Keywords

1 Introduction

In the traditional design and research procedures, researchers tend to base clear and objective fundamental principles on previous questionnaires. Then, researchers and designers would combine their subjective opinions to explore problem space and establish problem framework. By so doing, they give rise to design concepts and use them as the primary directions for design. For instance, in a research on the railway station signs comprehensible to the elderly conducted in 2000, Japan Railways carried out a satisfaction survey on 260 senior citizens in terms of font size, spacing, color in the signs by filling in questionnaires and answering questions orally [1]. In the design and research of signs for the elderly, [2] conducted a satisfaction survey on signs among the elderly in the residential areas in Beijing through questionnaires [2]. In the research on the subway station signs and evaluation research from the perspective of the elderly, [3] and her colleagues did a survey on the elderly passengers at the Guangzhou Subway Station through questionnaires. Evaluation and analysis of univariate descriptive statistics were made to offer suggestions for the sign design at the Guangzhou Subway Station [3]. Data can be collected within certain domain in a short span by means of such traditional survey questionnaires, which provides support for design later by analyzing objective data. However, quantified analysis lacks essential support for inaccurate and non-value data. Therefore, while analyzing quantified data in previous literature, this paper adopts Semantic Differential (SD) to analyze the feelings of the elderly after viewing the signs, set the scale of “sign semantic”, evaluate and analyze the descriptive parameters of all scales, describe the concept and construction of spatial targets with figures, and establish correlations between demands and sign design factors. This paper, with the elderly as the subjects, conducts SD-based cognitive evaluation of the Beijing subway signs by letting the elderly take the subway with questions provided and through distributing questionnaires, and constructs an “evaluation system” of the subway signs. Here is a list of some of our main proceedings series:

  • First, it points out that the traditional questionnaires for subway signs tend to be temporary and unfinished; whereas semantic analysis in the past is mostly objective and denotative when it comes to the cognition of the elderly toward subway signs. As a result, in this chapter, a guiding experiment from cognitive logic to systematic logic is carried out. Semantic terms summarizing and describing the subway signs are collected, analyzed and later selected, screened and matched by combining the previous literature. Adjective terms in the SD scale evaluation axis are made and evaluation scale designed, thus constructing a scale that targets evaluation of the elderly toward subway signs.

  • Next, native and non-native elderly people are invited to take the subway with questions provided. After that, based on the results of 60 SD evaluation data files, this paper uses factor analysis to extract the cultural, physiological and psychological factors. Combining that and the Functional Model Theory proposed by Kantowitz [7], this paper subdivides the cognitive process of the elderly toward subway signs into: cognitive system, processing, execution and maintenance system in the progressive order.

  • At length, variance analysis is adopted to identify the distinction between the native elderly and the non-native ones in their evaluation of the cultural factor. The research shows that due to the asymmetry in the cultural background, needs for information of the non-native elderly and designer model of the Beijing subway signs, their evaluation is comparatively low. Against the vacuum in the research of subway signs for the elderly, this chapter establishes the problem framework through the results and analysis of this research.

2 Empirical Study Based on SD

From March 1 to October 10, 2016, during the period of the survey, 60 elderly people were invited to take the subway, participate in a test and evaluate the SD scale. Prior to that, preparations for the survey were done, including specifying research objects, collecting adjectives in the factor axis and determining the scale of evaluation to complete the SD scale. Procedures of the experiment include evaluation of the basic characteristics of analysis on the average, application of factor analysis in pursuit of internal relevance and pattern within the semantic and then classification. Procedures are as follows (Fig. 1):

Fig. 1.
figure 1

(Made by the Writer)

Evaluation system of the subway signs

Full size image

Design of the evaluation scale based on SD means conducting semantic summarization and description toward the relevant factors of the subway signs in accordance with the characteristics and research targets of the signs. That is to say, via the guiding experiment procedures from cognitive logic to systematic logic, selecting the typical semantic space based on the semantic cognition of the elderly. First, in order to prevent subjective decisions made by the researchers in selecting semantic terms and use terms in line with the cognitive features, knowledge and experience of the elderly, this research deploys Evaluation Grid Method (EGM) [4]. It starts with the cognition and feelings of the elderly, and by way of comparing strong points and weaknesses of the design, identifies their initial comprehension of the signs. Then, it guides the elderly to divide their appeal into different layers, with comprehension emphasizing details at the bottom of the pyramid and that based on systematic feelings at the top [5]. Construction of user feelings is shown as below (Fig. 2).

Fig. 2.
figure 2

(Made by the Writer)

Construction of user cognition

Full size image

Experiment Approaches:

  • Researchers explain to the subjects about experiment procedures and rules;

  • Select five pictures of the good subway signs home and abroad as experiment materials through research on previous literature.

  • Ask the subjects to compare the pictures and rank them according to in sequence. As the subjects are fully engaged in the experiment, the researchers ask them to comment on the current Beijing subway signs and propose new requirements by listing “shortcomings” and “possible improvements” respectively.

  • After the subjects have finished the test, the researchers interview them and divide their answers into top, medium and bottom concepts. Positive as well as negative ones are determined, thus being used as the evaluation indicators in the scale.

Through the guiding experiment from cognitive logic to systematic logic, the elderly users help or independently design the cognitive dimensions at different levels, including the bottom concept emphasizing details and the top concept based on systematic feelings. Six unified, clear, simple, straightforward and bold adjectives indicating directions are obtained, which can be used in making the adjective terms in SD scale evaluation axis. Yet previous research shows that seven adjectives are still subjective and broad for the summarization and analysis of signs. While they may be insufficient to construct the SD scale evaluation axis, generally 15–20 will do. In addition, seven semantic terms clearly cannot summarize all the characteristics of the signs and there might be some downsides (using pictures as the experiment materials may fail to present complete systematic space to the subjects). For example, whether the circulation plan is smooth or not and the continuity of how the signs are set cannot be evaluated. Besides, how the elderly living with cataract view the signs cannot be observed.

According to results of the questionnaires and features of the Beijing subway signs, 15 evaluation items are selected and are converted into adjective phrases comprehensible to the elderly based on their cognitive traits. Eight of them are about the features of signs: planned locations of the signs, specification of the destinations, specification of the main roads, comprehensible marks, readable combinations of colors, understandable formats, straightforward information and sufficient lighting; The remaining seven are about the feelings of the elderly: sense of security, direction, memory, pleasure to the eye, relaxation, dependency and subjective evaluation (Table 1).

Table 1. SD semantic terms.
Full size table

This paper classifies the semantic terms into positive and negative categories and hence constituting adjective terms (Fig. 3). Evaluations are made based on the bi-polar principle in the hope of obtaining more objective data. The semantic evaluation factor axis is evaluated as per seven sections of an hierarchical structure, with the score ranging from −3 to +3. Take “lighting” as an example, the corresponding semantic terms are “dim” and “bright”. Based on the lowest score of −3 and the highest +3, the elderly subjects evaluate in the light of their personal feelings.

Fig. 3.
figure 3

(Made by the Writer)

Evaluation scale setting

Full size image

The evaluation scale is made through the aforementioned cognitive construction research, selection of semantic terms and evaluation criteria setting. The scale is primarily used to evaluate the feelings of the elderly after they have taken the subway with questions provided.

3 Survey and Research on Taking and Transferring Subway Trains by the Elderly

3.1 Survey Targets and Approaches

The subjects are required to take the subway with questions provided on the basis of the evaluation scale established according to the research above. The survey includes two parts: first, basic survey on the inside, entrance and exits of the Beijing subway stations; second, record and survey the behaviors of the elderly passengers while they take and transfer subway trains, and then analyze their needs for signs.

3.2 Selection of Subjects

  • Selection of the subjects participating in the test: use MMSE (Mini Mental State Examination) scale to conduct the test, thus ensuring that the subjects comply with the experiment requirements.

  • The famous American scholar Collier [6] points out that comparative analysis is a research method targeting at small samples or small cases. By focusing on the comparable samples and controlling the number of the variables, it increases the credibility of the experiment.

  • Confirmation of the Tested Sample Quantities: in user experience design, researchers always call for user-oriented design. But in actual work, due to time and resource constraints, they cannot but make decisions on their own without data or with partial data. Owing to problems such as “how many sample will suffice to reveal the problem”, “whether the result is reliable or not”, etc., it is difficult to search for user demands. This test chooses 60 samples (aged 67 on average) by applying the problem-spotting probability model to evaluate the sample quantities. By combining the realities of the elderly Chinese and the research target, this paper determines on the basis of probability that the age 65 is the boundary. Although some research argues that individual cases vary a lot, seen from probability, when people turn 65, their cognitive abilities and physiological functions degenerate. The boundary conforms to the standards using Nominal Group Technique (NGT).

  • Categorization of Samples: in this research, the elderly subjects are categorized into the native ones (30) and the non-native subgroup visiting Beijing (30). The native ones are also dubbed as the road-finding experts because they are familiar with the Beijing subways and the local culture. The non-natives visiting Beijing are those who take the subway for the first time and are not familiar with the Beijing Subways and the transfer lines. But the latter are required to be literate.

3.3 Cognitive Walkthrough

This paper adopts Cognitive Walkthrough while recording the behaviors and experience of the elderly who take the subway, as the results are unpredictable and the subjects might be missed out. During the test, the researchers treat themselves as XX and personally experience it without any hypothesis based on previous experience. Through months of actual survey, hundreds of data files are obtained.

3.4 Procedures of the Test

  • For each train-taking and transferring task, it is designed as choosing a destination in the subway map that the subjects have never been to. The subjects will need to transfer during the course (Fig. 4);

    Fig. 4.
    figure 4

    Hierarchical task analyses of the signs during the test

    Full size image

  • Each subject chooses a train-taking and transferring task randomly from the list. If they have been to the destination, then re-choose one;

  • To ensure objectivity and authenticity, the subjects are only allowed to arrive at the destinations with the guidance of the Beijing subway signs. They should not ask others for help, nor should they resort to maps and smartphones.

  • During the process, a researcher will record how each subject is doing with a smartphone to capture informal conversations. The subjects should describe their feelings toward the entire process to the researchers, which should be noted down. The researchers may ask appropriate questions, but should not help searching for information.

  • During the process or after the test is done, the subjects fill in the “SD Survey on Subway Signs Form” with the assistance of the researchers.

4 Results of the SD Survey

Input and calculate the data files regarding the 60 subjects and questionnaires with EXCEL. The average evaluation value of the two groups of subjects toward the subway sign sematic terms can be obtained (calculate the average value of each semantic term sample and then delete the largest deviation in the sample). Then input the value into the SD evaluation curve scale.

4.1 Factor Analysis

This paper contains as many as 15 summarizing and descriptive semantic terms as regards relevant factors of the subway signs. Excessive descriptive variables might complicate the research and analysis later. In a bid to analyze the evaluation of the elderly passengers toward the Beijing subway signs, this paper extracts the major common factors among the intricate adjective variables through factor analysis, input the original indicator data after standardizing it, and conduct factor analysis. Before that, indicator inspection shall be conducted to make sure that it is suitable for the analysis. This paper deploys KMO and Bartlett’s test for the inspection and the results are shown in Table 2 as follows:

Table 2. KMO and Bartlett’s Test
Full size table

Inspection shows that: KMO probability is 0.00 < 0.01, meaning the variables are interdependent. Therefore, it passed the KMO test and the KMO value is 0.519, indicating that it is suitable for factor analysis.

Judging from the aforesaid parts, the indicators in this research are suitable for factor analysis. Hence IBM SPSS Statistic (Version 21.0) is used subsequently to conduct analysis and Principal Component Analysis (PCA) is applied. Ultimately, three factors are extracted through cumulative variance explained rate and Screen Plot analysis factors. The characteristic value of the three factors are respectively 3.320, 2.036 and 1.861, while the rotated variance explained rate are 22.135%, 13.571% and 12.406% respectively. Besides, it can be told from the Screen Plot that when four factors are extracted, the fourth one will become comparatively stable. While if three factors are extracted, they will be steep. Therefore, it is more appropriate to extract three and the final accumulative variance explained rate is 48.112%. Lastly, the research rotates the factors using Varimax Rotation and the factor loading matrix thereby obtained is as follows:

Finally, name the three factors as physiological, psychological and cultural factors. However, such naming is not specific enough. Therefore, it requires a comparatively stable theoretical framework to summarize the research and describe the cognitive process of the elderly toward the signs. By combining factors and the Functional Model Theory proposed by Kantowitz [7], the cognitive process of the elderly toward the signs can be divided into: cognitive system, processing, execution and maintenance system in a progressive manner. Among them, the information cognitive system indicates the source of information regarding the cognition of human senses toward the basic attributes of objective things. Combining Table 3, it can be told that: Factor 2 corresponds to four indicators, comprised of the complexity of color identification, whether the signs can be clearly seen, whether the setting of the sign locations is suitable for the heights of the elderly, characteristics of the five senses, and the influence of lighting. Based on whether the subway signs meet the needs of the elderly passengers in terms of their physiological features and with information received about the signs through the cognitive system, procedures including cognition, interpretation and evaluation are established, which lays a solid foundation of cognition for information processing in the next step. As a result, they are named as physiological factors.

Table 3. Rotated Component Matrix a
Full size table

Nonetheless, divide arises in the bridge connecting the information processing system, indicating mismatch between the physiological functions of the elderly and the current signs. Hence, we need to target those living with eye conditions such as cataract, maculopathy, narrowing horizon, etc., and establish a “coding system” of comprehensible signs by means of quantifying the brightness and color of the current sign environment, conducting Cognitive Walkthrough research by experimenting and wearing cataract patients experience glasses and designing as well as comparing literature and standard theoretical research. In doing so, a cognition interface can be provided to activate the information processing system, thus avoiding negative feelings triggered while the elderly passengers are viewing the signs and realizing they are “aging” because of the degenerating physiological functions [8].

The information processing system means conducting context construction, establishing code of conduct and arriving at certain conclusions or decisions with the obtained information. It is the premise for the executive system to act. Factor 3 corresponds to five indicators, i.e. pleasure to the eye, continuity of the signs, subjective opinions, unity and dependency. They represent the psychological recognition of the elderly toward subway signs, hence named as psychological factors. Nevertheless, psychological changes in the crystallized intelligence and fluid intelligence of the elderly lead to a divide in the bridge connecting the executive system. Consequently, the executive system fails to act. In view of that, we need to focus our efforts on the crystallized as well as fluid intelligence [9] of the elderly, and establish a decoding system that strengthens the input, storage, guidance and free-forward control of the psychological schema. That will facilitate effective performance by the executive system.

Via the cognition interface established based on the physiological cognitive system, conduct psychological processing of the information and carry out the act through the information executive system. Factor 1 corresponds to six indicators, including sense of security, sense of direction, memorability, loading capacity of the information, comprehensibility and Straightforwardness. But variance analysis below shows that there is distinct difference between the evaluation of the native group and non-native group in the statistics sense. Specifically, the native group give a higher score to common factor 1, while the non-native group lower. That is because the latter are unfamiliar with the transfer lines and locations of the Beijing subways, thus getting lost, confused, wandering during the process. It results in a divide in the bridge connecting the executive system and is named as the cultural factor. Given that, a knowledge framework that elevates “cross-cultural adaptability” to the subway signs should be established. Building on the aforementioned conclusions, this paper also gives play to the scientific factor establishment and maintenance system, making sure that the three factors function soundly. To make that happen, smart facilities shall be applied to put forth the idea of designing “exclusive” signs for the elderly. In this manner, straightforward sign information can be provided to the elderly, their energy can be saved while trying to find the roads, and their sense of fulfillment can be enhanced.

4.2 Variance Analysis

This part aims to study the abovementioned two types of samples with two-way variance analysis. Analysis will be made on whether they are significantly different in terms of the three factors and if there is obvious variation between the average of each factor. Conclusion will be drawn for further study and made into Fig. 5.

Fig. 5.
figure 5

Results of variance analysis

Full size image

Figure 5 shows that there is no evident difference between the physiological and psychological factors. But the two groups give comparatively low score to the two factors, which speaks to the fact that the current sign design does not accord with the physiological and psychological features of the elderly. For instance, the sign information is discontinuous and signs as well as advertisement boards are mixed, compromising readability and causing visual interruptions. Moreover, the setting of color, height and lighting does not take into consideration the viewing distance of the elderly passengers [9]. As a consequence, it is difficult for the cognitive system to experience the basic attributes of the objective things relying on senses. Nor can it provide a fine cognition interface for the procession and execution of cognitive information later.

As a whole, noteworthy distinction is seen between the two groups in terms of the cultural factor, as p < 0.01. To make it more specific, the average score of the non-native group is −1.19 while the native 0.88. Furthermore, apparent distinction is seen among all the six indicators corresponding to factor 1. As for security, the native group score 2.20 on average while the non-native one −1.03, showing that the former give higher evaluation to security indicators. Similarly, variation is spotted in the following five indicators: discontinuous – continuous, forgettable – memorable, complicated – simple, abstract –straightforward, incomprehensible- comprehensible. Average score of the native group are respectively 0.47, 0.40, 0.80, 1.67, 0.53, while of the non-native group −1.63, −1.83, −1.67, 0.63, −1.60. The native group scores higher than the non-native one at all the six indicators, showing that they better recognize them. Study shows that due to the accumulation of experience and achievements in learning, the elderly can suffer from less cognitive pressure and can be less dependent on signs. Besides, it is helpful to lengthen their short-term memory and make them feel more secure. Therefore, the native group gives better evaluation to the Beijing subway signs than the non-natives in terms of the cultural factor. On the contrary, owing to a lack of user model cognition (cognition of the users toward the signs) toward the locations, subway stops as well as circulation design inside the subway stations in Beijing, there is asymmetry between the needs for information and designer model of the Beijing subway signs (As designers tend to name the stations according to local culture and construct the stations based on the geographical features, they may not understand or predict the difficulties the elderly encounter when taking the subway. One reason is that they are familiar with the design and have a detailed as well as complete Mental Model.). On top of that, degenerating cognition of the elderly passengers (such as memory, logical thinking and degenerating senses) make the non-native elderly unclearly and inaccurately comprehend the System Object Model and cannot continuously construct as well as renew correct psychological schema. Consequently, they get lost, confused or become resistant against taking the subway. The result coincides with the Cognition Friction Model established Norman [10].

5 Summary

In the past, designers opt to make judgment issues in designing based on their experience and inspiration, which can be unspecified and ambiguous. With the rapid development of the Internet technologies and the expansion of design objects, the dualism featuring traditional function and form as well as art and science gradually converts into problem-oriented research model, and resolution to problems in designing is acquired based on relevant knowledge. This paper, with the elderly as the subjects, conducts SD-based cognitive evaluation of the Beijing subway signs by letting the elderly take the subway with questions provided and through distributing questionnaires. Later, by analyzing the research results with the help of statistics, it specifies the problem framework and constructs an evaluation method that is systemic and readily available. That is helpful for the stable output of the design.