The current thesis builds upon developments in the field of Relational Frame Theory (RFT), which has proposed a behavioural re-examination of the widely used concept of intelligence in terms of derived relational responding (DRR). In the first chapter the concept of intelligence is explored theoretically from a RFT perspective. A framework for the construction of interventions to raise intelligence quotients as calculated by standardised IQ tests is also provided. Specifically, the current thesis proposes that training skills in DRR by utilising multiple exemplar training (MET) can improve intellectual skills. In Experiments 1 and 2, it is shown that, by employing a MET intervention for symmetry and transitivity, modest rises in full scale IQ on the WISC-IIIUK were generated for normally functioning children. In Experiment 3, the MET intervention methodology is further developed across a group of both children and adults to specifically improve the relational skills which appear to underlie intelligence as a behavioural repertoire. The newer methodology is shown to generate repertoires of Same and Opposite relational responding for experimental groups of adults and children, where these repertoires were previously weak or absent. In Experiments 4 and 5, the stimulus control of the intervention is further improved. Experiments 6 and 7 involve the addition of intervention protocols for More- Than/Less-Than relational responding. MET is again shown to facilitate the emergence of DRR for Same and Opposite (Experiments 4 and 5), and also for More-Than/Less- Than (Experiments 6 and 7) with both child and adult groups. However, Experiments 6 and 7 failed to clearly establish the functional dependence of More-Than/Less-Than responding on Opposite relational responding alone. In Experiment 8, participants with an extended history of MET across symmetry, transitivity, Same and Opposite showed rapid acquisition of More-Than/Less-Than DRR. Experiment 9 measured considerable rises in WISC-IIIUK scores across an extended MET intervention for four children. Importantly, similar rises were not seen for a matched control group who had no access to the intervention. In Experiment 10a, a relational abilities index (RAI) is developed for use as a baseline relational skills index. This baseline measure is then correlated with the WISC-IVUK and its subtests for a group of children with learning difficulties (Experiment 10b). Several interesting correlations between relational skills and intelligence are observed in the resulting analysis, although many theoretical and conceptual issues are also suggested by the data. In Experiment 11, a complete MET battery is administered to an educationally challenged child group. Both RAI and full scale IQ scores rise from pre to post intervention. In the closing chapter, the implications of these rises for intellectual disability, educational psychology and behaviour analysis are discussed.