Taking the right (or left?) turn: effects of landmark salience on the retrieval of route directions

ABSTRACT

We test the effect of a landmark’s visual and structural salience on memory retrieval of turning directions at intersections in a virtual environment. Across three studies, we find an increased decision correctness when the location of the visually salient landmark converges with the turning direction. This pattern is robust across various perspectives, including the return path. We find no performance differences between route repetition and return path. We find some evidence to turn toward a visually salient landmark even if this is not correct. Taken together, a visually salient landmark helps if located in the turning direction, but might be detrimental to a navigator’s ability to recall the correct direction if located on the opposite side.

KEYWORDS:

• Landmark information
• route memory
• route recall
• visual salience
• structural salience

Acknowledgments

We thank Benedikt Solf for implementing the design and collecting the data in all three studies.

Disclosure statement

The author(s) report no conflict of interests.

Data availability statement

Data are available upon request from the first author.

Notes

¹ We did not consider semantic (or cognitive) salience in our research, because we found no satisfying solution to operationalize this aspect to be equally salient and consistent across all participants, as well as showing no cross-interferences with the other two salience dimensions.

² Please note that Table 1 also provides information for Study 2 and Study 3 to allow an easier comparison across the studies.

³ A technical problem with the software code resulted in the first ten participants performing 31 rather 32 blocks. Furthermore, participants received wrong feedback about the correctness of their response in about 7% of trials.

⁴ Individual differences were not the primary focus of our research question. A preliminary inspection of these measures revealed no major relation with the overall correctness in the main experiment. We thus decided to refrain from further analysis of these data. We accounted for individual differences by including the participant id as a random intercept..