Socioeconomic inequalities in morbidity and mortality exist, but the psychobiological pathways that link social status and health are less clear. It has previously been reported that socioeconomic status is inversely associated with the magnitude of the cortisol response to awakening (CAR) in men and women of working age. In the present study, we tested whether larger cortisol responses would be present in an older retired population, and whether the CAR differed between men and women. The extent to which adherence to saliva sample timing also affects the CAR was investigated. Ninety three men and women aged 65-80 years took saliva samples on waking, and then 10, 20, 30 and 60 min after waking. Subjective social status was assessed using the 'ladder' measure devised by Adler et al. (2000). Non-compliance was defined as a reported delay of 10 min or more between waking and taking the first saliva sample. Cortisol levels on waking were significantly higher in the non-compliant individuals, and the CAR was blunted compared with that of compliant participants. With non-compliant participants eliminated from the analyses, we found that low social status was associated with a larger CAR after adjusting for gender, waist/hip ratio, body mass index, smoking, time of waking, chronic illness, prescription medication, education and financial strain. No association was found between CAR and education or financial strain. Women also had significantly larger CARs, independent of socioeconomic position. The results highlight the importance of controlling for non-compliance and are consistent with the notion that higher socioeconomic position protects against stress-related activation of psychobiological pathways which may contribute to variation in disease risk evident in old age.