russo

Hollywood

J. S.

Woods

Silberglitt

Drake

G. B.

Shaffer

J. S.

Zaydman

Chow

B. G.

(2015). Fostering innovation in community and institutional corrections: Identifying high-priority technology and other needs for the U.S. corrections sector. https://www.ncjrs.gov/pdffiles1/nij/grants/248580.pdf

29.

Kang-Brown

Montagnet

Heiss

(2021). People in jail and prison in spring 2021. Vera Institute of Justice. https://www.vera.org/publications/people-in-jail-and-prison-in-spring-2021

30.

King

C. M.

Heilbrun

Kim

N. Y.

McWilliams

Phillips

Barbera

Fretz

(2017). Tablet computers and forensic and correctional psychological assessment: A randomized controlled study. Law and Human Behavior, 41(5), 468–477. https://doi.org/10.1037/lhb0000245

31.

Kip

Bouman

(2021). A perspective on the integration of eHealth in treatment of offenders: Combining technology and the Risk-Need-Responsivity model. Frontiers in Psychiatry, 12, Article 703043. https://doi.org/10.3389/fpsyt.2021.703043

32.

Kip

Bouman

Kelders

S. M.

van Gemert-Pijnen

(2018). eHealth in treatment of offenders in forensic mental health: A review of the current state. Frontiers in Psychiatry, 9, Article 42. https://doi.org/10.3389/fpsyt.2018.00042

33.

Kip

Sieverink

van Gemert-Pijnen

L. J. E. W. C.

Bouman

Y. H. A.

Kelders

S. M.

(2020). Integrating people, context, and technology in the implementation of a web- based intervention in forensic mental health care: Mixed-methods study. Journal of Medical Internet Research, 22(5), Article e16906. https://doi.org/10.2196/16906

34.

Kirschstein

M. A.

Batastini

A. B.

Singh

J. P.

Graf

(2023). Technological innovations in forensic and correctional mental health: An introduction to the special issue. Criminal Justice and Behavior, 50(2), 167–174. https://doi.org/10.1177/00938548221141610

35.

Kirschstein

M. A.

Singh

J. P.

Rossegger

Endrass

Graf

(2023). International survey on the use of emerging technologies among forensic and correctional mental health professionals. Criminal Justice and Behavior, 50(2), 175–196. https://doi.org/10.1177/009385482110420

36.

Kroner

D. G.

(2014). Perceived Risk Inventory [Unpublished measure]. Department of Criminology and Criminal Justice, Southern Illinois University Carbondale.

37.

Kruzman

(2018, July 18). In U.S. prisons, tablets open window to the outside world. Reuters. https://www.reuters.com/article/us-usa-prisons-computers/in-u-s-prisons-tablets-open-window-to-the-outside-world-idUSKBN1K813D

38.

Lau

H. M.

Smit

J. H.

Fleming

T. M.

Riper

(2017). Serious games for mental health: Are they accessible, feasible, and effective? A systematic review and meta-analysis. Frontiers in Psychiatry, 7, Article 209. https://doi.org/10.3389/fpsyt.2016.00209

39.

Lombard

Ditton

T. B.

Weinstein

(2009, November). Measuring presence: The Temple Presence Inventory [Paper presentation]. Twelfth International Workshop on Presence, Los Angeles, California, USA. http://matthewlombard.com/ISPR/Proceedings/2009/Lombard_et_al.pdf

40.

Marcolino

M. S.

Oliveira

D’Agostino

Ribeiro

A. L.

Alkmim

Novillo-Ortiz

(2018). The impact of mHealth interventions: Systematic review of systematic reviews. JMIR mHealth and uHealth, 6(1), Article e23. https://doi.org/10.2196/mhealth.8873

41.

Marson

D. C.

Schmitt

F. A.

Ingram

K. K.

Harrell

L. E.

(1994). Determining the competency of Alzheimer patients to consent to treatment and research. Alzheimer Disease and Associated Disorders, 8(Suppl. 4), 5–18. https://doi.org/10.1097/00002093-199400000-00002

42.

Martey

R. M.

Kenski

Folkestad

Feldman

Gordis

Shaw

Stromer-Galley

Clegg

Zhang

Kaufman

Rabkin

A. N.

Shaikh

Strzalkowski

(2014). Measuring game engagement: Multiple methods and construct complexity. Simulation & Gaming, 45(4–5), 528–547. https://doi.org/10.1177/1046878114553575

43.

Maruschak

L. M.

Bronson

Alper

(2021). Alcohol and drug use and treatment reported by prisoners [NCJ 252641]. U.S. Department of Justice, Office of Justice Programs, Bureau of Justice Statistics. https://bjs.ojp.gov/sites/g/files/xyckuh236/files/media/document/adutrpspi16st.pdf

44.

Morgan

R. D.

King

C. M.

(n.d.). Project Choices [Mobile iOS]. SkylessGames.

45.

Morgan

R. D.

Kroner

D. G.

Mills

J. F.

(2018). A treatment manual for justice involved persons with mental illness: Changing lives and changing outcomes. Routledge; Taylor & Francis Group. https://doi.org/10.4324/9781315204994

46.

Olver

M. E.

Stockdale

K. C.

Wormith

S. J.

(2011). A meta-analysis of predictors of offender treatment attrition and its relationship to recidivism. Journal of Consulting and Clinical Psychology, 79(1), 6–21. https://doi.org/10.1037/a0022200

47.

Reynolds

L. M.

Davies

J. P.

Mann

Tulloch

Nidsjo

Hodge

Maiden

Simpson

(2017). StreetWise: Developing a serious game to support forensic mental health service users’ preparation for discharge: A feasibility study. Journal of Psychiatric and Mental Health Nursing, 24(4), 185–193. https://doi.org/10.1111/jpm.12340

48.

Rhine

E. E.

Mawhorr

T. L.

Parks

E. C.

(2006). Implementation: The bane of effective correctional programs. Criminology & Public Policy, 5(2), 347–358. https://doi.org/10.1111/j.1745-9133.2006.00382.x

49.

Ross

(2018). Policy, practice and regulatory issues in mobile technology treatment for forensic clients. European Journal of Probation, 10(1), 44–58. https://doi.org/10.1177/2066220318761382

50.

Russo

Drake

D. B.

Shaffer

J. S.

Jackson

B. A.

(2017). Envisioning an alternative future for the corrections sector within the U.S. criminal justice system. Office of Justice Programs. https://www.ojp.gov/pdffiles1/nij/grants/251357.pdf

51.

Salisbury

E. J.

Sundt

Boppre

(2019). Mapping the implementation landscape: Assessing the systemic capacity of statewide community corrections agencies to deliver evidence-based practices. Corrections, 4(1), 19–38. https://doi.org/10.1080/23774657.2018.1522279

52.

Scanlon

Lester

M. E.

Brace

Batastini

A. B.

Morgan

R. D.

(2023). Tried and true? A psychometric evaluation of the Psychological Inventory of Criminal Thinking Styles- Short Form. Assessment, 30(6), 1985–1997. https://doi.org/10.1177/10731911221132500

53.

Scanlon

Morgan

R. D.

(2024). Mental health services in jail: Identifying and quantifying barriers to implementation. Manuscript submitted for publication.

54.

Senn

(2002). Cross-over trials in clinical research (2nd ed.). Wiley.

55.

Taxman

F. S.

Pattavina

Caudy

(2014). Justice reinvestment in the United States: An empirical assessment of the potential impact of increased correctional programming on recidivism. Victims & Offenders, 9(1), 50–75. https://doi.org/10.1080/15564886.2013.860934

56.

Tetris Holding. (2021). Tetris (Version 5.0.2) [Mobile application software]. https://apps.apple.com/us/app/tetris/id1491074310

57.

Thompson

Baranowski

Buday

Baranowski

Thompson

Jago

Griffith

M. J.

(2010). Serious video games for health: How behavioral science guided the development of a serious video game. Simulation & Gaming, 41(4), 587–606. https://doi.org/10.1177/1046878108328087

58.

Walters

G. D.

(1995). The Psychological Inventory of Criminal Thinking Styles: Part I: Reliability and preliminary validity. Criminal Justice and Behavior, 22(3), 307–325. https://doi.org/10.1177/0093854895022003008

59.

Walters

G. D.

(2006). The Psychological Inventory of Criminal Thinking Styles (PICTS) professional manual. Center for Lifestyle Studies.

60.

Walters

G. D.

(2013). The Psychological Inventory of Criminal Thinking Styles (PICTS) professional manual. Center for Lifestyle Studies.

61.

WePC. (n.d.). Video game industry statistics, trends and data in 2021. https://www.wepc.com/news/video-game-statistics/

62.

Wickramasekera

Wright

Elsey

Murray

Tubeuf

(2015). Cost of crime: A systematic review. Journal of Criminal Justice, 43(3), 218–228. https://doi.org/10.1016/j.jcrimjus.2015.04.009

63.

Wilson

K. A.

Bedwell

W. L.

Lazzara

E. H.

Salas

Burke

C. S.

Estock

J. L.

Orvis

K. L.

Conkey

(2009). Relationships between game attributes and learning outcomes: Review and research proposals. Simulation & Gaming, 40(2), 217–266. https://doi.org/10.1177/1046878108321866

64.

Wormith

J. S.

Zidenberg

A. M.

(2018). The historical roots, current status, and future applications of the Risk-Need-Responsivity Model (RNR). In Jeglic

E. L.

Calkins

(Eds.), New frontiers in offender treatment (pp. 11–41). Springer. https://doi.org/10.1007/978-3-030-01030-0_2

65.

Yanev

(2021, October 3). Video game demographics—Who plays games in 2021. TechJury. https://techjury.net/blog/video-game-demographics/#gref