FAYETTEVILLE, Ark. — A second-grader who shoves other children on the playground and talks back to a teacher could be headed for increasingly serious problems in school and in life. But after aggressive children met for three semesters with a Lunch Buddy mentor, teachers reported to researchers that they could see changes.

Formerly disruptive children began to raise their hands in class or laugh with friends on the playground. What had happened? This was not the result the researchers had expected.

Timothy A. Cavell, professor of psychology at the University of Arkansas, in collaboration with Jan N. Hughes of Texas A&M University and graduate students, recently completed a study of the effect of school environment on mentoring programs that produced both this unanticipated outcome and questions to spark future research.

Evidence of the Lunch Buddy program’s effectiveness surprised researchers. Much has been written about the mentoring relationship as the mechanism by which mentoring actually benefits children, so Cavell and colleagues did not expect close mentoring relationships to form from a program set in a noisy school cafeteria with three different mentors.

“If the mentoring visits take place only in the lunchroom, and we bring in a new mentor every semester, how can strong relationships form? We had envisioned the lunch buddies to have little or no effect on our sample of highly aggressive children,” Cavell said. “We were wrong - either about there not being meaningful relationships or about the role of the relationship in mentoring.”

Cavell believes that the latter is true: “I believe that we helped these high-risk children because the consistent visits by a valued college student helped to improve how these children were viewed by their peers. Now, it may also be that the mentor supervised the children’s behavior during lunch or modeled positive peer interactions, but I suspect those were helpful only in the context of changing peers’ perceptions of a child that had been causing trouble because of his or her behavior.”

Although mentoring is now a popular approach to helping at-risk children, systematic studies of child mentoring programs are rare, and few have examined whether mentoring can help high-risk aggressive children. Cavell is trying to learn if - and how and why - child mentoring programs are effective and how the school environment can influence the outcomes of youth mentoring programs.

The study by Cavell and Hughes involved a range of schools. The researchers labeled some of them as high-adversity because of the level of playground aggression, the percentage of students entering and leaving the school during the year, and the percentage of students on federal free and reduced-cost lunch programs.

The researchers randomly assigned 174 second- and third-grade students identified as aggressive to one of two mentoring programs. The PrimeTime mentoring program was a structured program where the mentors received extensive training and supervision and mentored the same child -- usually outside of school hours -- for three successive semesters. The PrimeTime program also involved consultation with parents and teachers and social skills training for the children.

Lunch Buddy mentoring, on the other hand, was a stand-alone program and the mentors received no formal training before they met twice weekly with their mentees during lunch. Also, the lunch buddy who met with the child was different for each of the three semesters

Cavell and colleagues found that the two mentoring programs tended to produce different outcomes depending on the school context. The PrimeTime program did well in low-adversity schools, but the Lunch Buddy program fared much better in high-adversity schools.

The unanticipated effectiveness of the Lunch Buddy mentoring program has raised more questions. For instance, did the classmates sitting near a mentored student while a Lunch Buddy was present actually change how they viewed that child? And did those same classmates continue to be friendly on days when the mentor was not present? The researchers have just begun a study to see if interactions between mentored and non-mentored children are changed for the better on days when the Lunch Buddy is not there.

In addition to these questions about how mentoring actually works, Cavell emphasizes that many other questions are still to be answered in the field of child mentoring. Little research has been done to understand which problems are best addressed by mentoring and how to design mentoring programs that help rather than harm. In fact, in “Mentoring Children,” a chapter in the “Handbook of Youth Mentoring,” co-written with Anne-Marie Smith, Cavell calls for greater attention to the potential harm that can come from mentoring that is inconsistent or that ends prematurely. The two scholars also see the need for research that examines why mentoring works -- not simply if it works -- and that identifies which children are most likely to benefit from mentoring.

The article “Adverse School Context Moderates the Outcomes of Selective Interventions for Aggressive Children” by Jan N. Hughes, Timothy A. Cavell, Barbara T. Meehan, Duan Zhang and Claire Collie was published in the Journal of Consulting and Clinical Psychology. Cavell is a professor of psychology in the J. William Fulbright College of Arts and Sciences at the University of Arkansas.