The Importance of Cognitive andMental Factors on Prediction of Job Performance in Chinese High-Speed Railway Dispatchers

School of Transportation and Logistics, Southwest Jiaotong University, Chengdu 611756, China National United Engineering Laboratory of Integrated and Intelligent Transportation, Southwest Jiaotong University, Chengdu 610031, China National Engineering Laboratory for Comprehensive Transportation Big Date Application Technology, National Development and Reform Commission, Beijing 100000, China Dispatch Office of China Railway Chengdu Group Co. Ltd., Chengdu 610082, China School of Mathematics, Southwest Jiaotong University, Chengdu 611756, China Psychological Research and Counseling Center, Southwest Jiaotong University, Chengdu 611756, China


Introduction
With the share of China accounting for about 2/3 of the world high-speed railway (HSR) mileage [1], the daily transportation of China's high-speed trains is about 1.33 million, which occupies 25.7 percent of the overall passenger traffic [2]. e HSR dispatchers take full efficiency and safety responsibility in conducting the high-speed trains' running in the network. Railway Accident (2011) reports that in the Yong-wen railway, a severe accident of two trains' collision resulted in 40 deaths and 172 injuries partly due to the HSR dispatcher's overlook of the malfunctioned signalling equipment and his failure to deliver the correct movement authority to the trains [3]. Although aided by many automated equipment, the safe and efficient dispatching tasks still heavily rely on the HSR dispatchers' multiple capabilities and mental states in emergency situations [4,5].
Previous research studies on job performance of other occupations have listed possible impact factors: (1) at the level of organization, including the work environment, assessment system, incentive mechanism, and corresponding training [6,7]; (2) at the individual level, including but not limited to job-related knowledge, professional skills, capabilities, stress, and emotional state [8,9]. ese impact factors have different predictive effects according to the variety of occupations. As an emerging occupation, HSR dispatching has its own particularity. Hence, we discussed the predictive factors of HSR dispatchers' job performance from the perspective of work analysis. As the central nervous system of train operation, HSR dispatcher is not only responsible for daily adjustment of train operation, dispatching order issuing, tracking, and controlling the train route but also needs to dispose all kinds of emergencies quickly at any time. Highly complex and important work content, which reduces the prediction of job performance of ordinary tests concentrated on job-related knowledge, professional skills, and capabilities [10,11], makes the establishment of strict selection and training system as a must [12][13][14].
e predictive validity of the cognitive ability depends on the complexity of the job with the strongest validity coefficients observed for highly complex jobs [15,16]. Literature [17,18] and task analysis of Chinese HSR dispatchers both reveal that train dispatching is a highly cognitively demanding task [19]. We first summarized HSR dispatchers' job responsibilities and the related cognitive abilities in Table 1.
Among cognitive abilities, working memory and multiobject tracking caught our considerable attention given their highly frequent requirements in HSR dispatcher's job. Working memory can be defined as a temporary retention of representations that were just recently experienced (but no longer exist in the environment) or recalled from long-term memory [20]. Working memory ability is viewed as one of the most important psychological constructs extensively investigated in the cognitive psychology [21][22][23] and is more highly correlated with the job performance than others constructs like the intelligence [24]. Working memory has been proved to be a better predictor than others, especially under the similar job cognitive requirements situation of air traffic control system [25]. N-back task is often employed to measure the working memory ability [26,27]. e multiobject tracking ability is also repeatedly reported to be important in facilitating job efficiency by improving the adequate situational awareness on performance from monitoring tasks and multitasking activities [28,29]. e multiobject tracking ability is viewed as an important factor for air traffic controllers' selection and training [30]. e multiobject tracking (MOT) task is used generally to examine the multiobject tracking ability of various professional capabilities [31,32]. As seen from the job responsibilities of the HSR dispatchers, the multiobject tracking abilities are also a suitable factor to predict the job performance. Because many tasks such as routing, monitoring, and managing emergency require searching and tracking multisource information on the control panel, visual tracking of dynamic objects is an important component of the HSR dispatching. In conclusion, we assume that the multiobject tracking ability should be included into the cognitive abilities set.
In addition to cognitive abilities, a series of studies showed that mental health also impacts job performance. In the field of traffic safety, drivers with mental problems have much worse driving performance in both simulator and road tests, for depression and anxiety could slow down reaction speed and increase accident rates [33][34][35]. Similar results have also been found in nurses and operating workers [36,37]. ough it has been widely confirmed [33,36,38,39], little is known among HSR dispatchers. e particularities of HSR dispatching make mental factors to be the validity predictors of job performance. For example, HSR dispatchers had to work 12 hours in each shift and their attention and audio-visual perception must to be always highly alert while working. Moreover, the working environment is monotonous and the workplace is completely enclosed. ese above unique work characters could easily lead to emotional problems such as depression and anxiety. Depression is the most prevalent emotional problem, and there is no debate about its impact on job performance [40]. Similar to depression, anxiety also plays its key role. While the absence of positive affect is specific to depression, the physiological hyperarousal is specific to anxiety [41]. Anxiety could impair employees' memory ability, cause insomnia, and make them more aggressive when communicating with others [36]. Research investigated that metro dispatchers had worse mental health than other occupation groups in terms of depression and anxiety [42].
Besides, highly cognitive demanding and safety responsibility work also lead to a state of stress. Survey report from Federal Railroad Administration (FRA) has suggested that most dispatchers perceived stress in their daily work [43]. Perceived stress, as an important composition of mental health, is always associated with negative cognitions and negative mood on the one hand [44,45]. On the other hand, moderate stress is necessary for HSR dispatchers to keep physiological arousal that helps constant vigilance. Hence, perceived stress need to be test as a predictor of HSR dispatchers' performance. Besides, studies have shown that social support can regulate negative emotions [46,47] and this makes the ability to perceive social support also be a mental health factor. erefore, factors such as depression, anxiety, perceived stress, and perceived social support were all under consideration as different dimensions of mental health in this study to support the predictive model of HSR dispatcher's job performance.
Meanwhile, cognitive factors may interact with mental factors. Previous studies have shown that mental health affects cognitive ability [48,49]. Typically, anxiety has a negative relationship with working memory performance [50]; depression could alter reaction time of multiple cognitive abilities including working memory and attention [51]. Vice versa, cognitive ability could also affect mental factors [52,53]. Related research found that high cognitive performance has been associated with better mental health, e.g., lower risk of psychiatric disorders [54][55][56]; low working memory capacity leads to an increased susceptibility to anxiety [57]. erefore, cognitive factors including working memory and multiobject tracking and mental factors such as depression, anxiety, perceived stress, and perceived social support should be taken into account in the job performance model that needs holistic monitoring and comprehensive disposition, but to what extend these factors can explain the impact on performance remains unclear. To the authors' best knowledge, this research is one of the first studies to explore the quantifiable relationships between the cognitive and mental factors and the job performance in the context of HSR dispatching, especially on the prediction of the jobrelated specific cognitive abilities. It would lead to one of the theoretical and practical bases for the HSR dispatchers' candidate selection and personnel managing. e aim of this article is to get insights into the role of the cognitive abilities and related mental factors in the HSR dispatching environment and to offer a theoretical prediction model on the job performance of the HSR dispatchers. e remaining part of this article is structured as follows. e methodology section explains the sampling techniques, the methods for measurements, and research procedures. e subsequent section is dedicated to the presentation and analysis of the results. e final section draws conclusions along with a discussion.

Materials and Methods
e HSR dispatchers were recruited from a branch of China Railway. Cognitive abilities including the multiobject tracking (MOT) and the working memory (WM) of these HSR dispatchers were measured. And mental factors including depression, anxiety, perceived stress, and perceived social support were measured through questionnaire. e conduction of the tests was in a random order across the subjects to control the carry-out effects. eir job performances were measured by both subjective indicator and objective indicator, i.e., the overall evaluation and the delay time.

Cognitive Ability Tests.
Two E-prime 2.0 programmed tests were selected to measure working memory and visual multiobject tracking abilities.
2.1.1. Working Memory: 2-Back Task. One of the most frequently used methods to measure working memory ability is the n-back task [26,27]. e 2-back task was conducted in our study. Specifically, a series of white letters (font at 48) were presented on the black background; each of them was shown for 0.5 s and participants were asked to determine whether it is the same as the letter that was presented two items back in the series. If the answer is same, participants were required to press "Enter" key as soon as possible; otherwise, no button should be pressed ( Figure 1). e reaction time (RT, ms) and the correct accuracy (ACC) were recorded as the performance of working memory. Each participant did 6 practice trials and completed 176 trials in the formal test.

Multiobject Tracking (MOT) Task.
Multiobject tracking (MOT) paradigms [58] are widely used to study the capability of tracking multiple moving objects simultaneously [59]. In each trial (see Figure 2), 16 squares (1 * 1 cm) including targets (colored red) and nontargets (colored green) appear at random locations of a 15 * 15 cm box. e number of targets was randomly determined ranging from one to seven. In the 5 s phase of attentive tracking, all targets turned green just identical to the nontargets first, and all objects started moving. en, all objects stopped moving, and participants were asked to judge whether one randomly selected object (marked in white color) was a target or not by pressing buttons. Similar to the above test, performance was measured by the reaction time (RT) and the correct accuracy (ACC). Each participant completed 49 formal trials after they have really understood the whole task through practice trials (4-8 trials).

e Zung Self-Rating Depression/Anxiety Scale (SDS and SAS).
e SDS and SAS [60] are two widely used scales to quantify depression and anxiety symptoms. Each of them included 20 worded items that rated from 1-4 to assess the level of depression or anxiety, and higher scores correspond to more frequent symptoms. Both SDS and SAS have good reliability and validity [61,62], and Cronbach's alpha in the current study is 0.757 and 0.82, respectively.

e Perceived Stress Scale (PSS).
e PSS [63] is one of the most commonly used instruments to measure stress perception. 14 items in PSS were designed to measure the extent to which one's life is perceived as "unpredictable,

e Perceived Social Support Scale (PSSS). e PSSS [64] is the Chinese version of Zimet's Multidimensional
Scale of Perceived Social Support (MSPSS) [65] to measure actually perceived social support. It has 12 items rated on a 7-point Likert scale and is divided into three dimensions: family support, friend support, and other support. Cronbach's alpha is 0.88 in the current study.

Evaluation of Job Performance.
e HSR dispatcher's job performance was measured by both subjective indicator (overall evaluation) and objective indicator (delay time). e overall evaluation was made by their supervisors and the delay times were calculated by the HSR dispatching simulator.

Subjective Indicator: Overall Evaluation.
e participants' overall evaluation was given by five supervisors, including two dispatching directors (from two different shifts), a technical training director, a safety director, and an administration director from the HSR dispatching station. e overall evaluation was rated on a 9-point Likert scale (1, completely poor; 2, very poor; 3, poor; 4, rather poor; 5, neutral; 6, rather good; 7, good; 8, very good; 9, completely good) through the question "Please give an overall evaluation of the dispatcher's job performance." Before the question, there was a guide paragraph that specified the scoring criteria which should be rated from HSR dispatchers' five job responsibilities including schedule preparation, operation monitoring, information collection, emergency disposal, and data management. Similar one-item measure of overall job performance has been used in previous studies [66,67].
To avoid the inaccurate evaluation from the supervisors who were not familiar with the dispatchers, a 9-point Likert scale question "How well do you know the dispatcher's job performance of all five job responsibilities?" was also asked. Data from the supervisors with familiarity scores lower than 3 were excluded, which accounted for 3% of all the supervisor-dispatcher pairs. Afterwards, overall evaluation ratings from the remaining supervisors were averaged for each dispatcher.
e Pearson correlation analysis was used to examine the relations between the familiarity and the job performance (r � 0, p � 1), which showed that the supervisors' ratings on the job performance were not likely to be affected by their familiarity with the dispatchers. Also, the r WG values of the job performance ranged from 0.93 to 1 (mean � 0.98; median � 0.98), reaching the cutoff value of 0.7 generally recommended for good inter-rater agreement [68].

Objective Indicator: Delay Time.
Delay time is an objective indicator to reflect HSR dispatchers' job performance. In most circumstances, trains could run normal according to the operation diagram. But they would be stopped or be slowed down when encountering emergency scenarios such as equipment failure, natural conditions, or human causes, which would cause delay for minutes or even hours. Dispatchers need to ensure trains are running on time to the greatest extent when they enter the jurisdiction. Ensuring the running time of trains is an important part of dispatcher's daily shift and job performance.
To test this objective indicator while controlling all uncertain variables, we use a fully simulated dispatching simulator for data collection. e dispatching simulator was developed by the authors' team (see Figure 3), which was ey were required to avoid alcohol and caffeine on the day of experiment as well as to refrain from eating or exercising two hours before the study. Firstly, participants gave informed consent that participation was voluntary, and it would neither affect their performance evaluation nor have any other consequence in the company. Next, a questionnaire collecting demographic information was conducted, including variables that could be correlated to dispatching skills or cognitive abilities, such as age (mean � 34.76 and SD � 3.25 years), HSR dispatching experience ranging from 1 to 9 years (mean � 4.51 and SD � 2.22 years), and education (82% had a bachelor's or higher degree), and variables that could be possibility correlated to mental factors, such as marriage (89% were married). en, all the participants completed 4 mental health tests (integrated into a long questionnaire test) and 2 cognitive tests. e order in which participants took the mental health questionnaire and cognitive tests and the order of WM task and MOT task were all random and counterbalanced. A short break of about three minutes was provided between any two tests (see Figure 4).
Next, the simulator experiment was conducted. In this part, the participants firstly spent several minutes to be familiar with the simulator until they had mastered the basic operation. Secondly, participants were informed of the emergency scenario that needs disposal. e emergency scenario designed by highly experienced HSR dispatcher was the indicator of one of the rail switches turning into a wrong status. According the operation and safety regulations, eight process were included in this situation: (1) an alarm window was popped up on the control panel and the alarm sound was broadcasted; (2) the dispatcher identified the failure and commended the HSR driver to stop the train; (3) the dispatcher confirmed the failure and started the emergency procedure on the control panel; (4) the dispatcher informed the concerned personnel to make preparation; (5) the dispatcher adjusted the train operation plan for safety protection; (6) the station attendant sent a request for track-checking to the dispatcher who would issue the related dispatching order; (7) the dispatcher asked the on-site personnel about the situation; (8) the train operation was recovered after the failure was solved. Participants were asked to operate the simulator in actual situation to deal with the failure situation. ey did the trial twice including a practice and a formal test. e delay time of all the trains in jurisdiction was obtained from the formal test. Specific operation steps are illustrated in Figure 5.
Participants operated the simulator about 2 hours that depended on their dispatching ability. Each dispatcher participant was given a gift for their participation. At last, 70 questionnaires and simulator data were validated, giving a validity rate of 100%.

Results
On basis of the qualitative analysis, this section would explore the analytical relationship between the job performance and the possible variables from a quantitative perspective. As shown in Figure 6, firstly, we divided the dispatchers into the good performance group (Good) and the poor performance group (Poor) and established the threshold for them. en, in order to establish a predictive model of the dispatcher's job performance, the stepwise regression method was adopted into model of the relation between the HSR dispatchers' job performance and the possible variables. Finally, according to the threshold and the predicted delay time, the dispatchers were identified as Good and Poor. e validity and accuracy of the model would be tested by the comparison between the predicted group and the original group. Specific data analysis procedures are shown in Figure 6.

Descriptive Statistics of All Variables.
In the previous section, the job performance of HSR dispatchers was measured by the subjective and the objective indicators, i.e., the overall evaluation and the delay time. Four demographic variables (age, dispatching experience, education, and marriage), four cognitive variables (ACC of MOT, RT of MOT, ACC of WM, and RT of WM), and four mental variables (anxiety, depression, perceived stress, and social support) were the possible variables to predict the job performance. e mean and standard deviations of these variables are shown in Table 2. Among these demographic variables, "education" and "marriage" were not included in the following statistics for three reasons: (1) these two are dichotomous variables; (2) most of the answers are in one category (82% had a bachelor's or higher degree and 89%     Journal of Advanced Transportation were married); (3) independent t-test showed that they did not affect other variables (ps > 0.05).

Group Division and reshold
Determination. e overall evaluation was a comprehensive evaluation based on the dispatcher's aggregate performance which was susceptible to the personal preferences of the supervisors. erefore, the overall evaluation had certain subjectivity and arbitrariness. Additionally, the delay times indicated the dispatching ability of the participants to deal with emergency event by operating the simulator, which had certain volatility. As a result, the subjective indicator and the objective indicator were combined into the classification of the dispatchers and the threshold was jointly determined.
Firstly, the overall evaluation was used to stratify the groups. In order to make a large dispersion between groups, the dispatchers were stratified into three groups: Good (7-9 scores), General (general performance) (4-6 scores), and Poor (1-3 scores). In particular, 24 (34%) were Good, 20 (29%) were General, and 26 (37%) were Poor, respectively. e General group was a transition part between the Good and the Poor ones, with the possibility to interfere or to misjudge the results. In this case, the General group was removed from the groups. e participants were stratified by the overall evaluation into two groups and they were sorted according to the delay time from small to large. Since the subjective indicator had certain subjectivity and arbitrariness, the delay time was used to examine the group consistent with overall evaluation. e top 34% of the delay time was stratified into the Good group, and the bottom 37% was stratified into the Poor group. e participant would be removed when the group of delay time was inconsistent with the group of overall evaluation. In this case, a total of 2 (4%) participants were removed, indicating that the subjective indicator and the objective indicator are of practical unanimity (r � −0.957, p < 0.01). In the end, the final sample of 48 dispatchers was obtained from 70 valid data, including 23 Good and 25 Poor ones.
Finally, to make clear the predicted values obtained by the stepwise model in Section 3.4 belong to the Good or Poor group, it was necessary to determine the threshold. Given the groups had been consistently confirmed by overall evaluation and delay time, the maximum delay time (229 min) in the Good was adopted as the threshold. When the predicted value was 229 min or less, the participant was part of the Good group. In other words, when the predicted value was greater than 229 min, the participant was classified into the Poor group.
After grouping, the results of the difference in the demographic variables and the cognitive variables between Good and Poor are presented in Table 3. Two demographic variables, four cognitive variables, and four mental variables were significant differences between Good and Poor (ps < 0.05).

Correlation of Variables.
Two-tailed Pearson correlation analysis was performed to examine the relationship among independent variables (including demographic variables, cognitive variables, and mental variables) and dependent variables (both subjective and objective indicators). As shown in Table 4, the correlation between all independent variables and dependent variables was significant with absolute correlation coefficients ranging from 0.35 to 0.91 (ps < 0.05), which provided basis for establishing regression model for next step. And significant correlations were found between most independent variables. is indicated a multicollinearity relationship between these variables, and this provided basis for stepwise regression.
ere was a strong correlation between overall evaluation and delay time (r � −0.95, p < 0.01); therefore, we only used the delay time (objective indicator) as dependent variable in the predictive model.

Stepwise Regression Model.
In this study, the stepwise regression model was established to identify the variables contributing to the delay time. Eleven variables were entered in the process of stepwise and a total of five regression models were established. e model with the best fitting effect which is presented in Table 5 contained five variables: ACC of WM (β � −0.33, p < 0.01), RT of WM (β � 0.27, e result also showed the quadratic sum of regression and the residual. In addition, Table 5 revealed that lower ACC of WM (β � −0.33, p < 0.01), longer RT of WM (β � 0.27, p < 0.01), longer RT of MOT (β � 0.17, p < 0.05), higher depression (β � 0.41, p < 0.01), and lower perceived stress(β � −0.18, p < 0.01) were significantly correlated with longer delay time which indicates worse job performance. According to the stepwise regressions results, we use a statistical model that estimates the relationship between cognitive abilities, mental health, and job performance: where 128.722 denotes the constant. e vector X 4 is RT of MOT, X 5 is ACC of WM, X 6 is RT of WM, X 7 is depression, and X 9 is perceived stress.

Model Test.
In order to test the model of cognitive abilities, mental factors, and job performance on stepwise regression, five independent variables (ACC of WM, RT of WM, ACC of MOT, depression, and perceived stress) were considered back into the regression model. erefore, each participant got a new predicted value of delay time. As shown in Figure 7, the threshold value (229 min) and the delay time including the original value and the predicted value were presented. In Figure 7, we observed the predicted value wandered around original value, suggesting that there was no significant difference between them. Moreover, there was a significant jump of value in 229 min between Good and Poor. In fact, there might be a misclassification of Good and Poor. For example, there were two types of statistical errors:  Variable  the RWT errors and type B error, which are common errors.
In order to get a more detailed understanding of the accuracy of the stepwise regression model for the division of Good or Poor, the predicted value of the participant was grouped according to the threshold value. As shown in Table 6, 96% of the Good were correctly predicted, while 100% accurate for the Poor. From Figure 7 and Table 6, it was apparent that the number 21 was misclassified. In order to visually present the results of difference in participant no: 21 and groups between demographic variables, cognitive variables, and mental variables, the radar map was considered. Given the disunity of scale for all variables, the min-max normalization was conducted to solve the problem by the following formula: where i � 1, 2, 3, . . ., 48, j � 1, 2, . . ., 10. By this formula, all variables were unified from 0 to 1. Figure 8 presents the results of mean of each variable after unified scale of Good and Poor in full line and the number 21 in imaginary line. e features of number 21 were more consistent with Good. However, compared with the mean of Good, the participant's ACC of WM, RT of WM, RT of MOT, depression, and perceived stress were obviously different. e five indicators were used to predict the job performance. erefore, the participant was misclassified.

Discussion
e goal of the current study is to examine the relationship between the cognitive abilities, mental factors, and the job performance of the Chinese HSR dispatchers. Findings showed that the multiobject ability, the working memory ability, and 2 mental factors have strong relations with the job performance (indicated by delay time), and tremendous distinctions between groups of good and poor performance of the HSR dispatchers were revealed. In addition, the stepwise regression modelling is employed to construct the regression equation, in which the demographic data, the accuracy and the reaction time of working memory, the reaction time of multiobject ability, the depression, and the perceived stress appeared as the independent variables and the job performance appeared as the dependent variable. e current research findings can offer indicators and models in future HSR dispatchers' selection.
Consistent with existing studies [24,69], the positive relationship between cognitive abilities and dispatchers' job performance is supported in the current study. Working memory, which included both short-term storage and central executive control [29], is exactly what the dispatcher's job requires according to task analysis (Table 1). It is expected and understandable that working memory can significantly predict HSR dispatcher's key performance, just as working memory has demonstrated to be a good predictor for shooting behavior [70], school grades [71], and other performance [72]. HSR dispatchers should also dynamically monitor the operational status of multiple trains as well as the conditions of signals, stations, switchers, and weather, which involves dynamic visual tracking of multiobjects [73].
at leads to the ability of multiobject tracking, becoming an important predictor of job performance just as the similar job of air traffic controllers [30]. e result that mental factors were also predictors of the job performance of HSR dispatchers in this study is expected, which is consistent with results of other related occupations in previous studies [33,34,74]. Depression, as the most concerned mental health symptom, is associated with low self-esteem, reduced work motivation, and slower cognitive processing [33,75]. For HSR dispatchers, depression could directly affect their subjective and objective performance indicators on the one hand. On the other hand, it could damage cognitive abilities including both WM and MOT, indirectly affecting the job performance. Although overload pressure may always accompany with fatigue, insomnia, neurological disorders, and other symptoms that cause worse performance [76], moderate pressure could also promote work motivation which leads to even better performance [74]. In the current study, perceived stress was measured. e results of more perceived stress are associated with less delay time which means better job performance. It was most probably because of the particularities of HSR dispatching work that need continuous vigilance or the relatively moderate perceived stress score (mean � 18.02 and SD � 7.76). Inconsistent to other studies, anxiety and perceived social support did not enter into the predictive model of job performance [77,78]. is could be understandable from aspect of job analysis. In actual work, there are not many cases which require emergency dispatching though   constant vigilance is needed, while anxiety is mainly related to emergency situations or unmanageable events. Besides, the perceived social support scores obtained in this study were relatively concentrated in high section, while the ceiling effect makes this variable could not distinguish between the Good and Poor groups.
To summarize, the current study establishes a prediction model for the job performance of Chinese HSR dispatchers. e quantitative relationship between cognitive abilities, mental factors, and the job performance, to the best knowledge of the authors, has rarely been examined in previous study. It confirms that cognitive abilities including multiobject tracking and working memory and mental factors are important for the job performance among the existing qualified HSR dispatchers. e predictive variables including the ACC of WM, the RT of WM, the ACC of MOT, depression, and perceived stress explained an additional of 91% of the variance in the job performance. Moreover, the job performance model gets the accuracy up to 96%. ese findings suggest that tests of the 2 cognitive abilities and 2 mental factors may be useful tools complementary to the current practice of HSR dispatcher selection and assessment.
Although the traditional selection and assessment can investigate the dispatcher's dispatching knowledge and professional skills to a certain extent, the operational adaptability and mental factors might be more accurate potential indicators which have not yet been considered in terms of affecting the dispatcher's operational behavior safety especially in emergency scenario. Taking an emergency scenario of the dispatcher encountered an abnormal train noise, for example. However, in order to maximize the train operation, dispatchers should perform multiple tasks and organize railway personnel with different responsibilities quickly (even in seconds) and accurately. is makes dispatchers' basic cognitive abilities mainly working memory and multiobject tracking become the key factor of impacting their job performance. Regarding mental factors, the state of depression might lead to slower or less accurate reaction that would produce irreversible consequence, while perceived moderate stress would help to maintain alertness and promote more validity operation. Only when more comprehensive impact factors are included in actual HSR dispatcher's selection and assessment, our daily transportation will be safer. erefore, building the prediction model taking both cognitive and mental factors into account is not only of theoretical significance but also of practical significance.
A limitation of the current study is that the HSR dispatchers were all from one group, China Railway Chengdu Group. Future studies are needed to examine the predictive model among other dispatcher populations such as other geographical regions in China Railway and other countries. In addition to mental factors such as depression and perceived stress, other psychological factors such as personality and motivation may also be good predictors of job performance [79]. Future studies are needed to examine these additional factors and their joint influence on the HSR dispatchers' job performance.

Conclusions
In conclusion, both cognitive factors specifically multiobject visual tracking and working memory and mental factors such as depression and perceived stress are significant predictors of the HSR dispatchers' job performance.
e results from the current study provide a theoretical basis for the training of HSR dispatchers and can support future work towards establishing standardized methods and criteria for dispatcher selection and management. Additionally, the subjective and objective indicators have a high degree of consistency, which also provides a theoretical basis for the selection tools. In the next step, we will try other methods (data mining, artificial intelligence, etc.) to explore the relationship among these cognitive abilities and much more psychological factors. Data Availability e data used to support the findings of this study are available from the corresponding author upon request.

Conflicts of Interest
e authors declare that there are no conflicts of interest regarding the publication of this paper.