This study aims to research that within-the-bar bias depending on direction of axis, data and selective attention. Increasing amount of information, graph is used for conveying pattern information in big data and integrating knowledge effectively. How...
This study aims to research that within-the-bar bias depending on direction of axis, data and selective attention. Increasing amount of information, graph is used for conveying pattern information in big data and integrating knowledge effectively. However, according to past researches, when people understand graph, it causes systematic bias because of visual features such as form, characteristics and so on(Godau et al., 2016). Especially, bar graph which is frequently used in common, length judgement is affected by environmental features(Zacks et al, 1998). This paper conducted experiments that how mean bias is changed depending on data, axis and attention. For this research, using cumulative bar graph that has different achromatic color each parts to draw attention other direction, measured mean bias difference. In addition, limit response time for measuring cognitive bias. Setting up experiments by axis, data and attention as 2 × 2 × 2. As expected results, when attention goes to outer region, bias will be reduced and when attention goes to inner region, bias will be maintained. Thus, plus bar graph shows under estimated mean bias and minus bar graph shows over estimated mean. For the results through experiment, three way interaction is significant and main effect of data is also significant. In hence, according to hypothesis, attention goes on inner region, participants report mean bias as plus data to lower mean and minus data to higher mean than average regardless of direction of axis. When attention goes on outer region, in horizontal bar graph still shows mean bias as inner attention. However, in vertical bar graph, mean bias remarkably reduced and almost disappear. This difference imply that attention automatically assign to lower region according to lower region theory. This proves that visual feature draws attention to inner region and it causes within-the-bar bias.