Beyond translation: A deeper dive into the cultural considerations of risk-assessment in Latin America

Alicia Nijdam-Jones, Fordham University
Eric García-López, Instituto Nacional de Ciencias Penales
Maria Aparcero, Fordham University
Barry Rosenfeld, Fordham University

The ability to predict the likelihood of violent behavior is an important task that is not constrained by cultures, borders, or geographic regions. This qualitative study analyzed 15 semi-structured interviews with forensic professionals to understand violence risk assessment (VRA) processes and procedures in Argentina, Colombia, and Mexico. Participants reported a range of procedures for assessing violence risk, including the use of structured VRA tools. Most evaluators considered common risk factors, but several culturally relevant risk and protective factors emerged from the interviews. Perceived strengths and limitations of VRA measures were explored. Findings may inform risk assessments with culturally diverse adults.

Cross-cultural, forensic assessment, violence, Latin America, risk assessment

“Amidst a growing interest in violence risk assessment worldwide, Latin American clinicians have begun to use more structured approaches to risk assessment over the past three decades. However, the existing literature suggests that despite the advancements in the assessment of violence risk in Latin America, the application of structured professional judgment tools is only slowly increasing. One international survey of 2135 forensic professionals found that of the 57 South American forensic practitioners, fewer than half reported using any risk assessment tool in the previous year”. (p.1)

“In the latter study, only half the sample (n = 35, 49.3%) reported using any psychological measures to inform violence risk judgments, and the most commonly used tools were projective measures (n = 18, 25.4%); only 12 respondents (16.7%) reported using structured professional tools for violence risk assessment. These studies suggest that unstructured clinical risk assessment is still common in Latin America, despite advances in violence risk assessment worldwide. This leads to questions regarding the methods professionals use to evaluate risk of violence and why some professionals in these countries are still not utilizing structured approaches to violent risk assessment.” (p.2)
“Another consideration when conducting violence risk assessments in Latin America is that the available risk assessment tools may have weaker predictive accuracy than when used with predominantly non-Hispanic White samples. In fact, several studies raise questions about the accuracy of risk assessment tools when specifically applied to Latino/a adults in the United States. The reasons for this decreased accuracy are not clear but may include a failure to incorporate culturally relevant risk factors in the evaluation process.” (p.2)

“Most violence risk assessment measures have used an etic approach when used cross-culturally, whereby the measure is translated into another language and then evaluated to determine whether it has adequate psychometric properties. However, the etic approach to violence risk assessment risks omitting culturally relevant risk factors that may be important in understanding violent behavior, which would likely result in lower predictive validity when used in culturally diverse contexts. To maximize utility and accuracy, a violence risk assessment tool should also incorporate culturally relevant risk factors and adaptation. Thus, incorporating an emic perspective to instrument development and validation, in which existing tools are modified by local norms, values, beliefs and practices, may improve our understanding of violence risk.” (p.2)

“This study had three specific goals: (1) Understand the process and procedures participating clinicians use to assess violence risk with Spanish-speaking Latin
American adults; (2) Identify the culturally relevant violence risk and protective factors participants consider when completing these assessments; and (3) Elicit participants’ perspectives on the strengths and limitations of the violence risk assessment measures they use in their practice.”

Method
Participants
“Individuals were eligible to participate in this study if they: (1) were employed as a psychologist, psychiatrist, or social worker; (2) worked in Argentina, Colombia, or Mexico; (3) had completed a violence risk assessment in the preceding 12 months; and (4) agreed to have their interview audio-recorded for transcription purposes. A total of 15 participants were included in the study, and each country was equally represented.” (p.2)

Procedures
“The interview guide was collaboratively developed by all of the authors, and included a range of open and close-ended questions intended to elicit narratives relating to participants’ approach to violence risk assessment, the risk assessment tools they use, and the risk and protective factors they consider in their evaluations. In the first section, participants were asked information specific to their experience completing violence risk assessments with Spanish-speaking Latin American adults. This included open-ended questions about the context and process of the violence risk assessments they complete, as well as the risk and protective factors they consider in their evaluations. Participants were also asked if there were specific risk or protective factors they believed were unique to the population(s) they work with (i.e., based on cultural or geographic norms, values, and beliefs). In the second section, participants were queried on which risk assessment measures they use, why they selected those measures, and the strengths and limitations of each tool.” (p.3)

“Interviews lasted an average of 49.40 minutes (SD¼15.16, range: 30 to 79 minutes). To ensure accuracy, transcriptions were done by two native Spanish-speaking research assistants using a denaturalized approach that aimed to capture verbatim statements but not speech errors, pauses, or involuntary vocalizations” (p.3)

“This study used a hybrid process of inductive data driven codes and deductive theory-driven codes in the analysis of the qualitative data. Deductive codes were developed based on the violence risk assessment literature, including the Historical-Clinical-Risk Management-20 risk assessment scheme for violence risk factors and the Structured Assessment of Protective Factors for violence risk (SAPROF) for protective factors.” (p.4)

“Using a deductive etic approach to coding with the HCR-20 and SAPROF factors facilitated organization of the data and helped identify potentially universal risk and protective factors. However, aligned with the study’s emic goals, the descriptive exploratory analysis also required an inductive coding approach, and thus emerging patterns, themes, and concepts were coded and analyzed. Including both deductive and inductive codes allowed for the analysis to incorporate existing theories and current knowledge about violence risk and protective factors, while also facilitating the exploratory qualitative description that accurately reflects the participants’ culturally relevant knowledge and experiences” (p.4)

Results
“The first group included participants (n = 6, 40.0%) who indicated that they did not use not use any formal standardized violence risk assessment measures in their evaluations. These participants completed risk assessments in the context of government provided services (e.g., court-ordered evaluations, victim services), and they indicated that they relied on data derived from a clinical interview, with some noting that they used organizational forms to guide their interviews. These participants reported that their evaluations were often completed without access to external records to review, although contact with collateral informants was occasionally sought.” (p.5)

“The second group of participants (n = 9, 60.0%) reported having used a violence risk assessment instrument to inform their evaluations within the previous 12 months. These participants were typically psychologists, were usually retained by defense attorneys, and reported spending extensive periods of time gathering relevant data and writing reports. In addition to relying on clinical interviews, these clinicians typically reviewed clinical and criminal justice records, sought collateral informants, administered psychological testing (e.g., personality assessment measures), and used a structured professional judgment measure or actuarial assessment tool.” (p.5)

“Participants were asked what risk factors they considered in their evaluations of violence risk. Common risk factors listed included a history of violence, relationship problems, substance abuse, psychopathy or personality disorder, violent attitudes, instability, history of traumatic experiences, history of treatment for behavioral problems, and a limited social support network.” (p.5)

“…participants also identified several risk factors that were specifically relevant to the populations they worked with in Latin America. These included: (1) machismo, (2) problematic relationships with family and peers, (3) normalization of violence, and (4) economic disadvantage.”

“Participants were also asked about the protective factors that they consider when evaluating risk for violence. Social network and personal support were the protective factors most often mentioned by study participants, along with the individual’s living situation (e.g., removal from negative context or influences), work, and education. Similar to their discussion of risk factors, participants placed more emphasis on external factors that may reduce one’s risk of engaging in future violence rather than internal factors (e.g., coping, self-control). Although social support, employment, and education may be universal protective factors for violence, it was notable that participants put significant emphasis on the role of family within the context of Latin American cultures.” (p.7)

“In addition to machismo beliefs held by the individual, participants spoke about systemic sexism in the justice system, where cases of violence against women are often not prosecuted effectively, if the cases are even brought to the attention of the government. The limited enforcement of gender-based violence was viewed as a factor that perpetuated these cognitive distortions among men and foster continued gender-based violence in Latin American countries.” (p.6)

“In addition to modeling criminal or violent behavior, families were also described as directly encouraging some individuals to engage in antisocial behaviors: ‘In this city, the question of narcotics and cartels are very prominent right now, but I must also recognize that many times the families themselves encourage the subjects to commit crimes. Then this comes out during the interview, this information is provided and once we get it, we have the ability to determine whether the pattern will continue or if they will withdraw from this behavior.’ (Mexico #2)” (p.6)

“…participants reported that the extensive history of violence in both Mexico (e.g., the drug war) and Colombia (e.g., decades of civil conflict) has led to a normalization of violence in the daily lives of the people they evaluate.” (p.6)

“For some, violence was seen as a learned response, extending beyond simply what is taught through interactions with peers and family, but something that is taught through knowledge translation at the national level” (p.7)

“The majority of participants (n = 11, 73.3%) discussed aspects of familismo as a protective factor, referencing a strong sense of community, unity, and reliance on family for instrumental and social support. Large families that were free of conflict were viewed as a resource, helping the individual with physical, social, and financial burdens” (p.7)

“…participants reported difficulty accessing tools that have been developed outside of Latin America, as only tools that have been translated into Spanish were accessible to them. With the exception of Argentina where a group of researchers have translated and studied violence risk assessment measures, the majority of participants reported using tools that had been translated into Spanish by researchers in Spain. Participants noted that relying on Spanish translations filters the tools professionals have access to and delays their use in Latin America.” (p.8)

“Participants also noted that they were unaware of any evidence to support the predictive accuracy of these measures in their country (e.g., Mexico and
Colombia), raising concerns regarding their use in legal contexts.” (p.8)

“Concern was also voiced regarding the cross-cultural validity of some of the risk factors these tools incorporate to evaluate violence risk. For example, several participants noted that some risk factors on the HCR-20, such as problems with treatment or supervision response, cannot be reliably ascertained in their countries. Specifically, the resources available for community treatment and supervision are limited, and thus, information on whether or not a person has or would adhere to mandated services is not relevant to their evaluations of violence risk.” (p.8)

“…training in psychological assessment differs in Latin America from other countries, such the United States, Canada, and countries in Europe. These differences were perceived as impacting the conceptual equivalence of some risk factors when they are evaluated in Latin America. For example, when considering the relevance of personality disorder diagnoses, one participant noted that many psychologists in Argentina do not use standardized approaches for evaluating or diagnosing personality disorders, and unless similar empirically-based theories or criteria are used, this risk factor might be evaluated differently…”(p.8)

“More research is therefore needed to determine whether culturally relevant risk factors should be integrated into existing items (i.e., by modifying the rating criteria for different cultures) or whether “new” items should be added that have some overlap. Although the former may be more efficient, it may also increase the risk that these culturally specific risk factors are not given as much consideration as they would if treated as “new” risk factors.” (p.11)

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