Abstract
BACKGROUND:
Previous researchers examined the effects of either drivers’ personal characteristics or traffic sign design features on the usability of traffic signs. Their research indicated a connection existed between personal characteristics and usability and between design features and usability.
OBJECTIVES:
The focus of this study was to investigate which personal characteristics of drivers and which features of traffic sign design affect traffic sign usability the most.
METHODS:
The study was conducted in three stages. In the first stage, the participants filled out a questionnaire designed to record each driver’s personal characteristics. In the second stage, a System Usability Scale (SUS) was used to evaluate the subjective usability of traffic signs. The SUS had 10 statements that participants (N = 386) scored on a 5-point Likert-type scale from strongly agree to strongly disagree. In the third stage, these participants assigned from 0 to 100 points to 20 signs based on participant perception of the five design features of familiarity, concreteness, simplicity, meaningfulness, and semantic distance.
RESULTS:
The results showed that four of the five personal characteristics studied (age, education level, possession of a driving license, and formal driving experience) correlated significantly with traffic sign usability. The exception was gender, which did not correlate significantly. Additionally, it was found that the five traffic sign design features correlated in varying degrees of significance with each other and with specific traffic signs.
CONCLUSIONS:
Traffic sign usability depends mostly both on driver education level and age group and on the design feature of meaningfulness. These findings have implications for how drivers should be trained and how signs should be designed.
Introduction
Symbolic signs are important because they are often understood faster than text phrases and, when standardized, encourage comprehension beyond regional borders. Traffic signs that are symbolic hold much importance for road safety. These signs are essential tools for transmitting traffic messages and other information to motorists and cyclists who need to receive that information at a glance. These signs are also useful in transferring information to all road users who do not comprehend the text phrases. Whereas text on traffic signs is comprehensible to the local population, it has been shown to impair comprehension in foreigners [1]. Since traffic signs warn, regulate, and guide road users, understanding these signs is crucial for promoting driver compliance and preventing crashes [2]. Indeed, substantial evidence exists for the perils faced by foreign drivers unfamiliar with road signs and traffic regulations outside their countries (e.g. [3]).
Numerous studies have shown that such factors as design elements (color, shape, text addition to signs, and font style and size), legibility distance, and standardization influence guessability and usability or comprehension of traffic signs. The guessability of pictorial and abstract symbolic signs relates directly to their comprehensibility and determines the extent to which drivers respond to signs quickly and correctly [4]. One of the most crucial features of symbolic signs is their broad usability. Studying a product’s usability reveals how easily and/or quickly a user can utilize the product [5]. According to the International Organization for Standardization [6], usability is the “extent to which a system, product or service can be used by specified users to achieve specified goals with effectiveness, efficiency, and satisfaction in a specified context of use.” A product is usable if it allows users to learn, guess, re-use, and reach a steady skilled level of performance while spending minimal cost [7]. Traffic signs are therefore usable when road users understand the sign’s intended message and respond with speed and accuracy. The symbols used to design traffic signs must be selected to ensure that the user can recall the sign’s message after periods of limited or nonexistent interaction with the sign. If safety icons, such as the symbols on traffic signs, are not easily understood, they lose their communication value and may cause injury or loss of life. Their usability, therefore, is questionable at best. Design elements, such as color, stylization, and level of detail, affect icon usability, but to varying degrees [8]. Design elements are the cognitive features that are widely known to the general public.
In a study conducted by Mcdougall et al. [9], five essential features (familiarity, concreteness, complexity, meaningfulness, and semantic distance) of 239 symbols, including a few symbols pertinent to traffic signs, were measured and related norms were provided. Familiarity is the frequency with which signs are encountered. Concreteness means signs represent real objects or people. Complexity (or its counterpart, simplicity) indicates the amount of integrated details on a sign. Meaningfulness refers to how well users understand a sign’s definition. Semantic distance is the closeness of the sign’s graphical representation to the sign’s actual function and/or message. Semantic distance is not necessarily synonymous with meaningfulness. For example, some traffic signs may have meaningfulness to drivers, despite indicating high semantic distance, due to familiarity with the signs even though the drivers cannot interpret the signs based solely on their graphical representation [10]. The sign design features of familiarity, concreteness, simplicity, and meaningfulness were considered by Ng and Chan [11]. According to their findings, comprehension of traffic signs by licensed drivers correlates significantly with sign familiarity. Their study also assessed comprehension according to three age groups, two levels of education, and numerous gradations of driving experience. Some of their findings demonstrated that understanding symbolic signs depends on these personal characteristics, but, although significant, the differences between high and low comprehension were never more than a few percentage points. For example, they reported a significant relationship between the education level of participants and their comprehension of traffic signs: participants with a university education scored 6% higher in comprehending traffic signs than those without a university education did. Al-Madani and Al-Janahi [12] found additional personal characteristics significantly affecting comprehension of traffic signs: age, gender, education, driver experience, marital status, income, and nationality.
The findings of studies conducted for assessing the usability of traffic signs in rural and urban areas of Iran suggest that the majority of Iran’s traffic signs are marginally usable for the Iranian population and that traffic sign design should take into consideration ergonomic principles to increase usability and thereby reduce roadway crashes [13, 14]. The next step in the evolution of road safety is to identify the factors that interfere with traffic sign effectiveness. To that end, this study investigated which personal characteristics of drivers and traffic sign design features affect the usability of traffic signs the most.
Material and methods
Subjects
A self-reported survey was to be used to collect data on personal factors, sign usability scores, and ratings of sign design features. The intention was to sample participants having a variety of personal characteristics (age, gender, education level, driving license, and amount of driving experience), according to the number of people in the target population, which was a public medical university (N = 11,531). The cluster sampling method was employed to achieve this. Based on Cochran’s formula and on a 95% confidence level, the number of samples required for this study was determined to be 372. Based on proportional allocation, 43 samples from headquarters, 44 samples from the faculty, 59 samples from the health centers, 146 samples from the teaching hospitals, and 80 samples from the non-teaching hospitals were to be selected. Considering the usual loss of samples, 400 samples were sought. At the end, 386 questionnaires were completed, included male (N = 181) and female (N = 205) participants with a mean age of 39.21 (SD = 9.78, range = 21–60). Among participants having a car driving license, 35.9% of them had 1 to 5 years of driving experience. All participants had normal or corrected to normal vision, and no participant suffered from color vision deficiencies. The research was approved by the local Ethics Committee. Participants provided written informed consent.
Measurement procedures
In accordance with our previous studies [13, 14], the same 20 traffic signs were selected for evaluation in this study. Images of the traffic signs were displayed in color to fit within a maximum shape of 2×2 cm on white paper having no boundary. Table 1 presents the sign images with their definitions. The choice of signs represented the three sign categories of warning, regulatory, and informative.
Traffic signs used in the study
Traffic signs used in the study
Participants were briefed about the study’s purpose, the study’s procedure, definitions of the design feature terms, and the rating instructions. An interviewer was always present during the survey to answer participant questions and ensure the reliability of results. Traffic signs were presented to participants one by one, and each participant was separately evaluated. If a participant was unsure of a sign’s exact meaning, the definition was conveyed to him/her before the subjective ratings began. The required information was recorded on a paper questionnaire for all three stages of the study.
In the first stage, five personal characteristics of participants were gathered: age, gender, education level, possession of a car driving license, and amount of driving experience in years.
In the second stage, the Persian version of the Modified System Usability Scale (SUS) was used to measure the subjective usability of the traffic signs [14]. The scale consisted of 10 statements, shown in Table 2, to be scored on a 5-point Likert-type scale (strongly agree = 5 to strongly disagree = 1). For scoring the SUS, one score position is subtracted from each odd numbered statement (1, 3, 5, 7, 9) and the score position of each even numbered statement (2, 4, 6, 8, 10) is subtracted from five. The final score (0 to 100) is calculated by multiplying the aggregate total of the odd and even numbered statements by 2.5. Products with a usability score less than 50 are unacceptable, whereas a score of 50 to 70 is marginally acceptable, and a score higher than 70 is consideed acceptable [14, 15].
The modified SUS used in the study
In the third stage, participants, using the Persian-version of a cognitive features questionnaire [16], rated the sign design features for each traffic sign on a 100 point scale for familiarity (from 0: very unfamiliar to 100: very familiar), concreteness (from 0: quite abstract to 100: quite concrete), simplicity (from 0: very complex to 100: very simple), meaningfulness (from 0: completely meaningless to 100: completely meaningful), and semantic distance (from 0: no semantic proximity to 100: lots of semantic proximity). To ensure the results of the subjective ratings of traffic signs design features were as reliable as possible, the interviewer briefly explained the terms before the ratings began.
Table 3 presents statistics on participants and SUS score averages for the studied signs. A Spearman correlation analysis showed that age was weakly correlated with SUS score (r = 0.106, p = 0.01). No significant relationship was found between gender and SUS score (p > 0.05). An independent t-test determined that having a driving license and driving experience significantly influenced SUS score (p < 0.05). An analysis of variance (ANOVA) revealed that the SUS score depended on the education level of respondents (p < 0.05). A Tukey test suggested that level of education of participants from no diploma to a bachelor’s degree had significant effects on SUS score (p < 0.05), but there were no significant differences in SUS score among participants with a bachelor’s’ degree or higher degree (p > 0.05).
Mean and standard deviation (SD) of SUS scores according to participant characteristics
Mean and standard deviation (SD) of SUS scores according to participant characteristics
*No driving license means non-drivers.
Descriptive statistics on the ratings of sign design features for the 20 traffic signs appear in Table 4. According to the table, the study’s signs were perceived to be moderately familiar, concrete, simple, and meaningful. Signs with minimum and maximum scores for each design feature are displayed in Table 5. The sign for “No entry for vehicles more than 2 meters wide” had the highest ratings for familiarity, meaningfulness, and semantic distance. “No vehicles” was rated the most concrete sign and “Hotel” was rated the simplest sign. “Stop sign ahead” was assessed as an outlier with the lowest ratings on simplicity and meaningfulness. The signs for “Guarded railway crossing in 300 meters”, “Give away ahead”, and “Tram crossing ahead” were also deemed outliers with much lower ratings for familiarity, concreteness, and semantic distance, respectively.
Descriptive statistics of sign feature ratings for traffic signs
Interrelationships among the features were examined by Pearson correlation analysis, as shown in Table 6. Meaningfulness, semantic distance, and concreteness were significantly correlated, and familiarity and simplicity were not. The highest and lowest significant correlations were meaningfulness with semantic distance (r = 0.925, p < 0.001) and concreteness with simplicity (r = 0.350, p < 0.001), respectively.
Signs with maximum and minimum ratings of sign design features
Pearson’s correlation coefficients for sign design features and usability score
*Correlation is significant at the 0.001 level (2-tailed).
A Pearson correlation test was then conducted to verify the usability of signs with consideration given to familiarity, concreteness, simplicity, meaningfulness, and semantic distance. Results demonstrated that all the mentioned features were significantly associated with the SUS scores. The highest correlation with usability scores was for meaningfulness (r = 0.952, p < 0.001), while the lowest correlation was for simplicity (r = 0.366, p < 0.001).
Because comprehension and guessability are essential to usability, comparison of results from sign comprehension and guessability studies to sign usability studies is fruitful even though their methodologies differ. Comprehension studies present sign images to participants and determine the correctness of their responses based on either open-ended questions, multiple choice, true/false statements, simulated driving, or other test protocols. Guessability studies ask participants to predict a sign’s meaning and then rate the sign according to several ergonomic principles, i.e., cognitive sign design features. Usability studies usually inform participants of the sign meanings and then ask them to rate the signs according to several ergonomic principles. All three types of studies often analyze results too according to participant demographics.
This study was performed to examine the hypotheses that the usability of traffic signs relates to sign design features and to drivers’ personal characteristics. There are valuable findings from other studies concerning comprehension and guessability of symbolic signs, but no other studies combined testing both our hypotheses specifically for traffic sign usability with all the sign features and all the personal characteristics that this study examined. For example, Ng and Chan [17] tested 41 participants aged 18–27 on the guessability of 120 selected traffic signs and demonstrated that they were guessable as a function of their design features. Ng and Chan’s next study [11] tested comprehension based on traffic sign design features and three types of drivers’ characteristics for 109 drivers in Hong Kong. Their results varied substantially from those we obtained for a different geographical region and culture (Iran). Another study validated SUS as a reliable tool for evaluating the usability and comprehension of safety signs, and so, SUS may be considered also appropriate for assessing other types of graphical symbols [18].
Concerning personal characteristics, our study found a weak positive correlation between age and SUS score. However, our findings do not support previous research. Ng et al. [18] indicated that age group did not significantly influence SUS score. Our study had four age groups (21–30, 31–40, 41–50, and 51–60). The average usability score increased with age, except for the oldest group. The 51–60 age group had a lower score than the 41–50 age group. For the 21–30, 31–40 and 41–50 age groups, it may be said that the probability of encountering signs increases with age. Senior drivers are likely to find traffic signs that have been on the road for a long time more usable than junior drivers do. However, about the age group of 51–60, our results could theoretically compare to the results of the study by Schulz et al. [19], which examined the effect of age on traffic sign comprehension of older (> 65 years) and younger (15–35 years) participants. The older drivers had a lower processing speed in comprehending traffic signs than the younger drivers did. Additionally, among older drivers, traffic sign comprehension processing speed was relevant to semantic memory, which is a related factor for comprehension of general symbols by older adults. The se different results regarding age among the various studies may be explained by differences in study methodology, signs and symbols used, participant population samples, and study limitations –to name a few.
In this study, the effects of the five personal factors of age, gender, education level, driver’s license, and driving experience were evaluated to determine traffic sign usability. A Spearman correlation test indicated that, with increasing driver age, signs were identified as more usable, although in the study by Schulz et al. [19], regardless of age, both cognitive skills and regular updating of traffic sign knowledge were relevant to traffic sign comprehension.
Consistent with some previous findings [20, 21] on safety sign usability and comprehension, respectively, our study found gender had no significant effect on traffic sign usability scores. However, in the study by Taamneh and Alkheder [22], which evaluated traffic sign comprehension among Jordanian drivers, no significant differences were found between males and females for warning and regulatory signs, but a significant difference existed for guide signs. Contrary to this, Al-Madani and Al-Janahi [12] found that gender played a major role in traffic sign comprehension, with females averaging lower comprehension than males for warning and regulatory signs. In another study [23] on traffic sign comprehension, male drivers interpreted all seven tested signs better than female drivers did. Liu et al. [24] examined the guessability of traffic signs in all three categories and found no significant difference between the genders. They speculate that regional culture may explain why results vary. They also point out that not all researchers analyzed the results similarly, that is, some controlled for age and education level while others did not.
Our results indicated a significant relationship between the usability score of traffic signs and participant education level. Participants had five levels of education (<Diploma, Diploma, Bachelor, Master, and PhD). The mean score of sign usability rose as level of education increased. Participants lacking a secondary school diploma had the lowest scores and PhD graduates had the highest scores. Ng and Chan [11] reported a similar result for their two age groups: below university vs. university and above. Those with more education may have had experiences inducive to quick learning and memorization, and may also possess superior abilities for information processing [25].
Regarding driver license, the results indicated that participants who carried a driver license found the signs more usable than those without driver licenses. Similarly, in the study by Duarte et al. [21], possessing a driving license significantly affected the comprehension of safety signs. Since understanding traffic signs is a main requirement for obtaining a driving license, sign definitions are taught in driving schools. To be licensed, students must pass a driving test that includes a section on signs. Consequently, license holders are more familiar with traffic signs than non-licensed participants are [23]. This, therefore, leads licensed drivers to find traffic signs more usable.
Based on our findings, having driving experience significantly correlated with the usability score of traffic signs. The same results were found by Chan and Ng [20, 26], whose studies demonstrated that participants with a driving background had a higher guessability score for safety signs. They speculate that knowledge of traffic signs is transferable to safety signs for several reasons. Both types of signs have the same categories: warning, regulatory (prohibition and mandatory), and guidance. The categories of both have the same shapes: warning (triangle), regulatory (roundel), and guide (rectangle). Both often have the same colors, which is particularly true for regulatory signs: prohibition (red border and usually red slash) and mandatory (blue background). Both have the same general meanings, i.e., warning signs are alerts to hazards, prohibition signs forbid certain actions, mandatory signs convey orders to take certain actions, and guide signs provide directions to destinations. Drivers are also familiar with the concept of a sign system, one that must be learned, understood, and obeyed. The explanation above by Chan and Ng [26], we believe, validates the converse: our comparisons of findings from usability research on safety signs to findings from usability research on traffic signs.
The next part of this study rated sign designs based on familiarity, concreteness, simplicity, meaningfulness, and semantic distance. The score for each feature varied from sign to sign, but, with the exception of familiarity and simplicity, a significant correlation existed among the features. Similar studies have revealed a significant correlation respecting the design features of traffic signs. The findings of a study investigating the effects of driver characteristics and sign design features vs. the comprehensibility of traffic signs showed that the highest correlation was between meaningfulness and concreteness [11]. The later findings of Chan and Ng [20] also suggested a significant correlation among the design features of safety signs. The highest correlation was between semantic distance and meaningfulness (r = 0.947, p < 0.01). In this study, the highest correlation was between meaningfulness and semantic distance (r = 0.925, p < 0.001). This finding implies that, if the sign’s content relates closely to the information needing expression, then the sign will be more meaningful.
The results demonstrated that usability had a significant relationship with all the features in the following order: meaningfulness, semantic distance, concreteness, familiarity, and simplicity. In other words, meaningfulness had the highest correlation with usability, while simplicity had the least correlation with usability. Similar results from a recent study conducted by Saremi et al. [27] indicated that meaningfulness is the best predictor of guessability for pictorial signs [27]. In the study by Chan and Ng [20], safety sign design features significantly correlated with their guessability. The highest correlation was reported between guessability and semantic distance. Liu et al. [24] examined four factors: semantic distance, confidence in guessing, familiarity, and complexity (simplicity). They found semantic distance to be the most dominant factor in traffic sign guessability. As the results of our study showed, the highest usability score was for the “No entry for vehicles more than 2 meters wide” sign, which had the highest score for meaningfulness among the 20 signs tested. Likewise, usability had the highest correlation with meaningfulness when compared with the other features. Therefore, meaningfulness is the most important feature affecting sign usability in this study. In addition, in the study by Ben-Bassat et al. [28], experts in ergonomics and human factors rated alternative and currently used traffic signs according to three sign design features: familiarity, standardization, and message/symbol compatibility (i.e., meaningfulness). Their study showed that most ergonomically designed alternative signs were rated higher than currently used signs in terms of message/symbol compatibility. They recommend that universal sign improvement should focus mainly on message/symbol compatibility. Their evidence and ours can provide support that leads to improving sign design for better comprehension., but it does not imply that the features of familiarity, concreteness, simplicity, and semantic distance should be overlooked.
Conclusions
This study was conducted to determine what effects the design of traffic signs and the characteristics of drivers have on the usability of traffic signs. Its findings have shown that the drivers’ characteristics of age, education level, possession of a driving license, and driving experience significantly influence the usability of traffic signs and that traffic sign designs correlate significantly with the usability of traffic signs. Therefore, an effort to improve traffic sign usability is warranted. This should be accomplished through driving instruction that considers the personal characteristics of drivers and through sign redesign based on meaningful graphic content of traffic signs.
Conflict of interest
None to report.