As a result of an increasing interest in environmental issues, many scales have been developed to measure environmental attitudes. However, no agreement exists in relation to their dimensionality. They have been traditionally viewed as a uni-dimensional construct ranging from unconcerned at the low end, to concerned at the high end. Conversely, others argue that they have three correlated factors, these are concern for the self (Egoistic), other people (Altruistic), and the biosphere (Biospheric). Using combined factor analyses of items from previous environmental attitudes scales, and further factor analysis with the first-order factors that emerged on the first analysis, Wiseman and Bogner (2003) found two uncorrelated higher-order factors. They presented this as a Model of Ecological Values (MEV). This model has two orthogonal factors, Preservation and Utilisation. The former reflects conservation and protection of the environment, and the latter the utilisation of natural resources. This paper addressed two questions: (1) Do environmental attitudes form a two second-order structure as proposed by Wiseman and Bogner (2003)? If so, (2) Do these two factors differentially predict self-reported ecological behaviour and economic liberalism? A questionnaire-based study was conducted to test this higher-order structure, incorporating other environmental attitude measures into the model. Subjects were 455 undergraduate students (319 females; 136 males), with age ranging from 17 to 48 years (M = .19; SD = 4.31). Support for the higher-order model was accessed by factor analytic techniques through two procedures. First, 99 items from established measures of environmental attitudes were forced to a two-factor solution. Second, first-order factors were established, and then another factor analysis was performed to found the second-order factors. Goodness-of-fit indices from confirmatory factor analyses (using LISREL 8.54) indicated that a correlated two-factor solution, consisting of an Utilisation factor and a Preservation factor, provides the best fit to the data (_2 = 809.73; df = 393; _2/df= 2.06; RMSEA = .048; SRMR = .052; GFI = .89; CFI = .97). Additional results demonstrated that, after controlling for demographic variables, the Preservation factor predicted self-reported ecological behaviour (_ = .53, P