During past decades, effects of accessibility to growth have been considered extensively. In previous studies, however, matter of scale has been largely ignored. In this study, relevance of travel accessibility and essential socio-economic variables for explaining population change analysed with a multi-scalar study setting by the case of Finland. The analytical framework of the study relies on applied geographical information systems (GIS). Relationships between population change and explanatory variables were established with non-linear multiple regression, generalised additive models (GAMs). The data consist of population and socio-economic grid cell databases based on authentic records, transport network models, including complete digital road network database. Models were established at six resolutions between 2 km × 2 km and 24 km × 24 km. The main result is that the performance of the models, explaining and predicting population change, strongly relies on potential accessibility particularly at accurate resolutions. An important finding is also that it is possible to establish explain and predict population change tolerably at accurate resolutions and well at coarse resolutions, as the performance of models increases with scale.