The cross-sectional Biomarkers Study was undertaken in Canberra, Australia (2000-2002) to examine the role of psychosocial factors in the socioeconomic health gradient, via physiological changes consequent upon activation of the neuroendocrine stress response.¶ The study population was derived from healthy 40-44 year old men and women already participating in a longitudinal cohort study. Using data from the cohort study, four groups with similar occupational status were formed. The study sample was randomly selected within these groups, thus representing the socioeconomic spectrum.¶ A pilot study involved 60 participants with blood and saliva samples measured on two occasions. A further 302 people had blood and saliva samples taken on one occasion. Socioeconomic status was measured by occupational code and status, personal and household income, education and perceived position in the community and in Australia. Psychosocial and behavioural factors, including job strain, job security, coping style, anxiety, depression, optimism, self-esteem, sense of belonging and trust, social support, smoking, exercise and alcohol intake were assessed by selfreport. Five biological parameters: plasma fibrinogen, glycated haemoglobin, waisthip ratio, serum neopterin and salivary IgA were measured as outcome variables.Three hypotheses were tested:¶ 1. There is a socioeconomic gradient in measures of psychosocial stress, and of psychological resilience.¶ 2. There is a socioeconomic gradient in biological measures that have a plausible¶ association with future disease. 3. Psychosocial factors mediate the demonstrated association between socioeconomic status and the biological measures.¶ Data analysis confirmed a socioeconomic gradient in some psychosocial and behavioural variables: economic strain (r=-0.44, p&lt0.001), job demands (r=0.45, p&lt0.001), job control (r=0.26, p&lt0.001), active coping style (r=0.28, p &lt0.001), sense of optimism (r=0.24, p&lt0.001), social capital (r=0.26, p&lt0.001), job security (r=0.17, p=0.002), job marketability (r=-0.16, p=0.005), sense of belonging (r=0.22, p&lt0.001), number of adverse life events (r=-0.13, p=0.01) and positive interaction with family and friends (r=0.20, p&lt0.001 ), vigorous physical activity (r=-0.16, p=0.002), alcohol consumption (r=0.30, p&lt0.001) and smoking status (r=-0.25, p&lt0.001). There was no socioeconomic gradient in anxiety, depression, neuroticism, hostility, locus of control, self-esteem, perceived stress or mental health (SF-12). Four of the five biological markers varied with socioeconomic status: plasma fibrinogen (female (F): r=-0.26, p=0.002, male (M) r=-0.08, p=0.30), glycated haemoglobin (F: r=-0.23, p=0.01, M: r=-0.11, p=0.17), waist-hip ratio (F: r=-0.19, p=0.03, M: r=-0.27, p&lt0.001), serum neopterin (F: r=-0.21, p=0.009, M: r=-0.04, p=0.56), salivary IgA (F: r=-0.07, p=0.38, M: r=0.004, p=0.97). A more adverse biological profile was associated with lower socioeconomic status. Work characteristics, coping style, smoking and exercise were particularly important mediators of the association between the biological markers and socioeconomic status. Particular psychosocial factors were consistent mediators of the association between specific biomarkers and socioeconomic status (with little variation for different measures of socioeconomic status). However, the particular psychosocial factors providing significant mediation varied for the different markers.¶ In this sample of healthy 40-44 year olds, four out of five biological markers showed moderate socioeconomic variation with a more favourable profile associated with higher SES. The data provide limited support for the importance of psychosocial factors in the socioeconomic health gradient.