Page 1 of 8 3MM3_P1ReportPart2_03
Results
In order to determine whether prior visual or non-visual exposure had an effect on the distance walked ina blind walking task, data analysis of the degree of error under the ‘visiual’ and the ‘non-visual’ conditions must be calculated. The effect of time on distance estimation is also of interest. It is hypothesized that prior visual exposure will lead to longer distance estimations. In accordance with previous research, it is anticipated that as the time of the experiment prolongs, the distance estimations of the subjects will increase and subjects will walk longer. It is expected that the later ‘blocks’ of trials will have longer distance estimations than the earlier blocks. To determine relative accuracy across both conditions and distances, error was calculated based on the target distance and the response of the participants.
To provide a description of the results, the means of the subject’s error in distance estimateswere calculated. Subjects who had visual exposure prior to completing the blind walking task had an average distance estimation error of -0.132 meters (SE = 0.253). Subjects were more accurate in their distance estimation with prior visual exposure compared to prior non-visual exposure (M = -0.276 m, SE= .276). Figure 1 is a reflection of the average percent error of distance from the target in the visual and non-visual conditions, based on the target distance. Such a plot allowed for a general comparison of accuracy across distances and conditions. Figure 1 should show that participants with prior visual exposure walked longer in comparison to those who did not have visual exposure before completing the blind walking task.
For analysis of the effect of time on distance walked, the means of the subject’s error in distance estimates were compared based on the blocks during which the trials were walked.Through the first block of trials, subjects averaged an error in distance of 0.043 meters (SE= 0.270). During the second block of trials, participants walked a mean of 0.539 meters less than the target distance, with a standard error of 0.235. Finally, for the third block of trials, -0.117 meters was the average walking error of the subjects (SE= 0.301). Average walking error was the least during block 1 (as expected) and the highest during block 2 (no expected). Based on prior visual or non-visual exposure, average percent error was plotted against trial block in Figure 2. This graph would allow analysis of the impact of the length of the experiment on the distances walked by the subjects. The general trend should indicate that as trial blocks increase from one to three, subjects should walk farther. However, this was not the case.
In order to determine the significance of the experimental results, ANOVA were performed for both condition and trial block. To begin, results obtained for condition were evaluated. Multivariate tests proved to be insignificant and did not produce a main effect (F(31)=.198, p0.05). Second, data collected for block trials were analyzed. Multivariate tests again did not produce a main effect or significant results (F(30)=2.87, p>0.05).
Discussion
The relative importance of visual and locomotor cues in distance estimation was evaluated by comparing the accuracy with which subjects could complete a blind walking task. Subjects performed the task under two separate conditions – one with prior visual exposure, and the other without it. In the blind walking task, subjects were allowed to look at a target for three seconds before having to reach as close as they possibly could to the target while blindfolded. Subjects were blindfolded during the locomotor portion of this task in order to remove the contribution of optic flow information, as it has been found that subjects are able to make reasonably accurate distance estimates based on optic flow alone. The effects of interest are prior exposure (visual or non-visual) and time on the distance travelled. It was hypothesized that subjects with prior visual exposure would walk further than those who were not permitted to use vision before the blind walking task. It was also predicted that as trial blocks and the length of the experiment increased, subjects would walk longer distances than they initially did.
In this experiment, we were unable to replicate the results produced by previous experiments. Many researchers have found that prior visual exposure has a significant effect on the distance walked by participants in the blind walking task. Participants in the prior visual exposure condition should have walked longer distances than those in the prior non-visual exposure condition. This could be due to the fact that only 8 participants were used in our experiment, which is not a large enough N to produce a significant effect. By increasing N, it is possible to eliminate variation, and results obtained would be closer to average of the population. The budget and time constraints for this experiment limited the amount of data that was collected. Participants could only be selected from Dr. Sun’s Psychology class, which only has 15 potential subjects. From these 15, it was not possible for everyone to be a subject, since people were needed from to actually conduct the experiment. Also, having participants coming from the class through which the experiment was being conducted may have also caused an issue. Participants were aware of the past research, the hypotheses and the intent of the experiment. Normally, participants are not aware of the specific intentions of the research, so their efforts are not biased.
The second hypothesis that was not confirmed by our results was that subjects would walk farther as trial blocks increased from one to three. Previous research has shown that as the time of the experiment prolongs, subjects will tend to walk farther towards the end of the experiment than compared to the beginning of the experiment. Trials for each subject were categorized into three blocks, with each block consisting of 4 trials. Had our experimental results been reliable, the results would have shown that for each subject in each condition, as the block number increased from one to three, the distance the subject walked also increased. The data could be skewed because the time it took the experimenters to conduct the trials changed with time. The time required to run a trial decreased as the number of trials conducted increased because the experimenters became more efficient with the experiment. This indicates that the first set (and maybe the second set) of trials that were conducted took longer to perform then the last set of trials. If previous research is valid, subjects from our experiment could have walked longer in the earlier trials because the trials took longer to conduct.
The experiment conducted for this paper took into account similar research questions addressed in recent papers on human distance estimation. The findings here did not support the results found in previous research. This experiment examined only a fewpotential conditions in a very limited number of trials. Additional trials and more funding would allow for an increased number of trials and would certainly help to reproduce the results found by previous studies.
There are several potential sources of error in subjects’ locomotor estimates which could affect visual distance estimates. The first is heading deviation, how far subjects veered from the desired trajectory. Previous research has indicated that heading can have an effect on distance estimates in blind walking experiments. Most subjects tactilely directed themselves toward the desired visual target by positioning their feet directly toward the target. This was accomplished during the three seconds the subject was allowed to ‘learn the target’. However, subjects that did not position their feet parallel to the target during this period tended to veer away from the desired trajectory. Another variable specific to the locomotive conditions is walking speed, since subjects may have based their estimates on a constant speed of travel while in reality their speed varied as they walked.
In conclusion, the results of this experiment were not able to confirm the two hypotheses which motivated this study. The results lead us to believe that prior visual exposure does not have a significant effect on the distance walked during the blind walking task. The outcome also indicates that time does not influence the distance walked by participants. At this point, reflection on this experience can benefit the experimenters in many ways. For most of the experimenters, these were the first trials, the first experiment and the first research that most of them have ever conducted. Being amateur scientists, it is almost expected that mistakes will be made, since running an experiment is not a simple thing to do. Experimenter’s from Dr. Sun’s class should not be discouraged by the results of this experiment, but should instead use this experience as a guide for their future research. They should learn that conducting an experiment requires a lot of planning, labor and precise execution. If these experimenters learn from their mistakes and improving upon all the errors they made, they can ensure themselves a bright future in psychology research.
Figure Captions
Figure 1
A plot of the average percent error of each subject from each group against the target distance estimation in meters. The square, blue data points and dashed line represent subjects with prior visual exposure, where as the triangular, pink points connected by a solid line symbolize the non visual condition. This graph allows comparison of the visual and non visual conditions based the average distance estimation ability of the subjects.
Figure 2
A plot of the average percent error against the trial block 1, 2 or 3. Again, the blue data points represent subjects that were visually exposed before completing the blind walking task, and the pink points are for data from the subjects who were not visually exposed before the blind walking task. Figure 2 enables analysis of the trend of error based on the trial, or the time of the experiment.