Program Profile: Evaluative Conditioning for Adolescent at Risk for Violence (South Korea)

This program used conditioning methods to change the way at-risk adolescents implicitly think about violence. The goal was to train participants to associate violence with negative, instead of positive, objects or ideas. The program is rated Promising. Adolescents who received evaluative conditioning demonstrated a statistically significant increase in negative implicit attitudes toward violence, compared with the control group.

A Promising rating implies that implementing the program may result in the intended outcome(s).

A Promising rating implies that implementing the program may result in the intended outcome(s).

Program Goals/Target Population

Evaluative Conditioning (EC) for adolescents at risk for violence in South Korea was a strategy intended to modify adolescents’ implicit, or underlying, attitudes toward violence through a reflective conditioning process. Youth were identified as at-risk after receiving in-school counseling services for committing an act of school violence. Through this process, youth were trained to associate violence concepts (such as assaults, threats, or fights) with negative attributes (such as disgusting, terrible, or failed) and to associate peaceful concepts (such as harmony, care, or cooperation) with positive attributes (such as valuable, wonderful, or confident). The goal was to change positive attitudes toward violence into negative views by conditioning the youth to associate violence with negativity and peace with positivity. The target population comprised adolescents in middle and high schools in South Korea who demonstrated violence-supportive cognitions (Jeon, Lee, and Lee 2019).

Program Theory

A risk factor for violent behavior in adolescents is the maladaptive cognition of violence during adolescence (Huesmann and Guerra 1997). That is, some adolescents’ impulsive thoughts about violence may cause them to behave in ways that interfere with coping effectively in society. The dual-process, social-cognitive theory of aggression suggests that the cause and continuation of violent behavior can be explained through the existence of explicit and implicit cognition (Szlachcic et al. 2015). Explicit attitudes are thoughts and beliefs that are consciously formed, are deliberate, and can be reported, whereas implicit attitudes are formed on the subconscious level and are difficult to articulate or self-report. This theory suggests that implicit cognition that is supportive of violence works though automatic processing and increases access to aggressive thoughts when interpreting social situations (Greenwald and Banaji 1995). Thus, if individuals already have subconscious positive associations with violence, they may exhibit violent responses more easily and quickly than more peaceful responses in general social situations.

EC was based on the associate-propositional evaluation (APE) model, a dual-process model of attitude formation. The APE model suggests that explicit and implicit attitudes are rooted in two different types of mental processes. Explicit attitudes are formed through slow propositional processes, and implicit attitudes are formed though quick association (Gawronski and Bodenhausen 2011). According to this theory, explicit attitudes can be controlled with cognitive effort and often reflect what is socially acceptable in one’s environment. However, implicit attitudes are difficult to control or change because they are triggered by associative cues without any conscious awareness. Therefore, even if individuals learn to modify explicit attitudes through conscious reasoning, implicit violence-supportive cognitions could be evoked by exposure to cues associated with violence through an absence of conscious control.

EC was based on the principle of classical conditioning, whereby a conditioned response is elicited following repeated pairing of neutral stimuli with unconditioned stimuli. However, this program’s approach was to replace unconditioned stimuli with positive or negative stimuli to change implicit attitudes toward violence (Baeyens, Field, and Houwer 2005).

Program Activities

The program was conducted through a go/no-go task, so that participants did not notice the intent of the intervention. A go/no-go task is a test that requires the subject to either respond or not respond in order to measure inhibition. For this intervention, each adolescent was required to press the “space” key on a computer in a “go” trial, and to refrain from pressing the “space” key in a “no-go” trial. The program activity first consisted of seven blocks, each with different themes of items as go or no-go indicators. The themes included “animal” (which was a practice block), “fruits,” “country,” “city,” “sports,” “home appliances,” and “clothing.” Participants were instructed to either press or not press the space key when items appeared to correspond with the specific theme. Additionally, participants were shown conditioning pairs, which included one target word (i.e., violence-related words such as “threat,” or peace-related words such as “harmony”) and one word with negative or positive connotations (e.g., “nasty” or “beautiful,” respectively). When participants indicated a “go” for violence and positive connotation words, they were considered to have a positive implicit attitude toward violence. Conversely, when participants indicated a “go” for violence and negative connotation words, they were considered to have a negative implicit attitude toward violence. The idea was that the more quickly participants responded (i.e., by pressing the space bar), the stronger their implicit attitude toward violence. In other words, because the association already existed in the participants’ minds, their reaction time would be reduced.

Study

Jeon, Lee, and Lee (2019) conducted a randomized controlled trial to evaluate the effectiveness of evaluative conditioning (EC) on at-risk adolescents’ implicit attitudes toward violence. The study participants were drawn from middle and high schools in South Korea and were considered eligible for the study if they were between the ages of 13 and 18 and had been determined to be at-risk youth who had been called before the school violence committee at least once. The final sample for analysis consisted of 100 juvenile participants, 50 who have committed of an offense and 50 who have not. The 100 participants were randomly assigned to the treatment group (n = 50) or the control group (n = 50). There were no statistically significant differences between the groups on baseline characteristics.

Implicit attitudes toward violence were measured using the Implicit Association Test (IAT). The IAT measured automatic associations between concepts and attributes by instructing participants to perform classification tasks as quickly as possible. Participants categorized words into two paired groups of concepts and attributes as consistent with the theory that the more closely associated concepts and attributes are, the easier they could be categorized together quickly. For example, if a participant held a positive attitude toward peace but a negative attitude toward violence, the individual would quickly categorize peace and positive together rather than violence and positive because the association already existed. The strength of association in an IAT score was measured by difference in reaction time for two counterpart concepts.

For this study, a violence-IAT was designed with the target concepts of “violence” and “peace” and the attributes “positive” and “negative.” Forty-eight words related to violence, peace, negative, and positive (12 words per category) were selected from the Standard Korean Dictionary by the National Institute of the Korean Language. During the task, the category words (violence, peace, positive, and negative) were presented at the top left or right of the screen, and the related words were presented in the center of the screen in random order. A higher IAT score indicated negative attitudes toward violence. “Filler” pairs consisted of the affectively neutral words also selected from the Korean dictionary, which were paired with words from themes such as “animal,” “fruits,” and “home appliances.” The treatment group completed a series of 24 trials, which consisted of eight meaningful conditioning pairs (four violence-negative pairs and four peace-positive pairs) and 16 filler pairs (pairs of words related to the theme and neutral images). The control group completed an intervention that consisted only of filler pairs.

Participants completed the pre- and posttests immediately before and after completing EC. The implicit attitudes toward violence of the offender and nonoffender groups were compared using independent samples t tests. A three-way mixed analysis of variance (ANOVA) with an assessment point (pre- and posttest) as a within-subject variable group (offender and nonoffender) and condition (control and treatment) as between-subjects variables was conducted to assess the intervention effect of EC. The study authors did not conduct subgroup analyses.

These sources were used in the development of the program profile:

These sources were used in the development of the program profile:

Baeyens, Frank, Andy P. Field, and Jan De Houwer. 2005. “Associative Learning of Likes and Dislikes: Some Current Controversies and Possible Ways Forward.” Cognition and Emotion 19(2):161–74.

Gawronski, Bertram, and Galen V. Bodenhausen. 2011. “The Associative-Propositional Evaluation Model: Theory, Evidence and Open Questions.” Advances in Experimental Social Psychology 44:59–127.

Greenwald, Anthony G., and Mahzarin R. Banaji. 1995. “Implicit Social Cognition: Attitudes, Self-Esteem, and Stereotypes.” Psychological Review 102(1):4–27.

Huesmann, L. R., and N.G. Guerra. 1997. “Children's Normative Beliefs About Aggression and Aggressive Behavior.” Journal of Personality and Social Psychology 72(2):408–19.

Szlachcic, Rebecca, Simone Fox, Clare Conway, Alex Lord, and Alison Christie. 2015. “The Relationship Between Schemas and Offence Supportive Attitudes in Mentally Disordered Sexual Offenders.” Journal of Sexual Aggression 21(3):318–36.