In the current study, 349 teachers of all grades from 30 primary schools with rolls of over 200 pupils in a large Regional Authority were surveyed. In addition to a specially prepared questionnaire, measures of anxiety and depression, personality, and work environment were used. As well as employing a cross-sectional experimental design, a longitudinal component was introduced by carrying out two further surveys of anxiety and depression levels during the school year. Included in the third survey was a dependency grid designed to investigate the social support networks available to class teachers. Finally, semi-structured interviews were carried out in four schools, two in which a high proportion of teachers were reporting high levels of anxiety and two in which the overall level of anxiety was low. The principal finding was that over one third of teachers were reporting 'clinical' levels of anxiety whereas only one twentieth of the sample were returning 'clinical' depression scores. Univariate analysis indicated that, in the case of class teachers, stressors that correlated most highly with anxiety were organizational in the form of work load, time pressures, and role conflict. However, there were indications that supportive promoted staff were associated with lower levels of anxiety amongst class teachers and this finding was further substantiated by the results of the dependency grid and also a discriminant function analysis. Semi-structured interviews highlighted the differences in the social climates of high and low anxiety schools. Although individual characteristics such as personality factors in the form of neuroticism and, to a lesser extent, locus of control were associated with anxiety levels the primary finding was that the characteristics of a school as a social institution were powerful predictors of the anxiety levels being reported by class teachers and schools. Moreover, the promoted staff appeared to play a key role in the nature of a school's social climate.