Void fraction is one of the most important parameters to analyze various properties of multi-phase flows. Bubbles can arrange themselves in different manners and their distribution across the flow cross-section changes with time. Projection data (collected from a computerized tomographic scanner) for such a cross-section is not instantaneous in nature so time-averaging (over measurement interval) is required to obtain phase distributions. Two different types of averaging schemes are discussed in this work and it is shown that inappropriate averaging results in a significant dynamic bias effect (DB) leading to erroneous images. This analysis is performed on a three-phase bubble column reactor in which air, water and poly vinyl chloride (PVC) are used as representatives of gas, liquid and solid phases. Measurements have been performed for two different levels of this column. First Kanpur Theorem (KT-1) is implemented to select "good" projection data which is then used in the tomographic reconstruction step. Characterization of reconstructed cross-sections is done by Second Kanpur Theorem (KT-2). This approach provides a comprehensive strategy to compare quantitatively cross sectional void-fraction patterns obtained for different measurement levels. We observe that DB error is approximately 3 times more when air velocity is increased from 0.06 m/s to 0.14 m/s. (C) 2013 Elsevier Ltd. All rights reserved.