Traditional Culture Encyclopedia - Traditional stories - Man-machine dialogue in psychological test

Man-machine dialogue in psychological test

Man-machine dialogue in psychological test

Man-machine dialogue in psychological test I. Man-machine dialogue: new changes in the field of talent evaluation

Informatization is a revolution. It has brought about the change of interpersonal communication and the change of human thinking mode. Because "what is society (regardless of its form)? It is the product of interpersonal communication. " The rise of information society, network society and virtual space has strengthened people's communication in many aspects, promoted the formation of people's rich social relations and had a great impact on people's development. The "virtual reality" environment created by information technology provides an excellent place for people to practice their roles, in which people can "change roles", assume that they are different roles, experience the needs and emotions of different roles, and practice their roles according to their own understanding of the role norms in the network society. In a word, informatization has brought about changes in communication mode, survival mode and education mode, provided the possibility of further development for the improvement of personality and ability, and also brought about a major change in the paradigm of talent evaluation theory and method.

Informatization is manifested in the field of human resource management and is the "e" of HR process. The development and application of all kinds of talent evaluation software is one of the highlights. These evaluation tools based on artificial intelligence, information technology, advanced programming language and multimedia technology not only "e" the general evaluation methods, but also build them on a powerful technical platform, greatly improving the application scope of the general talent evaluation methods. Frankly speaking, man-machine dialogue is the realization of universal talent evaluation theory and method in computers and networks, which has important practical significance: when information technology replaces human operation, the evaluation task or process is automated; When information technology expands people's work, it makes the evaluation task or process informationized; When information technology is used in reorganization, it makes the evaluation task or process change fundamentally.

General talent evaluation needs to be realized under real and immersive stimulus-response conditions. However, in the actual talent evaluation, this is often beyond our power. Because in reality, many scenes can be met but not sought. For example, earthquakes, flash floods and other emergencies test the individual's emergency ability, which is unpredictable. Some scenes, even if they can be deliberately created, are expensive and threaten both sides of the test, such as the training and testing of pilots. These limitations also exist in other fields, and urge people to develop new ways to simulate scenes realistically. With the development of computer technology, this process can be popularized in virtual space. Man-machine dialogue evaluation is a response to the traditional evaluation theory and method of information challenge. Not only can the tested object gain self-awareness, but also can gain learning experience in the process of dialogue; It is not only a process of evaluation, but also a process of learning. Because man-machine dialogue has many advantages that other evaluation methods are difficult to possess, it has been widely recognized in all aspects. This can be seen from the popularity of network evaluation, which reflects the public's recognition and participation in human-computer dialogue evaluation. In recent years, many evaluation softwares have been launched in the field of talent evaluation for different evaluation objects. Psychological test, professional written test, system simulation and artificial intelligence expert system in evaluation technology have all been initially realized in man-machine dialogue. All these indicate the broad prospect of the application of man-machine dialogue in evaluation.

Second, man-machine dialogue: the promotion of universal talent evaluation methods

A common misunderstanding is that man-machine dialogue only transfers the method of talent evaluation to the computer, which seems to be no different from the application of general talent evaluation methods (based on the interaction between people). Indeed, in the past, the application of various talent evaluation methods was mostly based on direct interaction between people. Even the seemingly indirect interactive evaluation, such as paper-and-pencil evaluation, is only a "stimulus-response" evaluation with the help of one or several intermediaries. From the philosophical point of view, especially Marx's broad classification of human social development, the direct interaction basis of this evaluation has its limitations. Because the fundamental feature of traditional society is based on the dependence between people, in this society, I can really reflect myself in the mirror more accurately through the mirror of direct interaction between people. However, under the market economy conditions in the information age, the essential characteristics of society are transformed into people's dependence on things. In this society, the relationship between people is based on the interaction of things-people-things to some extent. On the surface, it reflects the relationship between things, but in fact it reflects the social relationship between people more deeply. This is the negation of negation. In the information society, this mirror is "virtual reality" (Qian Xuesen called it "spiritual mirror"). The change of evaluation means is the rise of man-machine dialogue evaluation method based on network and information. However, behind the man-machine dialogue and man-machine relationship, the social relationship between people is still working, but this social relationship is expressed through virtual space.

No matter from the external aspects such as design process, operating procedures, data collection and processing, interpretation of evaluation results and discourse system, or from the theoretical and technical basis of evaluation, it is not difficult to see that, compared with the general talent evaluation methods, the "man-machine dialogue" method is not a specific evaluation method, but a reproduction of many talent evaluation methods on the computer. It can realize the synthesis of general talent evaluation methods and provide a broader stage for them. This reappearance is not based on pure technology, but on grasping the new basic trend of social communication in the information age and changing the advantages and disadvantages of general evaluation methods. General evaluation methods, such as interview and document processing, have their own relative advantages and applicable fields, but they also have the same shortcomings. Man-machine dialogue can be improved in all aspects:

First, economic constraints. Some evaluation methods are considered suitable for the evaluation of "senior talents", not only because they are scientific, elegant and reliable, but also because of a very realistic consideration: economic reasons. For example, the evaluation center technology combines the evaluation methods such as group discussion, document processing and structured interview. The whole implementation process consumes a lot of manpower, material resources and time, and has high requirements for testers. In short, economic factors determine the difficulty of implementing many talent evaluation methods and limit their scope of application. Man-machine dialogue can solve similar difficulties to some extent. Man-machine dialogue evaluation system can simulate many evaluation methods in reality because of its powerful data processing and scenario simulation capabilities, so that evaluation can be carried out on a large scale anytime and anywhere. The scale effect first brings economic cost savings and opportunity cost reduction.

Second, the interference of human factors. The objectivity and scientificity of talent evaluation depend on the proper selection and correct application of evaluation methods to a certain extent. But some people think that interference has nothing to do with the method itself, but only with the technical platform on which the method runs. For example, the pursuit of reliability and validity in talent evaluation has been going on. How to realize value neutrality, objective process and fair procedure in evaluation to the maximum extent is often a headache. A typical example is the college entrance examination. In the meantime, there are not many institutional measures, but they are often unsatisfactory. Since the implementation of the network remote admission and online admission publicity system, those stories that have broken my heart have rarely happened. Man-machine dialogue has improved the realm of many evaluation methods, which is shown here. We can fully consider the interference of human factors in advance, take technical measures to shield and eliminate it, and "force" the whole evaluation concept and system to change; Another key benefit is that even if the interference of human factors cannot be completely ruled out, at least the reliability and validity of the evaluation and how to continuously improve it can be clearly defined. In short, in the face of inevitable fallacies, man-machine dialogue can make us understand more easily: how big the error is, how big the probability of making mistakes is, and where are the key measures to avoid mistakes. All these have troubled the general talent evaluation methods to varying degrees.

Third, the embarrassment of workshop-style management. General talent evaluation, even if it is a combination of various evaluation methods, often can not get rid of the workshop-style atmosphere. This is because: fundamentally speaking, the general talent evaluation method has to be based on experience and intuitive judgment. Experience and intuition are important, but not enough. For example, an old farmer tastes which corn is native corn and which corn is hybrid corn through his mouth, and a researcher seems to have similar results in identifying the two kinds of corn through genetic testing. They all identified the varieties of corn, and the former may be faster and more direct. But the technology content is different. Because these evaluations basically can't get accurate and detailed data, let alone establish a database of evaluation research. This determines that the implementation of traditional evaluation cannot be gradual, and the combination of evaluation methods is often very different, so it is difficult to obtain overall synergy. Performance in the field of evaluation, that is, the evaluation of each industry is fragmented, and the old and the dead are not in contact; It is difficult to establish the norms of all kinds of talent evaluation; Valuable data obtained from various evaluations are idle and cannot be enjoyed; Lack of quantitative comparison of talent evaluation in different genders, ages, regions and industries. Without the support of a strong talent evaluation data platform, workshop-style evaluation is difficult to connect with international standards and cannot communicate and compete with international standards. For example, several well-known evaluation and consulting institutions in the west are invincible, largely because their practices are supported by long-term accumulated data platforms. On the other hand, in China, sub-reviews, single reviews and game reviews abound. This situation restricts the sustainable development of evaluation. The application of "man-machine dialogue" evaluation method may gradually get rid of this sad situation. Because it can continuously collect evaluation data and form a powerful evaluation database; It can realize the exchange of online evaluation data, provide data support for evaluation, and make evaluation change from intermittent process to continuous process. In a word, man-machine dialogue is a prelude to the era of digital evaluation.

Fourth, deal with the confusion of the times. The arrival of the information society provides a new perspective and new thinking for talent evaluation, and also brings new problems and new thoughts on evaluation methods. For example, the popularity of online office, the establishment of digital community and the construction of information-based government have had a great impact on people's ideas. It requires new development in evaluation methods, contents and forms. Man-machine dialogue provides a new way to meet this challenge. General talent assessment methods, especially psychological tests, are widely used in man-machine dialogue. This is not only because the scale of psychological test is easy to realize man-machine dialogue, but also because psychological changes, as a reflection of the rapid changes in modern society, are often more deeply reflected in people themselves. In the transitional society, the most profound change is the change of people. To grasp the essence of talent evaluation, we must see that the essence of people is people themselves. Talent evaluation in the information society needs to adapt to the profound changes of people themselves, so it must respond to the dynamic, retrospective, humanized, service, convenience, conciseness, long-term and guidance of evaluation. And these new demands are just what ordinary evaluation methods can't cope with. The development of man-machine dialogue and its technology provides some answers to these questions in the field of evaluation. Dynamic network communication, interactive man-machine dialogue, increasingly humanized man-machine interface, and increasingly powerful computer data processing and simulation capabilities make the general evaluation method gain new connotations and new forms in man-machine dialogue. For example, the official document processing test of leadership quality evaluation should also be realized online. Because many government agencies have implemented paperless office, it is different from traditional document processing. In the long run, this is not only the ability to change the evaluation content, but also the way of man-machine dialogue.

Third, man-machine dialogue: theoretical basis and characteristics

Man-machine dialogue is not only a simple technical expression, but also has its own theoretical basis in its development. Recognizing this is conducive to a correct understanding of the relative advantages and possible disadvantages of man-machine dialogue evaluation. In addition to the theoretical assumptions of general talent assessment methods, the theoretical basis of man-machine dialogue also includes:

Theory of cognitive psychology. Cognitive psychology originated from the rebellion of behavioral psychology, which brought the new development of psychological test theory. It combines psychology with computer, and produces a brand-new field of artificial intelligence. It uses the viewpoint of information processing to explain people's psychological process, which is a big step forward than the hypothesis of behavioral psychology, and makes the hypothesis of evaluation change from "stimulus" to? Reactive "into" production "; It borrows the theories of information theory, cybernetics, computer, bionics and other emerging disciplines, and puts talent evaluation in a more advanced and scientific field; It makes the psychological test return to the forgotten tradition of semantic interpretation, re-examine the traditional evaluation point of view, and "resurrect" it on the computer. The application of cognitive psychology theory in the evaluation of man-machine dialogue broadens the horizon of evaluation and updates the way of evaluation, which is a sublation of positivism evaluation paradigm. All these make man-machine dialogue turn as an evaluation method and have a solid theoretical foundation.

Non-intelligence theory. The new trend of talent evaluation development is to pay more and more attention to the evaluation of non-intelligence factors. The concept of practical intelligence is an obvious example. According to Robert Jeffrey Sternberg's three intelligence theories, there are three types of human intelligence, namely: ① Synthetic intelligence. That is, the ability to interpret information in a structured and well-defined context; ② Empirical intelligence. That is, the ability to look at problems from different angles, explain information from changing scenarios and solve practical problems; ③ Situational intelligence. That is, the ability to adapt to changing environment and the ability to manipulate the negotiation system. Non-intelligence theory provides theoretical support for the application of man-machine dialogue: general evaluation methods can only assume scenarios on simple words and pictures, and are limited to the evaluation of synthetic intelligence and empirical intelligence, while man-machine dialogue makes it possible to reproduce and test non-intelligence factors, making talent evaluation more extensive, hierarchical and targeted.

Item response theory. In order to achieve the best evaluation design, the practice of project response theory is different from the classical measurement theory. It can estimate the estimated standard error and information function to be tested within the tested ability range in advance, and then select the questions of the tested ability level from the question bank. This lays a theoretical foundation for the personalization of man-machine dialogue. Computer automatically controls the test accuracy of evaluation instead of manual work, balances the content of questions, adjusts the test length, and realizes automatic topic selection; Man-machine dialogue can quickly achieve the optimization goal of evaluation: the shortest test length; The test information function is the biggest; The test deviation is the smallest; The test time is the shortest; The test reliability is the highest; The distribution of the obtained score has the best fitting degree with the distribution of the target score. Many advantages of man-machine dialogue are based on item response theory. For example, computerized adaptive evaluation can be realized, so that the presentation of the following evaluation questions depends on the answers of the previous questions at any time (the traditional adaptive testing method is to change the starting point and the end point, which is extremely clumsy); Different norms can be established according to different evaluation objects, and so on. Item response theory makes man-machine dialogue more scientific and practical.

Human-computer interaction theory. The subject and object of evaluation in man-machine dialogue can learn from each other, which greatly benefits from man-machine interaction. An important feature of man-machine dialogue is employing people? The immersion of human-computer interaction, according to the response of the evaluation object, constantly learn and choose different simulation scenarios, and apply the test items in a targeted manner. Simulated, real-time and active computer three-dimensional dynamic images and sounds make the test a dynamic interaction, giving the evaluation object a realistic effect; Evaluation can realize self-help, real-time monitoring, adapt to the requirements of self-help open training classroom, choose the time for self-help training, eliminate the psychological burden of the evaluation object and improve the evaluation quality. The theory of human-computer interaction dispels the prejudice that only direct face-to-face evaluation can be reliable and credible, and provides a new basis and defense for the legitimacy of human-computer dialogue.

The technical characteristics of man-machine dialogue depend on the characteristics that determine evaluation to some extent. But the characteristics of man-machine dialogue go far beyond the scope of pure technology. With the rapid development of information technology, man-machine dialogue is constantly changing. The progress of evaluation theory and technology is developing in the direction of scene, with strong realistic orientation and practical orientation, trying to evaluate people's appearance in real environment. The characteristics of man-machine dialogue are also around this direction. Specifically, the characteristics of the man-machine dialogue evaluation model are as follows:

Visualization of (1) evaluation. The application of expert artificial intelligence device and simulation system makes the man-machine dialogue of evaluation vivid and intuitive, improves the interest and enthusiasm of talents to participate in evaluation and examination, and overcomes the influence brought by the difference of evaluation subjects. (2) Simplicity of evaluation. Man-machine dialogue evaluation can extract the simplest evaluation items and cover the evaluation elements to the maximum extent. In particular, the simulation device can scientifically extract and summarize complex evaluation elements and projects, making them simple and scientific, easier to operate and evaluate, and realizing the unification of the complexity of evaluation design and the simplicity of evaluation results. (3) Safety of evaluation. The scale and items of man-machine conversation evaluation are based on a large number of question banks, and the evaluation data has hierarchical confidentiality; The content of polygraph can be added in the evaluation, and the scope of application and matters needing attention of the evaluation results should be indicated; (4) Scientific evaluation. Man-machine dialogue evaluation can test the reliability and effectiveness of evaluation at any time, ensure the scientificity and accuracy of evaluation data, eliminate human factors, make the evaluation results true and reliable, open, fair and just, and improve the quality of evaluation and diagnosis. A large number of data obtained through man-machine dialogue lay the foundation for the continuous improvement of evaluation in the future and maintain the continuity of evaluation. (5) The economy of evaluation. Man-machine dialogue evaluation can save a lot of manpower and material resources, save training time and reduce costs. (6) Effectiveness of evaluation. The evaluation of man-machine conversation can realize the learning of both sides in man-machine interaction. It can not only integrate evaluation and evaluation, but also print the evaluation results immediately after evaluation. It can also realize the separation of evaluation and supervision, and hand over the evaluation results to experts to ensure the reasonable, legal and applicable evaluation results.

Fourth, man-machine dialogue: problems that should be paid attention to

Man-machine dialogue is not everything. The imperfect theoretical basis and technical means on which man-machine dialogue itself is based, the limitation of evaluation of man-machine dialogue, the difference between man-machine simulation and practice, and the relationship between man and machine? Many factors, such as human factors, still dominate the computer system, which determines that there are still many problems in man-machine dialogue:

First, informatization is a double-edged sword. First of all, although computer adaptive testing simplifies the evaluation process, lightens the focus and reduces the variability, it also turns more and more work into digital and symbolic selection and marking. Evaluation is more and more becoming the explanation and explanation of symbols such as charts, and the world of evaluation is more and more restricted and melted by these materialized digital symbols. The interaction between the two sides of evaluation retreats behind these digital symbols, which will easily lead to different degrees of misunderstanding of the language ability, creative thinking ability and even moral judgment and choice ability of the evaluation subject. In addition, in the free electronic cyberspace, personality is inconsistent with real life, while in the same electronic cyberspace, this personality's "self-identity" may be broken, split and distorted in the evaluation of man-machine dialogue, which may lead to multiple personalities and personality splits in the evaluation and lead to disorder in the evaluation.

Second, the evaluation of information flooding and crowding out. Man-machine dialogue can produce a lot of data. If you can't handle these data skillfully and professionally, there will often be information confusion. In the man-machine dialogue, too much evaluation information inundates the evaluation subject and object, and the evaluation method dispels the evaluation purpose. In Rothzak's words, it is: "information is everywhere, but there is no thinking mind ... too much information will crowd out ideas and make people dazzled and confused in the face of empty and scattered facts." At this time, information "fetishism" and computer "fetishism" came into being, and information, as an alien force of the subject, was opposed to the subject. Information-based evaluation is easy to make people indulge in evaluation technology, ignore and avoid the feelings and communication between the two sides in the life world, confuse "man-machine relationship" and "interpersonal relationship", and confuse two different inspection standards of "real practice" and "virtual practice", resulting in the antinomy of evaluation purposes and means.

In view of some misunderstandings in the application of man-machine dialogue, it is urgent to correct the attitude towards man-machine dialogue:

First, we should pay attention to the possible problems in the evaluation software itself. The evaluation of man-machine dialogue is inseparable from the application of test software. Whether the software design is mature or not determines the maturity of man-machine dialogue. At present, many evaluation software, especially psychological evaluation software, are imported from the west, and their theories and methods are mostly centered on western psychology. The problem is that even the famous American psychologist G Murphy has to admit that most of the problems in western psychology exist only in western history? Western geographical, economic, military and scientific background? Within the scope of meaningful questions. According to statistics, there are as many as 15000 kinds of western talent evaluation software. If it is applied directly without modification on the basis of localization, problems will often arise. Even the self-developed evaluation software, because of the test, often appears problems such as bad use and inconsistent format; As for the problems in design ideas, they will be indirectly expressed in man-machine dialogue. According to the survey, most of the existing evaluation software developed by various units have not been fully realized in terms of ease of use, such as instant help, personal answering system, group data processing system, data screening system, etc., and the evaluation report is imperfect, the functional modules are incomplete, and the confidentiality is not strong, which needs to be solved urgently.

Second, we should prevent the tendency of being overly superstitious about evaluation software. Any evaluation software is based on a certain theoretical framework and specific technology, and it cannot be perfect. The so-called "wise men are fascinated by what they hear", man-machine dialogue transcends the "traditional" evaluation method, and at the same time sows the seeds of limitations. For example, in face-to-face interaction, the grasp of various details and attitude towards life are ignored; Blind trust in software procedures, ignoring the rigid thinking of the initiative of both subject and object; It has crossed the "language trap" in the paper-and-pencil exam, but it is in danger of falling into the "man-machine relationship" trap; The immersion of network simulation makes it possible for both parties to "be a guest in their dreams and make a mistake for their hometown"; Relying on the evaluation results obtained by computer, clinging to the data obtained by man-machine conversation, ignoring the common sense of evaluation. These tendencies are worthy of vigilance.

Third, we should prevent excessive commercialization orientation. Man-machine dialogue system has a long development cycle and high technical requirements. Due to various reasons, the realization rate of man-machine dialogue is very low at present, and the author or examiner can't get reasonable economic benefits, and often infringes copyright to a great extent and can't get the necessary legal protection. In this way, we have to rely too much on man-machine dialogue software and its evaluation to make profits, which will devour its scientificity and credibility, make the evaluation of man-machine dialogue quick and quick, make it difficult to form a good external environment, and guide the evaluation work to develop in a broader and deeper direction. In Marx's words, excessive commercialization tendency has brought too much profit to man-machine dialogue, but it has also cast too much shame on it. How to balance the professionalism, usability, standardization and commercialization of man-machine dialogue needs further study.

;