Abstract
Students with significant behavioral and social problems experience some of the poorest outcomes in school and beyond. It is imperative, therefore, that educational researchers and school-based professionals address the needs of students who exhibit maladaptive behavior to alter their poor outcome trajectory. Social problem-solving (SPS) instruction is a promising approach for improving social competence and changing problem behaviors. Despite documented outcomes for SPS instruction in school settings, Coleman, Wheeler, and Webber’s review appears to be the most up-to-date compilation of the SPS literature. Thus, the purpose of this article is to present a more current review of the literature on SPS interventions in school settings. We examine and summarize studies investigating SPS interventions in K–12 settings from 1993 to 2015 and discuss findings and implications for educational research and practice.
Researchers have identified evidence-based practices to support the social, emotional, and behavioral needs of students with significant behavior problems, yet these students continue to experience poor outcomes, both short- and long-term (Kern, Hilt-Panahon, & Sokol, 2009). Students who display maladaptive behaviors typically have lower grades, higher rates of school dropout, employment difficulties, and criminality (Bradley, Doolittle, & Bartolotta, 2008; Wagner, Kutash, Duchnowski, Epstein, & Sumi, 2005). Students with significant emotional and behavioral problems tend to demonstrate high levels of inappropriate and challenging behavior (e.g., aggression, verbal abuse, lack of compliance; Reid, Gonzalez, Nordness, Trout, & Epstein, 2004). By definition, students identified with emotional and behavioral disorders (EBD) have difficulties with interpersonal relationships, exhibiting deficits in the social skills necessary for appropriate social interactions and competence (Kavale, Mathur, & Mostert, 2004). All these deficits are prime targets for effective evidence-based practices to increase self-control and, ultimately, social, behavioral, and academic success.
Social competence, or the ability to appropriately select and use social behaviors and skills to accomplish social tasks (Gresham, Elliott, Cook, Vance, & Kettler, 2010), is necessary for success in school and in life. Students with EBD, who lack social competence, have difficulties forming and sustaining interpersonal relationships with both peers and teachers (Gresham, Cook, Crews, & Kern, 2004). As a consequence, these students may have difficulties displaying prosocial behaviors, diminishing their social acceptance among peers. What is more, deficits in the core skills necessary for solving social problems have been found in students with a wide variety of problem behaviors (e.g., Elias & Tobias, 1996; Fraser, Lee, Kupper, & Day, 2011; Shure & Spivack, 1982). For example, studies have found that children and adolescents with significant behavioral difficulties attend less to relevant social cues, are inaccurate when interpreting peer intentions, select inappropriate goals, and generate fewer and more aggressive responses to interpersonal problems (Dodge & Price, 1994; Tisdelle & St. Lawrence, 1986). Essentially, youth with significant behavior problems demonstrate a range of deficits in social competence, including social problem solving (SPS). Given the negative outcomes associated with social skills deficits, such as delinquency, school failure, marital problems, and psychopathology (Cook et al., 2008; Parker & Asher, 1987; Segrin & Flora, 2000), students with significant behavioral problems are in desperate need of effective school-based interventions that explicitly target SPS abilities.
Social-Emotional Learning
In an effort to decrease students’ problem behavior and increase social competence, there has been a recent focus on programming that promotes social-emotional learning (SEL; Jones & Bouffard, 2012). Although it can be conceptualized in different ways, SEL programming targets the essential social and emotional skills (i.e., self-awareness, social awareness, self-management, relationship management, responsible decision making) necessary to succeed in school and effectively manage everyday tasks, such as solving problems and developing relationships (Zins, Bloodworth, Weissberg, & Walberg, 2007). Importantly, researchers have found that SEL programming can improve students’ behavioral, social, and academic outcomes, with some evidence of maintenance of gains at least 6 months after intervention (Durlak, Weissberg, Dymnicki, Taylor, & Schellinger, 2011).
Though SEL programming has been studied widely, Jones and Bouffard (2012) and Durlak et al. (2011) assert the importance of investigating the relative contributions of SEL programmatic components and approaches. For instance, although Durlak et al. and others (e.g., Greenberg, Domitrovich, & Bumbarger, 2001; Payton et al., 2008; Wilson & Lipsey, 2007) have investigated programming to improve a variety of SEL skills and behavioral outcomes, these reviews have targeted a broad set of social-emotional competencies and an array of SEL programs. As students matriculate through school, however, and experience increasingly complex social demands in new environments they must rely heavily on one core SEL skill, solving social problems. Thus, a more directed examination of curricula uniquely focused on a specific social-emotional skill, such as SPS, is warranted to examine the features, delivery formats, and outcomes unique to that SEL approach. As such, SPS is the focus of the current review.
Social Problem Solving
According to Nezu (2004), SPS is the cognitive-behavioral process used to find successful solutions to everyday problems. There are typically five recognized steps in SPS: (a) identifying that a problem exists, (b) defining the problem, (c) generating solutions, (d) evaluating solutions and enacting the chosen solution, and (e) assessing the outcome (Smith & Daunic, 2006). Thus, SPS is a purposeful, effortful, systematic, step-by-step process. When used as an intervention in school settings, SPS programming typically includes explicit instruction in SPS steps and skills through a combination of cognitive and behavioral techniques (Smith & Daunic, 2006). By using this approach, children are taught how to think rather than what to think, and they benefit by learning strategies and cognitive processes to control their own behavior without as great a dependence on external reinforcement. An underlying assumption of SPS instruction is that generalization is built in by teaching cognitive processes that can be applied across situations and settings, rather than teaching content specific to given situations (Urbain & Kendall, 1980).
SPS has evolved from a well-established line of research and practice that began during psychology’s cognitive revolution (Dember, 1974), with the introduction of problem-solving therapy in 1971 by D’Zurilla and Goldfried. Since then, other researchers have added to this pivotal work, both theoretically (e.g., Crick & Dodge, 1994; D’Zurilla, Nezu, & Maydeu-Olivares, 2004; Spivack, Platt, & Shure, 1976) and empirically (e.g., Daunic et al., 2012; Daunic, Smith, Brank, & Penfield, 2006; Fraser, Galinsky, & Smokowski, 2005), and have continued to refine and advance the knowledge of SPS over the past four decades.
The theory and empirical findings of SPS and its use in school settings have advanced incrementally, yet current summaries, critical evaluations, and interpretations of SPS are conspicuously absent. We found four prior reviews (cf. Coleman, Wheeler, & Webber, 1993; Denham & Almeida, 1987; Tisdelle & St. Lawrence, 1986; Urbain & Kendall, 1980) that informed the field but lacked a clear focus on SPS interventions for children and adolescents in school settings. Given that it has been more than 20 years since the last review of this literature, and no review has exclusively focused on SPS instruction in schools, there is a need to assess the current state of SPS programming in school settings.
Purpose
Since the last review conducted by Coleman et al. (1993), new and more sophisticated theoretical understandings and innovative approaches for implementing SEL in school settings have emerged and, at the same time, propelled the need to revisit the contemporary SPS literature to help researchers, practitioners, and policymakers meet the needs of students with social and behavioral problems. Thus, the purpose of our article is to present an interpretive summary of the literature on SPS interventions from 1993 to 2015 and to suggest recommendations for practice and future research. Specifically, we review the SPS literature to (a) identify important instructional features (e.g., theoretical backgrounds, delivery formats), (b) summarize findings in four outcome domains (i.e., behavioral improvements, SPS skill acquisition, emotion regulation, peer acceptance), (c) discuss the current state of SPS instruction, and (d) provide future directions for research and practice.
Method
Search Procedures
To identify appropriate articles, we conducted a search of four electronic databases: Educational Resources Information Center (ERIC), Education Full Text, Academic Search Premier, and PsycINFO. Within each database, we used the following keywords: interpersonal problem solv* OR social problem solv* AND school OR class*. Our selection criteria included articles that (a) were published in a peer-reviewed, scholarly journal from January 1993 to June 2015 (i.e., from the last review by Coleman et al., 1993, which included studies published from 1980 to 1993, to the present) and written in English; (b) reported the results of an empirical, quantitative study investigating a school-based SPS intervention, curriculum, or program for students with a clear description that SPS was the primary focus and component; and (c) was implemented in schools during school hours (not before or in after-school programs) in the United States in grades K–12.
We limited our review to articles published in peer-reviewed journals as an attempt to locate high-quality studies judged by experts in the field to provide the most valid and reliable evidence of intervention effectiveness. Educational legislation and policies require teachers to use evidence-based practices, and the peer review process is designed to ensure methodological rigor in the identification of educational techniques grounded in scientific research. Thus, we only searched for refereed articles published in scholarly journals. We also restricted our search to studies conducted in the United States because we were seeking implications specific to public awareness, educational programming, and local, state, and national legislation considering national cultural, social, and socioeconomic nuances.
Our search produced 84 articles from ERIC and 403 articles from the remaining databases, which we searched using EBSCO Host. After eliminating duplicates, 380 articles remained. We first read the article titles and abstracts to identify relevant articles; 80 articles met our inclusion guidelines (i.e., articles that described an empirical study of an intervention involving SPS in K–12 school settings in the United States). Following this initial screening, we subsequently examined the methods described by researchers in the 80 articles to obtain a more detailed consideration of each study and accurately determine whether it met our criteria. After careful inspection of each article using our inclusion criteria, we narrowed the total to twelve.
Next, we conducted a hand-search for potential relevant articles in all issues from January 1993 to June 2015 of the journals that produced more than one article from our initial search, including Elementary School Guidance and Counseling, the Journal of School Psychology, and the Journal of Primary Prevention. From this hand-search we identified no additional articles. We then searched the previously listed electronic databases for studies on the program I Can Problem Solve, which is widely cited as one of the first well researched school-based SPS curricula, and the programs Making Choices, Tools for Getting Along, and STORIES, which were identified by more than one study in the initial electronic database search. Using these SPS program names in a search of the electronic databases produced four articles that met our criteria. Finally, a search of all reference lists of selected articles resulted in two additional articles. A final total of 18 articles met our inclusion criteria.
Results
In this section, we report important features of the 18 identified studies. We first describe the sample across studies, the theoretical backgrounds, and the program descriptions and delivery methods of the SPS interventions. We then present outcomes of SPS instruction in four distinct domains: behavioral improvements, SPS skill acquisition, emotion regulation, and peer acceptance. We chose to discuss the findings based on these domains because they have been included in prior reviews (e.g., Coleman et al., 1993; Durlak et al., 2011), are important indicators of social-emotional and behavioral well-being, and are supported by theoretical notions about SPS (e.g., Crick & Dodge, 1994; D’Zurilla et al., 2004; Lemerise & Arsenio, 2000; Spivack et al., 1976). The measures used to assess these four distinct domains, along with other study features, are summarized in Table 1. Finally, as we present significant findings in the following sections, we provide effect size values, when available. Researchers in only 11 of the 18 studies included in our review reported effect sizes within the four outcome domains, using a variety of effect size measures and approaches.
Methodological features of studies
Note. SPS = social problem solving; SIP = social information processing; beh = behavior; gr = grade; obs = observations; Em Reg = emotion regulation; Peer Acc = peer acceptance; RA = random assignment; M = male; EBD = students with emotional and behavioral disorders; Y = yes; N = no; TR = teacher report on student behaviors; SR = student self-report; Peer = peer nomination.
Fraser et al. (2011) is a 6-month follow-up study of Fraser et al. (2005), and Smith et al. (2014) is a 1-year follow-up study of Daunic et al. (2012).
Sample Description
The total number of participants across studies was 4,985. Students’ grade level ranged from kindergarten to eighth grade, with students in third grade studied most often. Across studies, 51.58% of participants were male, with four studies including only males (Guevremont & Foster, 1993; Nelson et al., 1996; Quinn, 2002; Robinson et al., 2002) and one study including only females (Leff et al., 2009). Of those studies reporting participants’ race, the overall sample included the following: 33.36% Caucasian, 28.02% African American, 25.14% Hispanic, and 12.66% other or not reported. The remaining participants (n = 41, 0.82% of participants) were included in studies that reported race for the overall school population (Elias & Hoover, 1997; Nelson et al., 1996) or did not report students’ racial background (Guevremont & Foster, 1993).
Theoretical Foundations
Several researchers specified the SPS theoretical underpinnings of their programs, with an emphasis on theories by Shure and colleagues (i.e., interpersonal cognitive problem-solving skills; Spivack et al., 1976) and Dodge and Crick (i.e., social information processing theory; Crick & Dodge, 1994). For example, Nelson et al. (1996) based their program on the work led by Shure that began in the 1970s with the training of interpersonal cognitive problem-solving skills, and on Goldfried’s conceptualization of SPS (e.g., D’Zurilla & Goldfried, 1971; Goldfried & Goldfried, 1975) and Kanfer and Locke’s theory of self-regulatory performance (e.g., Kanfer, 1975; Locke, 1968). More recently, Boyle and Hassett-Walker (2008) also studied the theoretical work of Shure and colleagues by investigating their I Can Problem Solve curriculum. Conversely, the programs Making Choices (Fraser et al., 2005; Fraser et al., 2011; Fraser et al., 2014; Smokowski et al., 2004; Terzian et al., 2015), STORIES (Rahill & Teglasi, 2003; Teglasi & Rothman, 2001), Tools for Getting Along (Daunic et al., 2006; Daunic et al., 2012; Smith et al., 2014), and Friend to Friend (Leff et al., 2009) were theoretically grounded in Dodge and Crick’s social information processing theory (Crick & Dodge, 1994; Dodge & Crick, 1990). In essence, we noted that although two programs relied on the SPS theoretical perspectives pioneered by Shure, more recent SPS programming was based on Dodge and Crick’s work.
Program Descriptions and Delivery Methods
Over half of the included studies delivered instruction universally (i.e., whole class) to prevent behavior problems, especially for at-risk students or those with emerging behavioral difficulties (Boyle & Hassett-Walker, 2008; Daunic et al., 2006; Daunic et al., 2012; Fraser et al., 2005; Fraser et al., 2011; Fraser et al., 2014; Quinn, 2002; Smith et al., 2014; Smokowski et al., 2004; Terzian et al., 2015). Of the 10 studies using a universal delivery method, two programs were studied frequently—Tools for Getting Along (Daunic et al., 2006; Daunic et al., 2012; Smith et al., 2014) and Making Choices (Fraser et al., 2005; Fraser et al., 2011; Fraser et al., 2014; Smokowski et al., 2004; Terzian et al., 2015)—which included instruction on sequenced social information-processing steps and emotions. Tools for Getting Along’s 26 lessons focused on the application of SPS steps (recognizing a problem, problem definition and goal setting, generating possible solutions, selecting a solution, enacting a solution, evaluating the outcome) within social situations that are emotionally charged, providing instruction including cognitive modeling and practice through role-plays, small group activities, booster lessons, and real-life problems. Similarly, the Making Choices program taught SPS steps and skills (encoding, interpreting cues, goal setting, generating solutions, selecting a solution, enactment) through group activities, discussion, and role-plays, and an augmented version (Making Choices Plus) included teacher- and parent-involvement activities (e.g., additional emotion regulation strategy instruction, behavior management, letters to parents).
Boyle and Hassett-Walker (2008) investigated universally delivered I Can Problem Solve, which was well studied by its developers in the early 1980s and was highly influential to the field of SPS. The 83 I Can Problem Solve lessons targeted interpersonal cognitive problem-solving skills, such as solution generation, means-ends thinking, and consequential thinking, as well as emotion identification. The curriculum incorporated role-plays and games and used problem-solving dialoguing techniques throughout the day. Finally, Quinn (2002) delivered universal SPS instruction using a cooperative learning approach to teach interpersonal problem-solving strategies, including lessons on communication skills, anger management, and SPS (problem identification and goal setting, generating solutions, identifying consequences, creating a plan, evaluating the plan and outcome). The 26 daily lessons in Quinn’s study included modeling, role-plays, corrective feedback, and social reinforcement.
Fewer researchers, however, investigated interventions at the selected (i.e., small group, secondary intervention) or indicated (i.e., individualized, tertiary intervention) levels. Leff et al.’s (2009) Friend to Friend program included 20 small group sessions delivered to girls who displayed relational aggression and girls who were positive role models with a ratio of four to one, respectively. The small group sessions were supplemented by eight whole-class sessions that the participating girls helped to facilitate. Friend to Friend lessons included instruction on types of problems, emotion regulation strategies, how to evaluate problem responses, and the application of SPS strategies to situations involving gossip and joining peer groups.
Nelson et al. (1996) used 8 weeks of small group SPS counseling sessions with students referred for externalizing behaviors, using techniques such as questioning, discussion, modeling, role-plays, reflective listening, paraphrasing, and feedback. Sessions focused on teaching students that they have control over problems, how to identify problems, and SPS skills (e.g., goal setting, generating alternative solutions, consequential thinking) and allowed students to apply skills to problems at school. Researchers studying the STORIES program used a small group delivery format in both general education classrooms (i.e., small group instruction including targeted aggressive students and nonaggressive students; Teglasi & Rothman, 2001) and self-contained settings for students with emotional disturbance (Rahill & Teglasi, 2003). Taking a different approach, the STORIES program used a story structure that incorporated a bullying theme and followed SPS steps to reorganize students’ schemas (i.e., organized cognitive information related to past experiences) through experiential learning and peer discussion. Using a story format, students explored an emotional experience, considered others’ perspectives, and thought about possible solutions for problem situations.
Conversely, Lochman et al. (1993) used both small-group and individualized sessions for groups of aggressive and aggressive-rejected students in a social relations training program, which consisted of 26 individualized sessions including video-based instruction and 8 small group sessions for skill practice. Organized around a SPS framework, Lochman et al.’s program included four components (SPS, positive play training, group-entry skill training, dealing effectively with negative feelings) that used a variety of techniques, such as role-plays, goal setting, and social reinforcement. The problem-solving component taught sequential skills including identifying problems, goal setting, inhibiting impulsivity, and considering solutions and consequences.
Robinson et al.’s (2002) Anger Control Curriculum also targeted students with significant behavioral needs within self-contained settings but included SPS instruction delivered to the class rather than small groups. The 15-session Anger Control Curriculum explicitly taught SPS skills (problem identification, choosing alternative solutions, evaluating the outcome) in a whole group setting through modeling and practice, along with lessons on anger management, communication, and relaxation. For self-monitoring and generalization of skills, students kept a log to identify anger-provoking situations. Conversely, Elias and Hoover (1997) provided individualized SPS instruction through computerized software for students referred for extra support, which was supplemental to the universal SPS instruction provided in students’ health classes (not targeted or measured in their study). The software promoted generalization of SPS principles through computer-facilitated counseling and practice in step-by-step skills (identifying problems, goal setting, generating solutions, determining consequences, selecting a solution, planning actions and potential obstacles, evaluating how it went). Last, unlike any of the other studies included in this review, Guevremont and Foster (1993) conducted the only single-case design, and they individualized SPS instruction according to each student’s needs and readiness level. Across 18 sessions, students were taught five SPS steps (problem identification, generation of alternative solutions, evaluation of consequences, means-ends thinking, implementation), with each step presented after the prior skill was attained, and instruction included vignettes, modeling, behavioral rehearsal, and feedback.
Behavioral Improvements
Researchers investigating the effectiveness of SPS instruction consistently examined student behavior as an outcome measure of interest. In fact, behavior was the most widely studied outcome domain, with all 18 studies including measures of behavioral change. Upper elementary students who received teacher-delivered Tools for Getting Along demonstrated behavioral improvements across three studies. In Daunic et al. (2006), treated students targeted by teachers as being disruptive or aggressive had lower proactive and reactive aggression at postassessment compared with targeted control students, with the intervention explaining .353 and .169 of the between-group variance in mean proactive and reactive aggression scores, respectively. In a subsequent randomized controlled trial, Daunic et al. (2012) found that treatment students who were high risk at pretest demonstrated lower risk at posttest than control students on teacher-rated proactive aggression (g = 0.11). In addition, Daunic et al. (2012) measured student behaviors related to executive function (EF) and found significant improvements for initially high-risk students on behavior regulation (i.e., inhibiting impulses to shift ways of thinking and control emotions and behavior; g = 0.03) and metacognition (i.e., self-management and monitoring; g = 0.12) indices.
In a 1-year follow-up study, Smith et al. (2014) found similar but slightly reduced positive outcomes. Compared with control students 1-year posttreatment, students who were high risk at pretest had improved scores on subscales of internalizing and externalizing behavior, as well as improvements on the behavior regulation index of EF-related behaviors. For each of these improved scores, the condition and condition by pretest interaction explained approximately 1.6% to 2.5% of the remaining variance. There were also nonsignificant improvements for proactive aggression and for the Working Memory and Initiate subscales of the metacognition index for students who were high risk at pretest compared with controls. These studies demonstrated improved behavior, particularly for proactive aggression, for students receiving universal SPS instruction.
The Making Choices curriculum, delivered universally to third graders by graduate students and program specialists, also was associated with improved behavioral outcomes. Smokowski et al. (2004) discovered a significant direct effect of the intervention on overt aggression, social contact (i.e., interactions with others), and cognitive concentration (i.e., self-management of school work completion), with effect sizes ranging from .27 to .57. They also found a significant treatment by pretest interaction for social contact, cognitive concentration, and social competence outcomes (but not for overt aggression), demonstrating higher gains for students with lower pretest scores. Effect sizes were higher for students with low (poorer) pretest scores compared with the overall sample, ranging from .46 to .88.
In another study of Making Choices (Fraser et al., 2005), the intervention group showed improved social and overt aggression, as well as social competence and social contact, compared with the control group, with a range of effect size values from .17 to .67. For the treatment group that received teacher- and parent-involvement activities in addition to Making Choices lessons (Making Choices Plus), positive outcomes also included improved cognitive concentration (effect size of .43), with social contact approaching significance. In a follow-up of Fraser et al.’s (2005) study, Fraser et al. (2011) found that students receiving Making Choices instruction demonstrated significantly lower levels of overt, physical, and social aggression than the control cohort 6 months after treatment, with effect sizes from .09 to .14. Interestingly, there were no differences at the follow-up time between the treatment group receiving Making Choices and the group receiving the curriculum plus teacher- and parent-involvement activities (i.e., Making Choices Plus).
In a fourth study of Making Choices, Fraser et al. (2014) examined behavioral outcomes in terms of response to intervention as measured by pre–post changes in risk-profile status. Combining student data from Fraser et al. (2005) and the comparison group data from Terzian et al. (2015) for one Making Choices treatment group and one business-as-usual control group, Fraser et al. (2014) used person-centered statistical methods (latent profile transition analysis) to group third graders into four risk profiles based on teacher ratings of prosocial skills and aggression: high risk (high risk for prosocial skills and the highest aggression), moderate risk and high aggression (high risk for prosocial skills and high levels of aggression), moderate risk and low aggression (high risk for prosocial skills and low aggression), and low risk (low risk for prosocial skills and minimal aggression). Results indicated a significant relationship between Making Choices instruction and transition patterns across risk profiles, with an effect size of V = .321. Specifically, Fraser et al. (2014) found significantly more students receiving Making Choices improved risk status to a lower-risk group or remained in the lowest risk groups and significantly fewer regressed to a higher-risk group than students in the comparison condition. Fraser et al.’s (2014) study, therefore, demonstrates that Making Choices serves to both prevent worsening risk levels and remediate high-risk levels.
Finally, in the most recent study of Making Choices, Terzian et al. (2015) found significant improvements in treatment males’ aggression relative to the comparison group with an effect size of d = −.898, but females did not differ across groups. In fact, males’ improvements in aggressive behavior did not differ between those receiving the Making Choices curriculum or those receiving Making Choices Plus, the augmented version with teacher- and parent-involvement activities. In all, the outcomes of these five studies illustrate that instruction in Making Choices was associated with decreases in students’ problem behavior, especially overt aggression, and increases in prosocial and academic behavior.
The remaining studies of universally delivered SPS instruction included investigations of I Can Problem Solve (Boyle & Hassett-Walker, 2008) and a cooperative learning program (Quinn, 2002), which were the only two studies to investigate outcomes for students in lower elementary school. Teachers in Boyle and Hassett-Walker’s (2008) study delivered I Can Problem Solve universally to kindergarten and first graders, and the researchers compared outcomes for students receiving 2 years of treatment, 1 year of treatment, and control students. On the teacher-rated Preschool Social Behavior Scale, treatment students who received 2 years of instruction had significantly better overall behavior, less relational aggression, and more prosocial behavior than those who received 1 year of instruction or were in the control condition. Two-year treatment students also had significantly lower overt aggression compared with control participants; yet 1-year treatment students did not differ significantly from either 2-year or control students. Effect sizes indicated that the treatment explained 3% to 7% of the variance in the scores on these behavioral scales. On another teacher-rated behavioral measure (Hahnemann Behavior Rating Scale), both 1-year and 2-year treatment students had significantly better overall and prosocial behavior, with effect sizes demonstrating that treatment explained 6% and 12% of behavioral scores, respectively; but there were no significant differences between treatment or control groups on aggression/impulsivity and passivity subscales. Overall, this study provides some evidence that students who received universal SPS instruction in I Can Problem Solve experienced behavioral improvements compared with control students, with better outcomes for students who had 2 years of instruction.
Quinn (2002) investigated the effects of teacher-delivered universal SPS cooperative learning instruction, and all analyses examined outcomes specifically for first-grade boys nominated by teachers for antisocial behavior. Behavioral observations showed that targeted treatment students were significantly more engaged in academic settings than control students. There were no significant differences, however, for observed peer social behavior on the playground or teacher-rated externalizing antisocial behavior. Thus, Quinn’s study had limited behavioral outcomes for younger students who received universal SPS instruction through cooperative learning.
There were also notable behavioral outcomes in three studies in which researchers investigated selected SPS instruction delivered to small groups of students (Leff et al., 2009; Nelson et al., 1996; Teglasi & Rothman, 2001). In their study of the Friend to Friend program, Leff et al. (2009) compared effect size differences in teacher-reported aggression and social behavior change scores between girls with relational aggression in treatment and control groups and found that girls with relational aggression who received Friend to Friend instruction had greater decreases than those in the control group in both relational aggression (effect size of .74) and physical aggression (effect size of .43). In Nelson et al.’s (1996) study of small group counseling for second-, fifth-, and sixth-grade students referred by teachers for externalizing behavior, overall scores on a teacher-rated behavioral scale, as well as all subscale scores, were significantly better at posttest for treatment students than control students. Importantly, scores for the treatment group fell into the normal range for behavioral adjustment, whereas control students’ scores remained more than one standard deviation above the mean, indicating significant behavior problems. In essence, Nelson et al. found that small group counseling sessions targeting SPS skills effectively improved behavioral outcomes for students displaying externalizing behavior.
In Teglasi and Rothman’s (2001) study, upper elementary students received 15 weeks of small group sessions in the STORIES program, although the authors did not specify who delivered the instruction. Each group consisted of four to six students, with one or two students identified as aggressive (i.e., nominated by teachers and the principal for exhibiting aggressive behaviors). When examining pre–post differences for all students receiving STORIES instruction, teachers reported an overall decrease in students’ externalizing behavior. An interaction effect, however, demonstrated that externalizing behavior increased for aggressive students but decreased for students not identified with aggression following treatment. Interestingly, on a pretest measure of externalizing behavior, teachers rated all students as having elevated aggressive behavior. Thus, although students nominated as having the most aggressive behavior did not improve behaviorally after STORIES instruction, other students with preexisting aggression did benefit from instruction. Nominated aggressive students, however, had lower teacher-reported externalizing behavior and self-reported normative beliefs about aggression after receiving STORIES instruction than comparable students in the wait-list control group. Thus, there were positive behavioral changes for all students after receiving small group instruction in STORIES, and among only the most aggressive students, the intervention significantly improved externalizing behavior.
Researchers in two studies in which SPS instruction was delivered individually (Elias & Hoover, 1997; Guevremont & Foster, 1993) also examined behavioral outcomes for students with behavior problems. Elias and Hoover (1997) used individualized SPS instruction through computerized software for 12 third graders who were referred to the lab by teachers or counselors. The researchers found that students receiving treatment and matched controls both demonstrated improved behavior control, with no significant differences between the groups. All students in the study received universal SPS instruction during weekly health classes (which was not targeted or measured), and the treatment group received the supplemental, individualized SPS instruction, making it difficult to draw conclusions about behavioral outcomes due to the SPS computer lab.
In a single case design study, Guevremont and Foster (1993) examined individualized training in SPS steps delivered by graduate students to five 11- and 12-year-old boys referred by teachers for aggressive and disruptive behavior. All five participants showed reductions in teacher-rated aggressive behavior after treatment, to such an extent that three students fell into the nonclinical range for aggressive behavior, and two fell into the nonclinical range for total problem behavior scores. The researchers also collected observational data on aggressive and disruptive behavior during baseline, SPS training, and the introduction of a SPS log consisting of students’ daily written reflections on their use of SPS across settings. Results of the observational data for the five participants were mixed: One student steadily improved aggressive and disruptive behavior immediately after SPS training; three students demonstrated improved behavior only after the SPS log (i.e., skill generalization) was introduced; and one student did not show any behavioral improvements. At a 6-month follow-up, only one of the five participants demonstrated maintenance of gains on aggression and behavior problems. Although conclusions are tentative due to limited studies, the mixed findings and lack of maintenance outcomes indicate less favorable behavioral outcomes for individualized SPS training compared with other delivery methods, yet some teacher-rated results were positive.
Lochman et al. (1993) examined behavioral outcomes for SPS instruction delivered through a combination of individualized and selected small group SPS instruction. In their study, students were screened through peer nomination as rejected and aggressive. Nonaggressive-rejected and aggressive-rejected students were randomly assigned to the treatment group and received instruction in a social relations training program using a SPS framework. Aggressive-rejected treatment students had significantly lower levels of teacher- and peer-rated aggression after receiving treatment than control group counterparts. Importantly, group differences were not observed for nonaggressive-rejected students. At 1-year follow-up, aggressive-rejected treatment students had lower levels of aggression and higher teacher-rated prosocial skills than comparable control participants. There were, however, no significant effects on peer ratings at follow-up or any significant differences for nonaggressive-rejected students. Thus, Lochman et al.’s study demonstrated that individualized SPS programming can be beneficial for aggressive students when supplemented with small group instruction.
Researchers in only two studies investigated outcomes for students receiving services in self-contained settings for more significant behavioral needs (Rahill & Teglasi, 2003; Robinson et al., 2002). Robinson et al. (2002) explicitly taught SPS skills in a whole group setting to middle school students with EBD. Students receiving treatment demonstrated significant differences compared with the control group on several measures related to self-reported anger and teacher-rated aggression, yet these findings were not maintained at a 4-week follow-up. In contrast, Rahill and Teglasi’s (2003) STORIES program used a story format in a small group setting to improve SPS skills and reorganize preexisting schemas for second through sixth graders with EBD. Of note is that Rahill and Teglasi are the only researchers who compared outcomes of the STORIES program with other treatments—the Skillstreaming social skills program and a nonspecific counseling group. Although STORIES students demonstrated significantly less problem behavior than students in the social skills group, they did not differ from students in a nonspecific counseling group. The researchers reported an effect size (η2) of .098 for overall problem behavior. The findings from these two studies suggest that students with more significant behavioral needs in self-contained settings may derive better benefit from explicit SPS skill instruction that includes practice opportunities.
Social Problem-Solving Skill Acquisition
The second most examined outcome domain was SPS. Not only is knowledge of SPS foundational to behavioral change (Bandura, 1999), but verifying improvements in students’ abilities to manage social problems as a result of intervention is necessary as well (Coleman et al., 1993). Although all studies included SPS instruction, only 10 of the 18 measured specific SPS knowledge or skills (Daunic et al., 2006; Daunic et al., 2012; Elias & Hoover, 1997; Fraser et al., 2005; Guevremont & Foster, 1993; Leff et al., 2009; Robinson et al., 2002; Smith et al., 2014; Teglasi & Rothman, 2001; Terzian et al., 2015). In a study of the universally delivered Tools for Getting Along, Daunic et al. (2006) found that treatment students scored higher than control students on curriculum-based knowledge about SPS, with an effect size indicating that instruction explained .418 of the between-group variance in mean SPS knowledge scores.
In a subsequent randomized controlled trial, Daunic et al. (2012) found a significant positive main effect on positive problem orientation (g = 0.11) and rational problem solving (g = 0.16) subscales of the Social Problem-Solving Inventory–Revised (D’Zurilla, Nezu, & Maydeu-Olivares, 2002). Daunic et al. (2012) also found that students with lower pretest curricular SPS knowledge benefited more from training than students with higher pretest knowledge (g = 1.88). In a 1-year follow-up study, Smith et al. (2014) found that although the treatment group continued to have better curriculum-based knowledge of SPS compared with control students (SPS knowledge scores explained 14.6% of the remaining variance), findings related to positive problem orientation and rational problem solving were not substantiated. In sum, studies on Tools for Getting Along show that this program can improve SPS knowledge and skills, with improvements in SPS curricular concepts maintained 1 year postintervention.
In a study of another universally delivered SPS intervention, Making Choices, Fraser et al. (2005) found that intervention students had better social information-processing skills (i.e., encoding cues, forming prosocial goals) on a story-based assessment of hypothetical situations than control students, with effect sizes of .82 for encoding and .28 for goal formulation. These findings were broadened to include other skills (hostile attribution, response decision) when the intervention included additional teacher- and parent-involvement activities, with effect sizes from .54 to .77. Similarly, using a different control group, Terzian et al. (2015) found improvements in several SPS skills for students receiving instruction in Making Choices, including greater response decision and lower hostile attribution for both boys and girls and improved goal formulation only for boys. Although all treatment students evidenced significant improvements, those students receiving the augmented version Making Choices Plus (curriculum supplemented with teacher- and parent-involvement activities) reported significantly greater improvements in all three SPS skills than those students receiving the standard curriculum. Effect sizes ranged from .213 to .469 for students receiving Making Choices and from .452 to .853 for students receiving Making Choices Plus. In essence, from studies of Making Choices and Tools for Getting Along, it appears that universally delivered SPS programs can improve students’ SPS related outcomes.
In addition to programs delivered universally, researchers investigating SPS programming delivered in selected and indicated formats found mixed results. For instance, Leff et al. (2009) found that after Friend to Friend instruction, relationally aggressive girls had greater improvements in hostile attributional biases in situations with relational problems compared with similar girls in the control group, with an effect size of .61. Additionally, Guevremont and Foster (1993), employing a single-case design and individualized SPS instruction, found that SPS skills could be acquired quickly when taught to aggressive and disruptive male students, as measured by student responses to vignettes of hypothetical problem scenarios. More specifically, they found that students demonstrated improved number and quality of generated solutions, and detailed and sequential means for achieving solutions, following training. Similarly, Robinson et al. (2002) found that on average, students with EBD who received whole-group SPS instruction in self-contained settings could recall 74.3% of the curricular content, suggesting gains in SPS knowledge.
Conversely, Teglasi and Rothman (2001) did not find a significant change in targeted aggressive students’ self-reported normative beliefs about aggression (i.e., cognitive standards on the acceptability of aggression) after students received small group STORIES instruction designed to change existing schemas that influence SPS. Elias and Hoover (1997) also did not find significant differences between treatment and control groups. Although Elias and Hoover found that treatment students attending the individualized SPS computer lab (supplemental to universal SPS instruction) improved in self-reported SPS skills, specifically in the areas of controlling behavior and emotions, they found that matched students who did not attend the lab also made gains. In sum, of the 10 studies measuring SPS outcomes, there is evidence for improved SPS knowledge and some support for improved SPS skills, especially when SPS instruction is delivered universally and taught through explicit instruction.
Emotion Regulation
Many researchers across studies view emotion regulation as an essential feature of SPS instruction, as evidenced by their inclusion of emotion identification and control techniques in SPS programming. In fact, researchers in 11 studies (Boyle & Hassett-Walker, 2008; Daunic et al., 2006; Daunic et al., 2012; Fraser et al., 2011; Leff et al., 2009; Lochman et al., 1993; Quinn, 2002; Robinson et al., 2002; Smith et al., 2014; Smokowski et al., 2004; Terzian et al., 2015) described an emotion regulation component as part of their intervention, specifically targeting participants’ anger and frustration. For example, Robinson et al.’s (2002) intervention incorporated lessons on anger, anger management, and relaxation techniques, and Boyle and Hassett-Walker’s (2008) study included instruction on identifying feelings, particularly those that produce problematic situations.
All other studies targeting emotion regulation included it as a major unit of instruction that occurred early on in the programs’ instructional sequences. For instance, the first unit of instruction in Making Choices (Fraser et al., 2005; Fraser et al., 2011; Fraser et al., 2014; Smokowski et al., 2004; Terzian et al., 2015) was on emotion regulation strategies. Similarly, Tools for Getting Along (Daunic et al., 2006; Daunic et al., 2012; Smith et al., 2014) included recognizing and controlling anger in the first two problem-solving steps within SPS instruction, and lessons specifically targeted situations that would be emotionally charged. Quinn (2002) also included lessons on anger management early in SPS instruction, with lessons occurring immediately prior to lessons on SPS skills, and Leff et al.’s (2009) Friend to Friend included instruction on identifying and managing aggression as an initial intervention component. Furthermore, Lochman et al.’s (1993) intervention provided emotion regulation instruction, which taught aggressive-rejected students how to effectively deal with powerful and negative emotions, immediately after the unit on SPS. Clearly, researchers made emotion regulation a primary component in SPS programming.
Although emotion identification and regulation skills were taught frequently as a part of SPS instruction, researchers in only six studies measured outcomes related to emotion regulation (Daunic et al., 2006; Daunic et al., 2012; Elias & Hoover, 1997; Robinson et al., 2002; Smith et al., 2014; Terzian et al., 2015), and some found significant improvements in students’ emotional control. For instance, Robinson et al. (2002) discovered significant differences between treatment and control students at posttest on self-reported state anger (i.e., intensity of anger), angry temperament, angry reaction, anger out (i.e., expressing anger), and anger control (i.e., attempting to control anger expression). Anger control, however, was the only significant outcome maintained at a 4-week follow-up. Similar to their findings for behavioral outcomes, Elias and Hoover (1997) found that both treatment students attending the SPS computer lab and matched control students who did not receive computerized instruction showed improved emotion control abilities, with no significant group differences.
Additionally, although Daunic et al. (2006) in their study of Tools for Getting Along did not find significant effects for student self-reported anger disposition and anger control, findings from Daunic et al.’s (2012) randomized controlled trial indicated significant improvements at postassessment for treatment students compared with controls on student-reported anger levels (g = 0.11) and outward anger expression (g = 0.04). A 1-year follow-up study, however, did not find evidence of improved anger ratings (Smith et al., 2014). Finally, Terzian et al. (2015) did not find significant outcomes for students receiving instruction in the Making Choices program on a measure of emotion regulation. Overall, these mixed findings demonstrate some evidence for SPS interventions to improve students’ emotion regulation, particularly anger.
Peer Acceptance
Researchers in six studies also included measures of peer status or acceptance (Elias & Hoover, 1997; Fraser et al., 2014; Guevremont & Foster, 1993; Leff et al., 2009; Lochman et al., 1993; Smokowski et al., 2004). Across these studies, both teachers and students reported improved acceptance by peers following SPS training. Smokowski et al. (2004) found that compared with control students, treatment students had higher teacher-rated peer acceptance after receiving universal SPS instruction in the Making Choices program, with an effect size of .45 for the overall sample and .69 for those at-risk students with poorer pretest scores. Furthermore, in another study of Making Choices, Fraser et al. (2014) found that instruction both prevented worsening risk levels and decreased high-risk levels for third-grade students on risk profiles comprised of several variables, including peer acceptance. Conclusions are tentative, however, given that teacher perceptions of acceptance by peers only constituted two items from the 37-item measure from which risk profiles were derived. Additionally, Leff et al. (2009) found that girls who displayed relational aggression had greater improvements in teacher-rated peer likeability after instruction in Friend to Friend than comparable control students, with an effect size of 1.73. Lochman et al. (1993) also found improvements for aggressive-rejected students in individual and small group sessions of SPS instruction as indicated by lower levels of teacher-rated rejection and more positive peer-rated social acceptance following treatment than the comparable control group.
There were less encouraging results for SPS instruction delivered individually. Three of the five students in Guevremont and Foster’s (1993) study demonstrated improved scores on a teacher-reported measure of social rejection, but at a 6-month follow-up, only one of the five participants demonstrated maintenance of gains on peer acceptance. Also, Elias and Hoover (1997) did not find significant differences between students who attended the SPS computer lab and matched comparison students on the sociometric measure (i.e., peers’ judgments of target students’ behavior). Similarly, Rahill and Telgasi’s (2003) findings on teacher- and peer-rated sociometrics were nonsignificant when comparing three interventions (SPS, social skills, nonspecific counseling) in self-contained settings for students with EBD. On the whole, although some studies that included peer sociometric ratings indicated that SPS instruction could improve peer acceptance and status, there is less support for SPS instruction delivered individually, in a computer lab, or in self-contained settings for students with EBD.
Discussion
Despite advances in instructional approaches and theoretical perspectives, it has been more than 20 years since scholars have examined the literature on SPS interventions. The purpose of our review, therefore, was to conduct an interpretive summary of the literature on SPS interventions from 1993 to 2015. As such, our review differs from prior research syntheses of SEL programming (e.g., Durlak et al., 2011) by focusing exclusively on the characteristics and outcomes unique to SPS interventions, thus contributing to a better understanding of programs targeting a specific social-emotional skill. Overall, the results from our review demonstrate positive outcomes related to the acquisition of SPS skills and mixed, yet encouraging, findings for improving behavioral, emotional, and peer acceptance outcomes. Specifically, the evidence suggests that training improves SPS knowledge, and there is emerging evidence that it improves problem-solving skills.
The findings also reveal encouraging behavioral outcomes, especially when delivered at the universal level. The strongest evidence was related to behavioral improvement for at-risk students, older students, and outcomes specifically targeting aggression. Though there was less conclusive evidence for improved behavioral outcomes resulting from selected or indicated SPS instruction, improvement was noted in small group counseling sessions or when individualized and small group instruction were combined. There is also some evidence that SPS interventions improved students’ emotion regulation, especially anger, and their acceptance by peers. In the studies investigating maintenance of gains across time, researchers discovered continued effects of SPS instruction up to 1 year later for some, but not all, outcomes. Finally, several studies showed that SPS programming appears to be most effective for those students who are low performing behaviorally and socially prior to intervention. Although some findings were unclear or inconsistent across these 18 studies, the review indicates that SPS is a promising intervention worthy of implementation and further research. Generally, SPS instruction improved students’ behavioral and social outcomes, especially for students with or at-risk for significant behavior problems.
Magnitude of Effect
Although the intent of our review was not to conduct a meta-analysis, it is important to examine the effect sizes reported by researchers to determine the magnitudes and strengths of SPS instruction. We found that the 11 studies that reported effect sizes were published after 2003 and varied in terms of which effect sizes were used (e.g., proportion of variance explained, standardized mean difference effect size, pseudo-R2 statistics, Hedges’ g), thereby limiting our ability to compare effects across studies (Durlak, 2009). Standardized effect sizes (excluding variance-accounted-for effects) tended to range from .09 to 1.73 across studies. Of the researchers that reported effect sizes, most were for behavioral and SPS outcomes. Conclusions about effect sizes for emotion regulation and peer acceptance outcomes, however, are tentative given that for these domains, only one or two studies, respectively, included effect sizes.
Researchers have asserted that interpreting effect sizes must be done in the context of other studies using similar interventions, samples, and outcomes rather than applying a one size fits all standard criteria (Durlak, 2009; Hill, Bloom, Black, & Lipsey, 2008; Lipsey et al., 2012). The effect sizes reported, however, in the included studies are comparable to other meta-analyses measuring similar constructs. For example, Lösel and Beelmann’s (2003) meta-analysis examined the effects of skills training programs (which included SPS) that aimed to prevent antisocial behavior for children and youth and found mean weighted effect sizes of .38 for overall outcome measures, .26 for antisocial behavior, .39 for social skills, and .40 for social-cognitive skills (including SPS).
Also, in their meta-analysis of universally delivered SEL interventions, Durlak et al. (2011) reported effect sizes of .57 for SEL skills, .24 for positive social behavior, .22 for conduct problems, and .24 for emotional distress. Finally, in a meta-analysis of psychosocial interventions delivered in school settings to improve students’ aggressive and disruptive behavior, Wilson and Lipsey (2007) found an overall mean effect size of .21 for aggressive/disruptive behavior, and effect sizes for other outcomes of interest (e.g., social skills, problem behaviors, knowledge and attitudes) ranged from .20 to .35. The findings of our review, therefore, indicate that when compared with prior studies of research with similar interventions and outcome variables, SPS interventions can result in a comparable magnitude of effect for social and behavioral outcomes, further supporting their use in school settings. To provide a more accurate estimate of effects and strengthen conclusions, however, it will be essential that researchers conduct meta-analyses of SPS interventions in future studies.
Instructional Features
From the review, we found key instructional elements that are essential for effective SPS programming, including step-by-step instruction, a focus on emotion regulation, and attention to generalization. Even though SPS can include components such as cooperative learning, social skills training, and goal setting, programmatic elements common among studies highlight the importance of the former three essential elements. For example, the majority of researchers directly taught SPS skills through an explicit step-by-step cognitive-behavioral process, using techniques such as structuring behavioral goals and contracts, using modeling, rehearsal and skill practice, and providing feedback. The review also demonstrates that emotion regulation instruction, in particular, appears to be an important part of SPS intervention, especially when included early on in the instructional sequence. Last, it is important to incorporate strategies to promote generalization and maintenance into SPS instruction, such as student journaling to record problematic situations encountered and booster lessons for the application of problem-solving skills to real-life problems and scenarios.
Although most SPS interventions incorporated similar instructional features, the developers of STORIES (Rahill & Teglasi, 2003; Teglasi & Rothman, 2001) took a somewhat different approach. The STORIES program did not include instruction on SPS steps and skills, but used peer interaction and discussion of a story structure involving SPS to change schemas that affect students’ social information-processing abilities. Thus, the STORIES intervention incorporated an approach that differed markedly from the other reviewed studies and resulted in less favorable outcomes. Though direct comparisons across studies are problematic, it appears that an explicit approach to teaching component skills may be more effective.
Instructional Delivery
Interestingly, of the delivery formats used in the 18 studies, most provided SPS instruction at the universal level. Our findings indicate that universal instruction is effective in preventing and improving students’ social and behavioral outcomes, especially for students at risk for behavior problems. In fact, researchers acknowledged the benefits of universal delivery, including (a) the prevention of behavior problems and poor outcomes for at-risk students, (b) avoidance of potential issues associated with pull-out programs (e.g., lack of natural reinforcement, hindered generalization, stigmatization, lack of positive peer models), and (c) the use of peer supports through modeling of socially appropriate and competent behavior (Daunic et al., 2006; Daunic et al., 2012; Quinn, 2002; Smokowski et al., 2004).
Although fewer studies used selected or indicated delivery methods, there is some evidence for the effectiveness of small group instruction but limited findings for individualized instruction. In fact, Guevremont and Foster (1993) cautioned that individually delivered instruction might hinder student gains, as SPS is a social activity and necessitates involvement with peers. Researchers found encouraging results, however, from a combination of individualized instruction with small group practice. Given that rehearsal of SPS skills during training requires interaction with others, supplementing individualized instruction with activities involving peers, such as small group practice, may enhance outcomes.
Instructional Personnel
In general, when researchers used universal delivery methods, they provided training for the teachers who implemented the intervention. Graduate students and program specialists were also used to deliver universal SPS instruction, however, limiting the ability to make conclusions about the feasibility of delivery by school-based personnel. Similarly, studies of instructional delivery at selected or indicated levels typically involved nonschool personnel (i.e., graduate students) or did not specify who the interventionist was. It is essential to investigate studies incorporating school personnel as interventionists to assess the feasibility of implementation within the school environment (Barnes, Smith, & Miller, 2014). Researchers, therefore, must continue to examine the effectiveness of SPS interventions delivered by school personnel, especially at the selected and indicated levels.
Student Characteristics
The findings of our review suggest that student characteristics are important to consider for SPS programming. First, 6 of the 18 studies included mostly male participants. Although students with EBD are typically male (Bradley, Henderson, & Monfore, 2004), less is known about SPS outcomes for female students with or at-risk for behavior problems. Leff et al. (2009) included only female participants to examine improvements in relational aggression. Additionally, it is noteworthy that even though studies included a range of participant ages (Grades K–8), most studies included students in Grades 3 to 5, with third graders studied most frequently. The findings of the review for younger students, therefore, are tentative, given that the only two studies including participants in lower grade levels (K–1) had limited outcomes. Importantly, SPS programming using a cognitive-behavioral approach is typically more effective for students beginning around age eight (O’Connor & Creswell, 2008); thus, cognitive-behavioral interventions may be more effective at the upper elementary level. In sum, much remains to be known about SPS instruction across gender and age groups.
Theoretical Advances
Social Information Processing Theory
Since the last review of SPS interventions by Coleman et al. in 1993, there have been important theoretical advances. Of the studies included in the review that described the intervention’s theoretical background, social information processing (SIP) theory was most prevalent. Crick and Dodge (1994) described the cognitive processes that lead to children’s behavioral responses, and these processes are well aligned with the SPS steps. According to SIP theory, competent performance in a social situation is the product of skillful cognitive processing of social cues across six sequential mental steps. Lemerise and Arsenio (2000) stressed the role of emotion within SIP, which is reflected in SPS interventions that include instruction on emotion identification and regulation. The findings of our review thus demonstrate that theoretical advances have informed developments in SPS instruction that reflect a shift over time from the early work of Shure and colleagues (e.g., Spivack et al., 1976) to a more advanced and comprehensive model that links cognitive and emotion control processes to behavioral outcomes.
Self-Regulation and Executive Function
A nascent line of research indicates that executive functions are foundational for effective self-regulation programming such as SPS (e.g., Blair & Diamond, 2008; Liew, 2012; White, Jarrett, & Ollendick, 2013). Instruction in self-regulation, or the ability to intentionally monitor, modify, and manage emotions, thoughts, attention, and behavior (Blair & Diamond, 2008; Gardner, Dishion, & Connell, 2008), is thus a promising approach, as researchers have found that students with behavioral difficulties often have self-regulation deficits (e.g., Kavale et al., 2004; Polsgrove & Smith, 2004; White et al., 2013). Executive functioning (EF) is an umbrella term for neurocognitive abilities that are recruited when an individual engages in intentional, goal-directed behavior and problem solving (Gioia, Isquith, Kentworthy, & Barton, 2002), central to self-regulation. Essentially, when using SPS skills, an individual engages in self-regulation and simultaneously recruits EF (e.g., shifting mental sets, inhibiting prepotent responses, holding and using information in working memory). Given the abundance of recent research and the theoretical link to SPS, self-regulation and associated EF offer a novel framework from which to investigate SPS.
The findings from this review offer insights into interventions fostering self-regulation and EF skills through SPS instruction. In two recently published studies of SPS instruction, Daunic et al. (2012) and Smith et al. (2014) included measures of EF and found improvements in several student EF-related abilities (e.g., self-management of behaviors, impulse control, shifting cognition, initiating tasks) after receiving instruction in Tools for Getting Along, with some gains maintained 1 year later. These results indicate the potential positive impact SPS instruction can have on contextualized EF that stems from underlying neurocognitive mechanisms associated with self-regulation, and they provide evidence that researchers must continue to investigate the relationship between SPS instruction and EF.
Limitations of Current Research
Assessment
In 1993, Coleman et al. emphasized the importance of including measures of SPS skills, behavioral ratings, and behavioral observations to examine outcomes of SPS training, yet only one study included in our review followed these guidelines and included all three types of measures (Guevremont & Foster, 1993). Despite the fact that all studies included behavioral ratings, only two studies conducted behavioral observations. Furthermore, as noted by Smokowski et al. (2004) and Daunic et al. (2012), studies should include a variety of informants to provide comprehensive support for findings. Most studies incorporated teacher-reports and some included student- and peer-ratings, yet others included measures from only one informant source. Furthermore, none of the studies we reviewed included parent-reported measures about the transfer of school-based instruction to other settings and occasions. Thus, there is a need for studies on SPS that incorporate multiple measures and informant sources.
Social Problem-Solving Measures
Many researchers measured outcomes related to SPS abilities, yet the measures used varied considerably. Notably, only one study (Daunic et al., 2012) and its 1-year follow-up (Smith et al., 2014) assessed SPS skills with a widely used SPS instrument with adequate psychometric properties (Social Problem-Solving Inventory–Revised; D’Zurilla et al., 2002). Although three studies (Fraser et al., 2005; Guevremont & Foster, 1993; Leff et al., 2009) included vignettes adapted from previously used measures, most of the SPS measures across studies were created or adapted by researchers for purposes of a specific study. Nezu (2004) acknowledged the common practice of developing measures of SPS for research, yet D’Zurilla et al. (2004) pointed out that many researchers fail to describe adequately how SPS measures were constructed or provide information on their psychometric properties. It is essential, therefore, that researchers continue to use measures with strong psychometric properties and to investigate appropriate instruments to gauge students’ SPS abilities.
It is striking that we found only 18 studies conducted over the past 22 years that fit our selection criteria, examining SPS interventions for students in K–12 school settings in the United States; yet we acknowledge that our review is not without limitations. We constrained our review to articles published in peer-reviewed journals as an effort to find only high-quality studies evaluated by experts in the field, thereby excluding gray literature (i.e., unpublished studies, doctoral dissertations, conference reports, book chapters) and introducing a potential for publication bias. Additionally, our search strategies, although designed to adequately capture all applicable studies, may have unintentionally missed relevant studies. Thus, there may be further knowledge about SPS interventions that is not included in the results of our review. We believe, however, that despite these limitations, our interpretive review provides an updated synthesis of what is currently known about SPS interventions delivered in K–12 schools in the United States. As such, the findings of this review offer important implications for research and practice.
Implications for Research and Practice
Although the findings of this review allow for positive conclusions about the effects of SPS programming in school settings, some questions remain. Researchers must continue to investigate SPS interventions to meet the needs of students with social and behavioral problems. First, theoretical advances should continue to inform SPS interventions. Li, Fraser, and Wike (2013) suggested that intervention research must move from the early SPS theories to SIP theory, which more fully explains the effects of sequenced cognitive processes on behavior. Furthermore, researchers should continue to identify and investigate interventions that improve the self-regulatory abilities and underlying EF skills of students with social and behavioral problems. Given the growing research base on self-regulation and associated EF, researchers should investigate the use of direct training of EF skills within SPS interventions. Emerging research demonstrates the potential ability of direct training to improve EF (e.g., Diamond & Lee, 2011; Klingberg, 2010), which may transfer to SPS skills as well. According to Denham and Almeida (1987), students with more significant behavioral needs (e.g., aggression, impulsivity) might need modifications to SPS programming, and a focus on enhancing EF skills may provide an effective enhancement.
In future studies, researchers must also address current limitations within the literature. Studies involving female participants and students from a variety of ages would contribute to findings, as would investigations of the role of gender, race, socioeconomic status, and/or academic achievement in student outcomes, as these characteristics may impact the effects of SPS intervention (Daunic et al., 2012). Additionally, less is known about the impact of SPS programming on outcomes for students with the most significant behavioral problems. Students receiving services in self-contained settings, who may need more intensive interventions to remediate social and behavioral deficits, seemed to benefit more from explicit instruction in SPS and emotion regulation strategies than instruction designed to alter preexisting schemas through a story structure and peer discussion. Given the limited number of studies involving these students, however, more research including participants with EBD in self-contained settings is needed to determine SPS components most effective for these students.
In future investigations of SPS instruction, researchers should also study a variety of intervention methods to increase understanding of SPS programming. For instance, researchers could examine various delivery formats for SPS programming, especially those using small groups. For students with or at-risk for behavioral and social problems, more intense and targeted intervention may be required. Thus, researchers should examine the instructional features, delivery methods, and programmatic components most effective for students who need additional supports. Moreover, researchers should continue to conduct follow-up and longitudinal studies to investigate long-term effects of SPS instruction. To address the efficacy of SPS training and support for implementation in school settings, researchers should include assessments of professional development and social validity (i.e., participants’ reports of an intervention’s acceptability).
Additionally, studies should include adequate intervention descriptions to allow for replication and measures of treatment fidelity to ensure that outcomes can be accurately and appropriately attributed to the intervention (Smith, Daunic, & Taylor, 2007). Researchers should also pay particular attention to Coleman et al. (1993) who advised using a variety of assessments. It is important to continue to create and validate SPS measures with adequate psychometric properties, as well as to include assessment of student SPS deficits prior to intervention. Finally, it is essential that researchers calculate and report effect sizes to help understand the magnitude of the effect of SPS instruction.
The current literature on SPS also provides implications for practice. First, interventions should be incorporated within multi-tiered systems of support to fully address students’ social and behavioral needs. Administrators and teachers can use SPS curricula across tiers of instruction to create a school and classroom climate that supports skill development and meets various levels of student needs. Importantly, given the variety of personnel used to deliver instruction, collaboration among school professionals is important for effective SPS programming across tiers of intervention. In the classroom, skills to support the management of social problems can be taught using role-plays, brainstorming sessions, and problem-solving journals. Teachers can use cognitive modeling with their students to think aloud how to solve real-world problems within natural contexts and settings and thus foster generalization. Additionally, programming should not only include a focus on emotion identification and regulation but also foster EF skills. Finally, pre- and inservice trainings should adequately prepare teachers to use cognitive-behavioral interventions, specifically those targeting SPS. Essentially, by incorporating these suggestions into practice, the needs of students with behavior problems will be more adequately supported, which will ultimately result in better student outcomes.
Conclusion
It is clear that SPS can be used effectively to support the social and behavioral needs of students with or at-risk for significant behavior problems. We agree with Coleman et al. (1993) who posit that researchers encounter many barriers when examining the effectiveness of SPS interventions. Yet we assert that researchers can and should tackle these challenges. It is vital that researchers and practitioners continue to revisit SPS as a potential tool to remediate and prevent students’ problem behaviors and social deficits. Past and current research on SPS interventions may hold the key to improving the future for students with significant social and behavioral needs.
Footnotes
Authors
KRISTEN L. MERRILL is a doctoral candidate in special education at the University of Florida, PO Box 117050, Gainesville, FL 32611, USA; email:
STEPHEN W. SMITH is a professor of special education at the University of Florida, PO Box 117050, Gainesville, FL 32611, USA; email:
MICHELLE M. CUMMING is a doctoral candidate at the University of Florida, PO Box 117050, Gainesville, FL, 32611, USA; email:
ANN P. DAUNIC is a scholar in the College of Education at the University of Florida, PO Box 117050, Gainesville, FL 32611, USA; email:
