The current thesis examined the use of the Implicit Relational Assessment Procedure (IRAP) as a measure of spider fear, approach, and avoidance. Participants were drawn from a normative sample of university undergraduates. Study 1employed The Fear of Spiders Questionnaire (FSQ), a Behavioral Approach Task (BAT) using a spider moult, and a Rule Focussed IRAP (RF-IRAP). Both spider fear and approach/avoidance were measured simultaneously. Overall, patterns of responding revealed predictable IRAP effects with an implicit negative bias recorded for spider fear and also spider approach. However, the RF-IRAP had limited predictive validity in terms of the self-report measure and also in terms of the BAT. The next two studies therefore employed the traditional IRAP. Study 2 employed two IRAPs, one targeting spider fear, the other targeting spider approach/avoidance. The Fear of Spiders Questionnaire (FSQ) and a Behavioral Approach Task (BAT) using a spider moult were also employed. Negative response biases for spider fear and avoidance, but not for approach, were recorded. The bias for fear was significantly stronger than for avoidance and approach. Both IRAPs failed to provide evidence for the predicative validity of the IRAP in terms of the BAT. Study 3 was a partial replication of Study 2 but using a live house spider instead of a moult for the BAT. A similar pattern of results was obtained across the two IRAPs but one specific trial-type (Spider-Approach) predicted approach responses on the BAT. The research thus replicated a previously published study by Nicholson and Barnes-Holmes (2012), thus supporting the predictive validity of the IRAP but, at a level of precision not provided in the earlier studies. Implications for applied research are considered.