This pilot study examined the effectiveness of child-centered play therapy (CCPT) with 54 low income preschool children (42% African American, 39% Hispanic, and 18% Caucasian) identified with clinical levels of disruptive behaviors. Children were assigned to the experimental treatment, CCPT, or to the active control, reading mentoring (RM), through block randomization. Results from a 2 (group) x 3 (measures) split plot analysis of variance indicated that according to teachers blinded to children's treatment group assignment, experimental group children demonstrated a statistically significant decrease (p .05) in disruptive behaviors compared with the RM group across 3 points of measurement. Post hoc analysis showed that children receiving CCPT demonstrated statistically significant decreases (p .025) in aggression and attention problems. Regarding practical significance of the findings, CCPT demonstrated large [partial eta squared] treatment effects ranging from .17 to .34. Of the 27 children receiving CCPT, 21 children moved from clinical levels of behavioral concern at pretest to more normative levels of functioning following treatment, thus establishing the clinical significance of CCPT as an early mental health intervention for at-risk children in Head Start. (author abstract)