SAVRY Predictive Validity in Asian Youth

Shanna M. Y. Li, Simon Fraser University
Jodi L. Viljoen, Simon Fraser University
Aisha K. Christiansen, Simon Fraser University
Nicole M. Muir, Simon Fraser University; University of Toronto

Objective: Although past studies suggest that the Structured Assessment of Violence Risk in Youth (SAVRY) has moderate predictive validity, its predictive validity with Asian youth in Western countries is unknown. We therefore compared the SAVRY’s predictive validity in a sample of Asian Canadian versus White Canadian youth. Hypotheses: Given that the SAVRY is normed on samples comprising mostly youth who are White, we expected its predictive validity for recidivism would be lower for Asian Canadians than White Canadians. Method: We examined youth probation officers’ SAVRY assessments for 573 youth (445 White Canadians, 56 East/Southeast Asian Canadians, and 72 South Asian Canadians) on community supervision (i.e. probation) in a Canadian province. Youth were prospectively followed for an average of 1.97 years (SD = 0.56 years) to determine if they were subsequently charged with violent or nonviolent offenses. Results: Asian Canadians scored significantly lower on risk total scores compared to White Canadians. Predictive validity for violent and nonviolent recidivism fell in the medium to large range for East/Southeast Asian Canadians (AUCs = .69 to .89) and South Asian Canadians (AUCs = .64 to .83). In comparison, predictive validity for White Canadians was generally lower (AUCs = .63 to .77; small to large range). Risk total scores and nonviolent risk ratings significantly predicted nonviolent recidivism better for East/Southeast Asian Canadians (AUCs = .89 and .87, respectively) than White Canadians (AUCs = .77 and .71, respectively). Despite few significant differences between Asian subgroups, predictive validity for nonviolent risk ratings was significantly higher in East/Southeast Asian Canadians (AUC = .87) than South Asian Canadians (AUC = .64). Conclusions: The SAVRY may be a useful tool for predicting recidivism with Asian Canadians. However, future research should examine the SAVRY’s predictive validity for youth of Asian descent in different countries and contexts.

SAVRY, risk assessment, predictive validity, recidivism, Asian Canadian

“Risk assessment tools are often used to assess the risk of recidivism among youth and adults in the justice system. Particularly in Western countries, these tools are used to guide numerous legal decisions, such as what treatment services youth receive and whether youth are detained prior to trial, transferred to adult court, and given a custodial placement following adjudication. Studies have found that risk assessment tools are more accurate in predicting reoffending than unstructured risk judgments. In addition, researchers and policymakers hypothesize that tools can help improve treatment- planning, reduce incarceration rates, and even reduce violence” (p. 485-486).

“Despite the potential benefits of risk assessment tools, policy-makers, legal scholars, and courts have expressed growing concerns that risk assessment tools may exacerbate racial and ethnic biases. For instance, Eric Holder (2014), a former U.S. attorney general, stated that risk assessment tools may ‘exacerbate unwarranted and unjust disparities that are already far too common in our criminal justice system and in our society’. Similar concerns have been raised in Canada. Recently, the Supreme Court of Canada ruled that federal correctional services are obligated to ensure that only appropriate and validated risk assessment tools are used with people who are Indigenous. In addition, a recent article in the Globe and Mail reported that Indigenous and Black Canadians are more likely than other groups to receive maximum risk scores upon federal prison entry, resulting in harsher restrictions and reduced access to rehabilitative services while in custody” (p. 486).

“Although there are growing concerns about the validity and impact of risk assessments with diverse populations, the use of risk assessment tools with Asian populations, particularly those in Western countries, have rarely been examined. To help address this gap, the present study examined the predictive validity of the Structured Assessment of Violence Risk in Youth (SAVRY) with a sample of Asian Canadian youth” (p. 486).

“As hypothesized, East/Southeast Asian Canadians and South Asian Canadians scored significantly lower than White Canadians on risk total, historical risk, and individual/clinical risk scores. However, even though YPOs [youth probation officers] rated Asian Canadian youth as lower risk, they were not significantly less likely than White Canadian youth to reoffend. As such, it is possible that YPOs may sometimes rate Asian Canadians as lower risk than warranted due to contemporary perceptions and myths about individuals of Asian descent. For instance, consistent with the model minority myth, YPOs rated Asian Canadians as showing better school achievement and commitment to school than White Canadians. The lower risk scores for Asian Canadian youth could be due, in part, to a cultural emphasis on family privacy. For in- stance, East, Southeast, and South Asian families might be less likely to disclose negative information, such as a history of child maltreatment” (p. 497).

“However, even if these factors may have led to attenuated risk scores for Asian Canadians, the SAVRY nevertheless showed strong predictive validity for the Asian subgroups. Contrary to our hypothesis, SAVRY risk total and summary risk ratings for East/ Southeast and South Asian Canadian youth generally fell in the medium to large range and there was a pattern toward higher predictive validity for Asian Canadian versus White Canadian youth. Further, South Asian Canadians with higher nonviolent ratings reoffended sooner than White Canadians. Thus, overall, YPOs were better at identifying which Asian Canadian youth would reoffend than they were for White Canadian youth. Al- though the reasons for this finding are unclear, the SAVRY may be picking up on cues that are especially salient for reoffending in Asian Canadian youth. For instance, although cultural factors such as high individualism are associated with delinquency among Asian American youth, peer delinquency (i.e. an item on the SAVRY) is a partial mediator between this association, and it remains one of the strongest predictors of delinquency among youth of Asian descent” (p. 496).

“In addition, much like Canada’s multicultural population, YPOs within the province had diverse backgrounds and included those who are Asian Canadian. Although we did not examine the racial/ethnic background of YPOs, it is possible that the YPOs in our sample may have been particularly sensitive to diverse cultural norms (e.g., body language) and thus better able to build rapport and interview Asian Canadian youth to elicit relevant information. A recent American study found that SAVRY risk scores did not differ based on whether youths’ race/ethnicity matched their YPOs’; however, it is not known if this has an impact on predictive validity” (p. 496).

“Furthermore, our predictive validity results were somewhat stronger than previous SAVRY research with Asian youth in Singapore and in China. For predicting violent recidivism, AUC values for SAVRY risk total scores and summary risk ratings fell within the small to medium range for youth in Singapore and China, while in our study, these AUCs for East/Southeast Asian Canadian youth fell in the large range” (p. 496).

“Although we hypothesized that YPOs would be more likely to incorrectly rate Asian Canadian youth as low risk, indicating an underestimation of risk, this hypothesis was not supported. For instance, nearly all the Asian Canadians that YPOs rated as low risk for violent recidivism did not commit a violent offense. Thus, even if the general public or YPOs hold stereotypes that Asian Canadian youth may be less likely to reoffend given the model minority myth, the use of the SAVRY might help to counteract these stereotypes by providing a standardized structure for evaluating risk” (p. 496).

“In general, the SAVRY’s predictive validity was similar for East/Southeast and South Asian Canadians. When the AUCs for these groups were compared, only one of the 12 comparisons reached significance. Specifically, nonviolent summary risk ratings were significantly higher in East/Southeast Asian Canadians. In addition, for predictions of nonviolent recidivism, NPVs (i.e. rated as low risk and did not reoffend) and PPVs (i.e. rated as high risk and went on to reoffend) were higher for East/Southeast Asian Canadians than for South Asian Canadians. However, as NPVs and PPVs are dependent on recidivism base rates, it is unclear if these findings would be replicated in other studies and it will be important to test whether results vary with Asian Canadian samples that have higher base rates” (p. 496).

“Based on the results of this study, the SAVRY appears to be a reasonable instrument to use for assessing recidivism risk among East, Southeast, and South Asian Canadian youth. However, despite results showing good predictive validity for Asian Canadian youth in general, this does not necessarily indicate that the distal causes for the SAVRY’s risk factors are the same among and within East Asian, Southeast Asian, South Asian, and White Canadian youth. There may be unique racial, ethnic, and cultural risk factors that result in the risk items that are subsequently captured within the SAVRY. For instance, with a Cambodian Canadian youth, the experience of being a refugee from a war-torn country may result in a high rating for stress and poor coping directly as a result from trauma. As such, in developing case formulations, it is important for assessors to try to identify the root causes of risk factors to allow for more effective intervention plans” (p. 497).

“Further, to help ensure that risk assessments do not inadvertently lead to ‘indirect discrimination’, assessors should remain vigilant of stereotypes that they might hold and work to counteract them. If assessors incorrectly assume that Asian Canadians are low risk due to the model minority myth, youth may not receive the supervision and interventions they need. Alternatively, if assessors inaccurately overestimate risk for Asian Canadians as a result of stereotypes about Asian gang involvement, youth may receive unwarranted restrictions (e.g., custodial placements). Although the SAVRY’s structured rating criteria might help assessors minimize the influence of stereotypes as compared to unstructured judgments, research has yet to directly test this possibility. Further, professionals with limited cultural knowledge may still introduce biases while using assessment tools” (p. 497).

“Thus, as recommended by Choi and Severson (2009), professionals who work with individuals of Asian descent must aim to increase cultural competency through education about cultural differences that depart from Western norms. Cross-cultural training may help to reduce stereotypes and promote more inclusive environments. For instance, involvement with crime is viewed as shameful among many Asian families, and parents may blame their child’s offending on parental failure. Given that collectivist orientations are prevalent among Asian cultures, family cohesion is often paramount. Consistent with recommended practices for adolescent risk assessments, assessors should take steps to ensure that parents/guardians are actively involved during interviews with youth justice professionals. This may include the use of interpreters for those who are not proficient in English” (p. 497).

“Rather than assuming that individuals of Asian descent make up a single homogenous group, a key direction for future research will be to attend to the diversity within the Asian subgroups. For instance, rather than grouping East and Southeast Asian Canadian youth together, researchers should examine these groups separately or, better yet, compare youth with different self-identified nationalities or regions of origin (e.g., Korean Canadians), pro- vided that sample sizes are sufficient to do so. In addition, al- though many Asian Canadians were born in Canada and have lived there all their life, some youth may have immigrated recently. Thus, researchers should attempt to incorporate acculturation factors into research, as acculturation may be an additional hurdle during adolescence for youth of Asian descent who reside in Western countries. Prior studies have demonstrated that first-generation immigrants, who are less acculturated, are less likely to reoffend or report delinquent behaviors compared to second-generation immigrants” (p. 497).

As always, please join the discussion below if you have thoughts or comments to add!