Outcomes for methods to accelerate thinking skills involving some peer interaction have been more consistently positive than those for purely teacher-directed or materials-led methods. However, methods involving mainly or only peer interaction are rare. This paper describes and evaluates such a method for peer tutoring in thinking skills, which scaffolds interactive discourse based on a differentiated real book the tutorial pair has chosen to read together. This pilot study aimed to partial the impact on quality of thinking of a peer-tutored thinking intervention from that of a peer-tutored reading intervention, controlling for time on task and amount of peer interactivity. Experimental peer tutees were a whole class (n =28) of seven-year-olds; experimental tutors a whole class (n =31) of 11-year-olds. Comparison tutees were a whole class (n =27) of seven-year-olds; comparison tutors a whole class (n =30) of 11-year-olds. Classes/teachers within the same school were randomly assigned to conditions. In Phase 1, a paired reading intervention was implemented for six weeks for all groups. In Phase 2, the experimental classes of tutors and tutees engaged in the "paired thinking" (PT) method for 10 weeks, while the comparison group continued with paired reading. Both treatments involved one 20-minute session weekly. Pre- and post-test assessment of thinking skills and attitude to reading for all participants was conducted and post hoc subjective feedback gathered from participants. The experimental (PT) tutees showed significantly better performance in thinking skills than comparison (PR only) tutees, and some evidence of improved attitudes to reading. However, this was not true for the experimental (PT) tutors. Subjective feedback was very positive from the PT tutees and class teachers, but less positive from the PT tutors. Given the brevity and low cost in time and resources of the treatment, the finding of significant differences in measured thinking skills for the PT tutees is considered encouraging. Recommendations for refining organizational aspects of the implementation of PT and for future research are made.