Introduction This study was designed to investigate the links between defence mechanisms and the increase of elderly patient’s depression levels in nursing homes. This research is motivated by the common assumption that the depression level of elderly persons increases in nursing homes (see the AGED study) yet we still have only a few insights about the psychological mechanisms involved. We focused on unconscious defence mechanism models to examine the assumption that the increase in depression levels in nursing homes could reflect the involvement of elderly persons in an adaptive process helping them to deal with the psychological consequences of institutionalisation. Based on the hierarchical and empirical description of defence mechanisms, we postulated that institutionalisation promotes both an increase in depression levels and in immature defence styles. In regard to previous works, we also explored the links between anxiety personality traits, depression levels and defence mechanisms. Method Our sample ( N = 39) was subdivided into two subgroups (group 1: nursing home vs group 2: elderly persons living in communities). Assessments included the French versions of the Geriatric Depression Scale (GDS), the State and Trait Anxiety Inventory (STAI) and the Defence Style Questionnaire (DSQ). Statistical analyses We used the Mann-Whitney formula to compare the two independent groups, and we used the Spearman-Pearson method to perform correlation analyses. Results Two-group comparisons revealed significant increases in depression and anxiety state in nursing homes. Furthermore, elderly persons who are institutionalised exhibit a greater tendency to use regression and help-rejecting complaining defence mechanisms. We performed two distinct correlation analyses matrices with regard to the subjects’ status (nursing-home vs living at home). In both groups, the anxiety-trait is positively correlated with anxiety-state and depression. In elderly person groups living in communities the regression defence mechanism is positively linked with anxiety-trait and depression, while data from aged people groups who are institutionalised reveal a positive link on one hand, between anxiety-trait and regression and on the other hand, between depression and omnipotence defence mechanisms. Discussion Our results confirm that nursing-home placement can constitute a risk-factor for an increase in depression and anxiety state for elderly persons. Institutionalisation favours the use of unconscious immature defences altering social relationships by a tendency to retire within oneself (regression) and by expressing complaints. Interestingly, neither regression nor help-rejecting complaints are significantly correlated with psychiatric outcomes in nursing homes. However, a lower depression level is associated with a higher tendency to overvalue one’s ability to influence one’s life (omnipotence). This negative link could reflect either the adaptive value of omnipotence that restores the sense of control of aged people, or the GDS limits in rating major signs of elderly persons’ depression. In this case, it would be of interest to assess one’s hypomania tendency and to explore links with an increase in use of omnipotence. Based on the correlations reported between anxiety trait, defence mechanisms and psychiatric outcomes, we discussed the interactions between personality and environment in determining behaviour and psychological mechanisms in nursing homes.