We describe an experiment that was designed to explore differences between active and passive travelers in a way-finding task. In this study, we examined the effect of active travel mode on spatial and visual memory for a built environment. After completing a way finding task in a university campus, we tested participant's memory for the test route using sketch map, mirror-image discrimination, and scene recognition tests. In addition, we tracked participant's eye movements during the scene recognition test. Results were consistent with the hypothesis that active travelers had enhanced memory for the built environment. Our data also provide some evidence for qualitative differences between active and passive travelers in the visual cues they used to recognize scenes. Based on our findings, we suggest that travel mode is an important consideration when designing built environments.
Way-finding involves interaction between the traveler and the environment, and a
variety of cognitive processes are engaged by this task. For example, successful
navigation requires the generation of accurate cognitive maps which requires memorizing,
recognizing, and decoding spatial information and location attributes, as well as
forming an action chain of spatial knowledge [
Despite the focus on the process of cognitive mapping in previous navigation
research [
We hypothesized that active travelers would have more accurate memory for spatial layout of study routes than passive travelers (i.e. cognitive maps), but they would also have better memory for scenes encountered during the way-finding process. 2
2.1
One-hundred-and-eight participants (52 Females) were randomly assigned to active or passive traveler groups (54 participants in each group). Participants’ ages ranged from 17 years to 58 years (M=23.9; SD = 7.07).
2.2
Participants first completed the way-finding task. They were given a map showing the predetermined route and three ‘key landmarks’ to answer the questions relating to each landmark. Passive participants were asked to follow the experimenter through the test route, whereas the active groups led the experimenter through the test route. Both groups were instructed to pay attention to their surroundings.
After the test route had been completed, participants filled a Spatial Ability Questionnaire and Sketch Map test. In the questionnaire, participants made subjective judgments of their own spatial ability, their level of familiarity with the campus, and spatial layout of the campus. In the sketch map test, participants were given a partially completed map of campus containing only start point of the test route, and were instructed to draw the test route and locate the key landmarks as accurately as possible.
Next, participants were presented with scenes containing main buildings and landmarks on and off the study route for completing the Mirror-image Discrimination test (MD), and Scene Recognition test (SR). MD was designed to test the memory for orientation of scenes – participants were shown images of scenes and mirror-reversed versions of these side-by-side on a computer screen. They had to select the correct orientation of the images. We predicted that orientation of scenes would be encoded in memory more accurately by participants in the active traveler group.
Finally, for the scene recognition test, participants were shown fifty-four images (half on-route, half off-route). They had to indicate whether the scene was encountered on the route or not. In addition to measuring the accuracy, we also recorded their eye-movements whilst they completed this task by a static eye-tracker (Tobii TX300). 3 3.1
Recognition memory performance. Active travelers were more accurate in this test (Mean correct = 85.5; SD =11.1) than passive travelers (M = 81.7; SD =15.2). An independent sample t-test confirmed that this difference was statistically significant, t (96) = 2.00, p < 0.05.
Eye tracking analysis. We analysed eye movements by defining Areas of Interest
(AOIs) for each of the images in the Scene Recognition Test (see Figure 1). The two
dependent variables were total fixation count per AOI, and mean gaze duration per
AOI (Table 1). Total fixation count measures the number of fixations to an AOI; and
mean gaze duration is the length of time for one visit in an AOI from entry to exit [
We were only interested in interaction between Travel Mode and AOIs. This interaction was non-significant for Fixation Count F2, 212 = 1.66, p > 0.05, η2 = .01. However, for Gaze Duration the interaction was significant, F2, 212 = 3.22, p < 0.05, η2 = .03. This interaction indicates that travel mode affects visual cues encoded in memory by participants during way-finding.
To investigate the interaction effect in gaze duration data, we carried out planned comparison t-tests between active and passive groups for each AOI. This test revealed greater reliance on First Floors among passive travelers compared with active travellers, t (106) = -2.71, p < 0.05 (not in upper floors, t (106) = 0.34, p > 0.05, and nonbuilding, t (106) = 0.12, p > 0.05).
Fig. 1. Examples of heatmaps and defined AOIs from aggregate data of both groups of active and passive travelers; Blue: First Floor, Orange: Upper Floors, Green: Non-buildings (Left image shows heat maps; Right image shows the AOIs).
We analyzed accuracy data on the Mirror Discrimination Test with 2x3 repeated measure ANOVA with factors Travel Mode (active/ passive) and Image Type (onroute campus, off-route campus, off campus). There was a significant main effect of travel mode, F1, 103 = 4.15, p < 0.05, η2 = .040, and of image type, F1.9,198 = 42.63, p < 0.05, η2 = .040. However, the interaction between factors was not statistically significant, F1.9, 198 = 0.48, p > 0.05, η2 = .005. This result shows that the ability to correctly Active
Passive 3.3
On Route Campus 86.5 (11.3) 81.4 (12.7) identify the orientation of visual scenes was enhanced for the active travelers, but that this benefit was not specific to images from the test route. It is not possible based on the data from this study to determine whether this effect was caused by the experimental manipulation used in this study, but we intend to follow this up in future research.
Sketch maps were geo-referenced and analyzed separately in ArcMap 10. The sketch maps were analyzed according to six factors, namely: route dislocation, landmark and building dislocation, open space dislocation, number of landmarks, number of spaces, and number of details. An independent t-test reveals that there was a statistically significant difference between groups in open space dislocation, with passive travelers were scoring higher than active travelers, -41.6 (95% CI, 80 to 3), t (91) = 2.1, p < 0.05. All other comparisons were non-significant. Table 3 presents the mean and standard deviation for each six items.
SAQ scales were derived from Santa Barbara Sense of Direction Scale [
Our results suggest that active travelers had better memory for the built
environment than passive travellers. This memory benefit manifested in two ways. First, in
line with previous studies [
Our data also provide evidence that memory for built environments differs qualitatively as a function of travel mode. Specifically, eye tracking analysis showed that passive travelers looked more at first floor of buildings, perhaps suggesting that they were more concerned with visual details immediately in front of them during the wayfinding task. In future research it will be important to reveal which characteristics of the visual environment are most beneficial to the process of way-finding. Acknowledgments. We thank Associate Professor Samsung Lim for providing the UNSW campus Shapefile used in the experiment for analyzing the sketch maps.