During typical childhood interactions with screen-media, two features are prominent. First, input is dominated by audio-visual signals and second, these predominately provide children with ready-made images, potentially negating effortful mental imagery construction. We present a two-year longitudinal study on a sample of 109 preschool children. We endeavoured to measure media usage and mental imagery development in a differentiated manner, also taking account of control variables and purpose of media use (learning vs. entertainment). Results indicated that children who viewed more media had worse mental imagery skill. Active media usage (e.g. gaming, tablets) and total screen time linked to lower mental imagery performance. Further, both mental images in the visual and haptic modalities appeared equally affected. Findings are discussed in terms of shaping early educational experiences with respect to virtual and three dimensional reality.