In the last decade, the use of eye tracking in visual marketing and consumer choice studies has rapidly increased (see Wedel and Pieters 2008 for a review). The majority of these consumer studies are performed in laboratories with products presented on a monitor (Wedel and Pieters, 2008) or as projections (Tonkin, Ouzts and Duchowski, 2011). The findings in these settings are often generalized to more natural environments but before this can be done, there are several issues to be considered. The central bias is one of them.
It has been shown that when presented with a naturalistic scene on a computer monitor, viewers tend to look in the centre of the screen. This central bias occurs irrespectively of both the task and the distribution of features in the scenes (Tatler, 2007). The central bias has mostly been investigated in scene viewing and visual search studies. The question that we want to address in this study is whether these findings are restricted to only the computer screen or if they can be translated to naturalistic settings with stimuli presented within a frame such as a product shelf in the supermarket? We also address the temporal aspect of this bias and whether it is more pronounced in early stages of the process.
We investigated the central bias in a naturalistic environment by studying the visual behaviour of consumers making decisions in a supermarket. 20 consumers were recruited in a supermarket and were fitted with the SMI Eye Tracking Glasses, recording binocularly at 30Hz. Participants were asked to select a fruit syrup that they would consider buying. This was done during their regular shopping. Both the original and a reorganized setup of the shelf was used. The reorganization was made so that the products in the middle of the shelf was moved as far out as possible. This was done to ensure that the distribution of fixations in the scene did not depend on the setup of the shelf and the placement of the most popular products.
The results did not show a clear central bias tendency of the consumers visual attention to the product shelf. Interestingly, the placement of the products had only minor effect on consumers visual attention to them. These results are an interesting contrast to earlier monitor studies clearly showing a central fixation bias. The question is how these results could be generalized to naturalistic settings and in what kind of situations this bias would occur. The results from this study show that the central bias tendency is not that robust in naturalistic setting. Static pictures have several shortcomings compared to real world environments. In real supermarkets consumers can move around and may place themselves so that the most interesting features of the environment are placed at the centre of their visual field. These environments also force people to move around since it is often difficult to look at the whole shelf and examine all products from a single viewpoint. These results are important for the interpretation of data from visual search and scene perception studies in general and consumer decision making studies performed on computer monitors specifically.