The combination of cutting-edge molecular technology and high-throughput phenotyping tools will not bring the expected contribution to the pre-clinical study of anxiety if not paralleled by an increase in our capacity to interpret behavioral data. Here, previous views about the multidimensional nature of emotional behaviors will be expanded and the psychological meaning and behavioral overlaps of widely used anxiety tests such as the open field, elevated plus maze and light-dark box will be discussed. It is proposed here that short-term, intra-individual variations in emotionality, although normally overlooked, constitute an important factor in the study of anxiety and can lead to unreliable estimates of the similarities between tests. The physical integration of different current tests in one single apparatus, in such a way that the emotional status of an animal becomes assessable through a series of distinct tasks, could contribute to increase reliability, rapidity and comprehensiveness in behavioral testing.