Abstract Evidence from analysing and evaluating examples of three voluntary SEAs prepared in Scotland in the mid-late 1990s showed that different spatial and temporal scales were used when providing a baseline environment description. The SEAs analysed were prepared for: a wind farm siting programme that looked at national and short-term impacts; a land use plan that looked at regional and short-term impacts; and a transport plan that examined local and medium-term impacts. It was found that the two SEAs prepared by local government only considered impacts on the baseline environment within their jurisdictional boundaries whilst the SEA prepared by the private business considered impacts on the national baseline. A mixture of baseline data about planning, economic, environmental and social issues were included in the SEAs, however, evidence suggested that each SEA only focussed on those baseline features that might be significantly affected by the proposal. Each SEA also made extensive use of existing baseline information available from a variety of sources including local, and central government records and information from statutory bodies. All of the SEAs acknowledged that baseline data deficiencies existed and in certain cases steps were taken to obtain primary field data to help address these, however, it was also acknowledged that resource restrictions and decision-making deadlines limited the amount of primary baseline data that could be collected.