Attentional biases reflect an individual's selective attention to salient stimuli within their environment, for example an experience of back pain. Eysenck suggests that different personality types show different attentional biases to threatening information. This study is the first to test Eysenck's theory within a chronic back pain population by investigating the attentional biases of four different personality types using a back pain specific dot-probe paradigm. Participants were 70 volunteers (45 female) recruited from a back rehabilitation program at an NHS Trust. The four groups were selected on their trait anxiety and defensiveness scores: defensive high-anxious; high-anxious; repressor and non-extreme. Participants completed a dot probe task comprising 20 practice trials and 250 experimental trials. The experimental trials contained 100 threat-neutral pairs, 100 positive-neutral pairs and 50 neutral-neutral image pairings. The threat images were taken from the Photograph Series of Daily Activities (PHODA) and the neutral and positive images from the International Affective Picture System (IAPS) image bank. The results provided partial support for Eysenck's theory; defensive high-anxious individuals showed an attentional bias for threatening information compared to high-anxious individuals who demonstrated no bias. Repressors showed an avoidant bias to threatening images and an attentional bias to positive stimuli relative to neutral images. The clear difference in responses demonstrated by high-anxious individuals who vary in defensiveness highlight the need for separate investigation of these heterogeneous groups and help to explain the cognitive processes of defensive high-anxious individuals within a pain population. The demonstration of an attentional bias in this group to threatening information could explain why defensive high-anxious individuals are more likely to re-present for treatment.