Abstract
The logic of technical governance goes as follows: the knowledge of society can be obtained by the state via technology, and thus social problems are identified and solutions are found. Questions have been raised about whether technical governance would ever work. For many the answer is negative and doubtful. However, one defense remains: technical governance fails not because the idea is inherently flawed but because the technology is not good. Would technical governance succeed with better methodology and more technical rigor? In order to challenge this defense, this paper examines the operation of opinion polls—a form of technical governance supported by rigorous quantitative social research methodology—run by a sub-district government in the city ‘S’. In particular, this paper asks whether it is possible for a government-run poll to reflect manipulated public opinion, despite the strictest compliance with quantitative polling methodology. The finding of this paper gives an affirmative answer. It argues that on the surface, polls are statistical surveys, but in actuality they are a political process controlled by the government despite their compliance with all statistical requirements. The power structure of the local government determines the questionnaire items, their multiple-choice answers (the screening, compressing, and quantifying of social scenarios), and the final make-up of the public opinion index. The rigorousness of methodology does not guarantee the authenticity of ‘public opinion’ in final poll figures. More likely, the outcome is controlled by those who organize polls. Hence, quantifiable technical governance presents a contradiction: the state manufactures biased public opinions precisely when it is looking for unbiased public opinions. In the end, the government constructs an image of society that is its own reflection.
The two branches of government that Boyle and Hobbes develop, each on his own side, possess authority only if they are clearly separated: Hobbes’s State is impotent without science and technology; Boyle’s science is impotent without a precise delimitation of the religious, political and scientific spheres. Latour (1993: 27–28)
Introduction
Technical governance has been a ‘haunted’ subject in almost every field of sociological studies. It cannot be ignored, yet it is difficult to understand. Ever since Jingdong Qu et al. (2009) pronounced their judgment that ‘state–society’ relations are transforming into relations with more technical governance, more attention has been drawn to technical governance, and a trace of it can be found in every theme of social science study. Technical governance provides methodological confidence in theories like fine governance, cybernetic governance, and data governance; it leaves traces and clues in theoretical observation of project-based governance (Qu, 2012), mobilized governance (XG Zhou, 2012), and administrative subcontract (LA Zhou, 2014); it came into academic attention for its applications in E-governance (XC Huang, 2010), targeted poverty alleviation (Li, 2017; Wang, 2016), land resources management (Du, 2017), and the management and monitoring of migrant populations (Han, 2017). What, then, is technical governance? Is it a concept, a kind of social phenomenon, a trend, or just a buzzy terminology? The quest for the answer to these questions is mingled with Foucault’s governance theory, the Frankford class, political conservatism, and even technocracy, which make it more difficult to understand technical governance.
Meanwhile, the dilemma of social governance is yearning for new and better solutions. With grassroots governance, for example, one dilemma concerns how personal preference gets reflected in public preference. It is difficult for public opinion to be expressed correctly and in timely fashion and be responded to with effective feedback. This fundamental fact is conveyed in those difficult problems of governance, such as the dual differentiation of residents (Min, 2009); the middle classes’ lack of interest in participation in community affairs (Xiong, 2012); the increasingly fierce community contest (DY Huang 2013; XX Huang, 2012); and the fraud and misconduct of some neighborhood committees (Yang and Yu, 2012). The mission of technical governance is to identity and solve these upcoming problems so that the whole picture of our society can be explicitly presented to the state. Is this mission possible? Technical governance has been criticized for its generalization of society and its imposing power on individuals (Foucault, 1988: 403–417). It attempts to ‘relieve institutional and structural problems by administrative techniques', yet gets stuck in the existing framework of governance (Huang and Ji, 2016) and foresees its own failure (Scott, 1998). In fact, no theoretical resources can provide answers to these questions: are the techniques destined to be manipulated or are they flawed in their logic? Are the techniques doomed to failure or are they insufficiently developed? To answer them, this paper presents an extreme case 1 study in hopes of maintaining technical rigor and maturity as well as independence of procedure; we used systematic mathematics, especially principles of statistics, as the methodology of public opinion investigation so as to explore the logic and destiny of technical governance, which would be perceived as an empirical issue. In other words, by uncovering the mysterious process of public opinion formation, we tried to understand why the opinion poll carried out by grassroots governments reflected low public credibility in most cases and was even questioned as being a trusted source for reflecting what people really think. If the opinion poll was a failure, then why?
Theory and literature review
Mathematical management in grassroots governments
Ray Huang believed that mathematical management, a Weber-style proposition, is the solution to problems in social governance in China. The ‘calculable legal system’ and the ‘administration in terms of formal rules’ are signs that the state’s social relations are becoming rationalized (Weber, 2001: xxxviii). Imperial governance in ancient China mainly relied on the administrative art of Wenshi (civil servants) which was based on their ethics and philosophical thinking after they mastered the Confucian classics. Coin-making, measuring, taxation, land registration, military administration, residents’ registration methods, and other metrology techniques were either backward or too difficult to implement, so the administration management and judicial authorities were sloppy and weak (Weber, 1995: 47–59, 84–87, 112, 117–142). Hence R Huang’s (1997a: 55) judgment that ‘mathematical management can work in modern western countries, but not with Chinese traditional bureaucratic governments’. The following paragraph explains this: Thus, Chinese society cannot be managed mathematically. For a long time, law was replaced by morality. When conflicts happened, local governors were all required to stifle them. On the one hand, the work load in local administration was reduced; on the other hand, equal and free exchanges of economic elements in lower-class society were blocked. (Only the most primitive exchange could be acknowledged by the public; the division of labor and complicated contract relations were not guaranteed under such social circumstances.) (R Huang, 1997b: 25)
However, T’ung-tsu Ch’u (2011) showed us different scenarios in his work, Local Government in China under the Ch'ing. At least in the Qing Dynasty, mathematical management was not rare as an administrative technique, and it was even popular in grassroots governance. Calendar, map, and name list constructed a space-time coordinate for the empire and yielded information and data; techniques like systems of bureaucratic establishment, salary, punishment and penalty, and promotion standard supported each other and formed a behavior space for grassroots officials. Taking the most concerned issue, taxation, for example, Fu Yi Quan-shu (The Complete Books of Taxes and Labor Services), Qing Hui Dian (Collected Statues of the Qing Dynasty), and Hubu Zeli (Regulations by the Board of Revenue) all made detailed specifications on ratio and quota for collecting di-ding-yin (land taxes) and cao-liang (tribute grain) according to the estimated population and land scale in every province of the country. The specifications on tax collection included: the revenue allocation proportion between the central government and local governments; the agenda of tax collection and the principle of installment payment of taxes; the appraisal method of tribute money and grain; officials’ performance evaluation method considering their tax collection amount; and solutions to overdue payments, etc. (Ch’u, 2011: 203–225). Civilians have the right to know. ‘Officials shall carve the category of tax and tax rate on the monument for the public’ (Ch’u, 2011: 210). In tax collection, ‘a bi-bu (brochure of tax procuration) was made recording every tax payer’s tax payable, tax paid-in, and tax owed … the most indebted payer was listed on the top’ (Ch’u, 2011: 217). The data base for the brochure was made up of tablets such as the xian-dan (notice for taxpayers to pay a tax within a specified time), the transaction list of tax collection, the tax receipt, the bi-zhao (reference for tax payment), and/or xiao-zhao (cancelation of tax payment) (Ch’u, 2011: 217). From the above we can see that with the techniques of land measurement, identity recognition, categorization, sequencing, and accountancy, taxes were collected in a digital and quantified manner with logical rigor and order.
If we observe this process in a micro political scenario, however, we find that techniques were misused or distorted in many ways. When tribute grains were collected, the actual amount the taxpayer paid was about 250% of the statutory amount; if taxpayers wanted to replace tribute grains with cash, they had to pay 300–400% of the statutory amount in Jiangsu’s case (Ch’u, 2011: 222–223). How could it be possible that the tax burden was so heavy under such a rigid taxation system? The reason is that every level of taxation authority tried to ‘take an advantage’. The dou-ji (officer in charge of weighing grain) would shake the containers to pack more grain, or extort more from the taxpayers by claiming that their grains did not meet the acceptance criteria; some low-ranking officers paid the tribute grains for taxpayers without their consent, then demanded that they repay the grains with increased interest; some shu-li (officers in charge of tribute grain documentation) would procrastinate about the date of tax payment, which increased the cost of looking after the grains and then asked taxpayers for bribes; some officers demanded that taxpayers buy ‘refined’ grains as tribute grain at certain grain shops and shared the profits with those shops; dou-ji and shu-li collaborated with subordinate officers of County Head to cover up their crimes, etc. (Ch’u, 2011: 80–81, 105–106, 138).
Strict rules could not prevent shocking corruption from happening. Throughout the Qing Dynasty, local administrative regulations were continually revised and maintained consistency, but they were all ‘technical and procedural without any substantive significance’ (Ch’u, 2011: 315). Why did technical and procedural rules not work? Why did mathematical management blur the social image rather than present a clear social image? The above questions led us to ask about the nature of technical governance.
Governance, techniques, and technical governance
Just as the word ‘governance’ became a buzzword (Fukuyama, 2013; Offe, 2009), the term ‘technical governance’ can be found in many academic literatures, and sometimes even in rhetoric. What exactly is technical governance? The answer remains to be determined. The original meaning of ‘governance’ was to ‘steer’, ‘manipulate’, and ‘conduct’ (Jessop, 1998), with an implied meaning of techniques. Arguing with Syracuse about justice, Socrates once said: ‘We call him steersman, not because he steers the ship, but because he has his own technical skill, and he is able to manage sailors’ (Plato, 1986: 23). The steersman manages to ‘govern’ the ship by means of his sailing skills and knowledge. Likewise with shepherds, doctors, architects, and musicians who ‘govern’ their sheep, their patients, their architectures, and their instruments by using their particular techniques and skills (Plato, 1986: 28–30, 35). ‘Governance’ implies techniques which are used to govern. The purpose of techniques is to make the governed happy, as when shepherds take care of sheep: count them, find pastures for them, and watch out for weather (Foucault, 2016a: 320). As the modern western countries came into being, the meaning of the governor’s role underwent a series of evolutionary processes, from sheriff to herdsman, religious pastor to king, and finally settled on the concept of ‘governing the country by monarch’ (Foucault, 2010: 104–112).
If ‘governance’ implies ‘techniques’, then what does ‘techniques’ mean? According to Plato's The Symposium, ‘techniques’ are used to achieve ‘presence’: ‘Every occasion for whatever passes over and goes forward into presence from that which is not presence, is poiesis, is bringing-forth’ (Symposium, 205b, quoted in Heidegger, 1977: 10). Techniques turn the state of things from ‘being potential’ to ‘being actual’. According to Aristotle (1959: 62) ‘Potential is prior to actualization, while not every potential thing has to be actual.’ ‘Being potential’ means infinite possibilities. Once something becomes ‘actual’, its possibilities are narrowed down and its meaning is presented (Aristotle, 1959: 106–107, 176, 182). Therefore, the manner in which techniques present the meaning of things is a process of simplification. 2
When the word ‘governance’ became an entitled noun used for society by the state, technical governance could be viewed as a kind of simplification of social complexity by the state. The techniques of language and concept such as naming and categorizing would simplify complicated and vague social contexts into crime, income gap, plague, invasion, and other upcoming issues to be solved by the state. Public security, taxation, public health, national defense, legal system, and other institutional techniques would simplify the possibilities that a person can achieve, by creating pairs of identities such as law-abider/law-breaker, tax-payer/tax-evader, the infected/the healthy, soldier/civilian, etc. Quantitative techniques such as counting, accounting, statistics, and geometry generalize society and people into data of household registration, population, land area, tax, etc. Therefore, the process by which technical governance simplifies the complexity of society can be divided into three stages: social context, problem, and data. Under technical governance, from social context 3 to problem and data, the dimensions of social and human possibilities are simplified step by step, and the amount of information is compressed gradually until it becomes a figure; that is, a dot on the open quantitative ruler in some dimension.
Undoubtedly, the logic of technical governance has confronted many doubts and questions. The first question is about the social image distortion caused by simplification. Is the simplified figure able to restore the original social context? Does the deleted information cause social image distortion? When observing society, the state has to rely on a pair of glasses—that is, techniques—to conduct the procedures of simplification, such as identifying social problems, defining target groups, measuring the demand, and analyzing the cost and revenue. During this process, the vague information of society is categorized or deleted by techniques for being placed outside of the order (Oakeshott, 1985: 212–219). Here is an example. For the purpose of increasing economic returns, German scientific forestry organizations have simplified their complex forests into single tree species of tall coniferous forests like Norwegian spruce and have developed a sophisticated technical process based on the analysis of cost and revenue on seed selection, tree growth, and deforestation. Other factors are excluded or assumed to remain unchanged (Scott, 1998: 12). However, the data generated by technical governance could only be used most effectively when they are used in a social context. After the data are generated, they cannot be restored to the state when they are not compressed due to the lack of dimensions and the reduction of information. The technical procedures of scientific forestry were ‘revenged’ for not restoring the original state: the soil became barren because shrubs and dead trees were cleaned up; trees were more easily blown down by the wind because trees were planted separately by age; and the lack of other trees caused pests to ravage them (Scott, 1998: 18–22). Similar examples could be found in Soviet farms, Tanzanian agricultural projects, the city of Brasilia, etc. (Scott, 1998: 193, 223, 103); in New York, Chicago, and other major cities in the US, after converting some sidewalks, old streets, and old buildings in urban planning, social possibilities and inclusiveness were found to be constrained (Jacobs, 2005: 244–282). The fundamental reason why simplification causes social distortion is that ‘in all worlds, the political world may be the last one that would withstand the rationalist scrutiny. Politics is always rooted in tradition, coincidence and transience’ (Oakeshott, 1962: 3); the social world also consists of all things that are incommensurable, including things that are ‘undefinable, incoherent, inconsistent, uncompromisable, illogical, irrational, ambiguous, vague, indecisive, and contradictory’ (Bauman, 2013: 11). Hayek questioned the feasibility and effectiveness of technical governance and concluded that the national plan would definitely fail because of the inability to collect information on social tendencies (Hayek, 1945, 1997: 48–58). The most reliable social monitoring technique is price. As a kind of up-to-date social public knowledge, price is stable and dynamic. Other social monitoring techniques such as bureaucracy and large-scale economic model calculations may lead to social distortion caused by information distortion and omission, and limitations in calculating capacity (Hayek, 1945).
The second question is about the power that hides in the process of simplification because it limits the possibilities of society and people. Through the establishment of the concept of unity and command hierarchy through scientific knowledge, rationality is enlightened to achieve the control of the possibility of the world from the single power center of reason (Horkheimer and Adorno, 2006: 10). The logic of governance is that the state sets up the form/frame with techniques to shape people into the way it needs them to be. The operating principle of governance consists of a whole set of systems, procedures, calculations, and analyses. Governance is a series of complicated operations of power with political economy as its knowledge resource and security configuration as its tool (Foucault, 2010: 45–64; 2016a, 2016b: 251–305). In China, a series of techniques are widely used in local governance such as urbanization, community building, eco-city construction, religion, and sexual health (Jeffreys, 2009). The logic behind this is three-layered: the first layer of logic is the self-development of the capability of governing a modern country. The ‘Great Leap Forward’, the urban population regulation movement in the 1960s, and the movement of Chinese educated urban youth to work in the countryside and mountain areas all demonstrate that the overall planning, procedures, and techniques of population management have gradually improved (P Zheng, 2014); the second layer of logic is the doctrine of economic development. Economic development drives decisions about whether middle and primary schools are to be closed or merged. Basically speaking, economic development rationally regulates and allocates almost every aspect of people’s migration, settlement, work, and living (Ye, 2017); the third layer of logic is the central government’s control over local governments. ‘The central government has gradually established a national land information system and become increasingly reliant on visualized technical knowledge to monitor local governments’ (Du, 2017). In addition, politics has also penetrated into the process of technical design: the height of a flyover restricts the passage of some oversize vehicles, which is quite discriminative and socially divisive; the broad design of Paris Avenue was to prevent street fighting like the French Revolution; the application of a new type of harvester was to dissolve a labor union (Winner, 1980). In short, under the simplified effect of techniques, relations between people and relations between the state and society have been ‘materialized’ and have lost their ‘mystery’ and become manipulable and controllable (Marcuse, 2014: 139–143).
The third question confronting the logic of technical governance is whether its over-simplifying process will inevitably go to extremes and create new complexities. In China, the rule of bureaucratic operation is loyalty to the superior, not Weber’s standard definition implying that workers should have specialized knowledge and be independent so that they can counter the dual pressures from monarchs/political ruling groups and civilians (Weber, 2010: 327–328, 330–332). China has a particular form of political response in that the public’s calls will only be valued if they are heard by the superior level, and only when the superior level attaches importance to the issue can it be attended and responded to by the lower level (Wen and Huang, 2016). Therefore, the principle of simplification of China’s technical governance can only work when it is relatively independent of the existing system; for example, in the form of projects, teams, and offices. The book Soulstealers: The Chinese Sorcery Scare of 1768 relates a story in which Emperor Hongli of the middle Qing Dynasty realized the severity of the sorcery scare case, then immediately called off routine bureaucratic procedures and initiated direct correspondence with local officials, reducing levels of information transmission (Kuhn, 2014: 161–163). The ‘secret report channel’ by which a confidential memo would be commented on by the emperor with a red-inked brush was an institutional innovation of the emperor. It eliminated routine bureaucratic procedures, a work reporting system that consisted of routine channels and secret channels, and avoided the layer-by-layer filtering and processing of information so that the emperor could obtain first-hand and controllable social information directly. Many governance practices in China were successfully conducted by setting up a special fund and working independently without constraints from the bureaucratic system (FZ Zhou, 2012), as in the following examples: technical procedures of the petition system, including registration, transfer, inspection, reminders, and filing (Feng, 2012); a series of rational procedures of the project system, including project establishment, declaration, review, supervision, assessment, acceptance, evaluation, and rewards and punishments (Qu, 2012; XG Zhou, 2012); and targeted poverty alleviation work, including a cadre system in poverty-stricken villages, archives from poverty-stricken families, and records of poverty alleviation work (Li, 2017; Wang, 2016). However, the simplification procedure of technical governance complicated the relations between state and society in practice. First, in the interaction with government organizations, techniques remained in reserve and became part of the governance system. For example, the Network Office, the Poverty Alleviation Office, and the Letters and Calls Office, which were set up at all levels, became parts of the bureaucratic system. Second, all levels of the bureaucracy would implant their wills during technical governance. Under the operating principle of ‘the superior assigning, the subordinate subcontracting, and the grassroots executing’, all levels of government involved in the project system would ‘slightly modify the centralized framework and bureaucratic logic, and add more intent and self-interest so as to acquire more power of their own’ (Zhe and Chen, 2011); they even ‘borrowed chickens for eggs’, trying to integrate and bond their own development strategies with projects and to transform those projects (XG Zhou, 2015). The corresponding authorities thus gained access to power rent-seeking due to having a monopoly of resources, information, and power, and might gradually evolve into a ‘special independent kingdom’ (Qu, 2012).
The power will inevitably be brought forth by techniques, which means the second question cannot be resolved. What, then, can we do about the technical problems in technical governance? is the first and the third question. Will the simplification of society cause social distortion? Is it caused by the insufficient development of techniques? Social simulation techniques based on computer programming and social simulation modeling by complex systems are so advanced that the Walras general equilibrium model of the Hayek era cannot compete. Because of the support of big data and artificial intelligence technology, people’s information keeps entering the database when people shop, travel, receive medical treatment, find a place to live, and surf on the internet. All the information is like puzzle pieces and will eventually form an identifiable individual ‘face’ under technical governance. When the facial expression can be accurately identified and instantly responded to, has the simplified task of technical governance not been completed?
Is the simplification of technical governance destined to be distorted? Will distortion be caused by insufficient technical rigor? After all, the project system, petition system, and targeted poverty alleviation were designed by the superiors. They were based on ‘primary’ techniques such as taxonomy, elementary algebra, and linear computing. Techniques are segmented from each other and do not form strict logic. There is much room for improvement. Is there, then, a sophisticated methodology to support a poll? On the basis of statistical principles such as the law of large numbers/central limit theorem and linear optimization, the sampling randomness, the reliability of the questionnaire design and the validity index are set with simplified standards, which together are regarded as the technical solutions for addressing issues of heteroscedasticity, collinearity, and self-correlation, ensuring the scientific authority of the opinion poll. Then, would the simplification of technical governance be distorted?
The failure of mathematical management in local governments in the Qing Dynasty reflects a division of the analytical perspectives of Ray Huang and T'ung-tsu Ch’u. If the division is left unexplained, it will remain a haze over the grand ambition of technical governance. Thus, this study sets political factors (not to mention the limitation of possibilities of society and people) and technical factors (assuming better and more rigorous technology) aside and deals only with the process of how techniques simplify social complexity to see whether techniques will make it.
Background and principle of the poll
Background
Process of public opinion survey in three stages.
Note: As a matter of fact, as a project, the whole poll includes procedures of application of financial funds, project design, and project summary. Only the implementation stage of poll is discussed in this paper.
Principle
Public opinion does not come from nowhere; rather, a kind of mechanism is required to develop, mold, and make it meaningful. When the respondents are expressing their views, they upgrade public opinion from an ambiguous and complicated ‘potential’ status to a ‘substantial’ status. The public opinions presented by polls are also simplified public opinions. The polls conducted by grassroots governments embody the simplification process of ‘social context-problems-figures’, which can be divided into three stages, namely, questionnaire design, questionnaire filling, and index system (Table 1).
1. How to identify contexts?
Context screening: the street selects representative contexts based on certain principles, namely, scenes of life as the prototypes for each question in the questionnaire.
2. How to ask questions?
Question defining: when a representative context is expressed as a question in the questionnaire, the countless possible dimensions in the context are limited to finite ones, to form a kind of narrative logic.
Question quantifying: in each question, the street may set corresponding options for residents to choose. The defined questions correspond to their selection domain, and the options are the scale values in the selection domain.
3. How to do sampling?
Determination of questionnaire respondents: the street cooperates with the neighborhood committee to select and determine the residents who shall fill in the questionnaire.
4. How to generate figures?
Figure generating: residents as voices of public opinions select the options according to the stipulation of the selection domain.
Figure totaling: the street takes community as the unit to calculate and totals up the scale values. The poll indicators stipulate the method of totaling, and an indicator serves as a scale for community ranking. Based on the ranking, the street may grade the communities. In this way, the street obtains the results of the poll.
Subjects of grassroots government in the poll.
The related subjects in polls are the street, functional departments and offices, residents, and technical staff. The street and all functional departments and offices are responsible for designing questionnaires and setting indexes; neighborhood committees are responsible for coordinating with the street in doing sampling and organizing the poll; residents are responsible for filling out the questionnaires; and technical staff, along with third-party academic institutions, provide technological guidance to ensure the scientificity of the poll.
The street and functional officials and staff are concerned about what contexts are selected into the question domain or the principle of selection; the direction of defining contexts and forming them into questions and the setting of question options; and the setting of indexes to form the ranking system and the ranking of communities. The neighborhood committees are concerned about what groups of residents should be chosen to fill out the questionnaires, how to draft the questions, and what they should do to improve the score, etc. Serving as a technical staff member, I had the direct impression that during the public opinion poll, almost all steps were strictly kept in line with scientific principles, that there was no case of overstepping the limit; however, the process was always awkward or uncomfortable. The final results of the public opinion poll turned out to create a peculiar situation in that ‘there was no counterfeit, but the credibility was low’ (except for that within the government system).
Given the interactions among all parties, the public opinion poll seemed to be a statistical process, but in fact it turned out to be a political process. In describing the public opinion poll by a grassroots government, this paper will gradually present the following question in connection with case studies: how was the public opinion ‘fabricated’ and how were the wills of all parties embedded in the formation process of public opinion? In a specific operation, I designed two contrast mechanisms: comparing the thoughts of all parties in the grassroots government and other possibilities provided by technical staff; or comparing the changes of the poll scheme in 2016 with that in 2015. On this basis, I reviewed the different designs and adjustments in poll schemes over the years and analyzed the reasons behind them.
Simplification procedures of public opinions: the case process
From ‘social contexts’ to ‘questions’
Questionnaire design: how to screen context?
A questionnaire is a collection of community contexts; that is to say, it collects public opinions and confines them within a certain range. As the executor and organizer of the public opinion poll, the street is responsible for screening representative contexts to form a questionnaire among numerous possible social contexts from 35 communities. How, then, to select a context? The goal of a public opinion poll should serve as guidance for context screening and, accordingly, construct the questionnaire. When people go to stalls at the entrance of a community every morning to eat breakfast, we call this a context. But only when there are cases of people doing business by occupying roads, food safety problems, or noise disturbances by people selling midnight snacks can such contexts become objects of a public opinion poll by the street. Therefore, the question emerges: what kind of empirical criterion influences the screening principle? At the poll mobilization meeting of the street, street Secretary J emphasized that [t]he public opinion poll we are engaged in cannot cover every detail—it is not checking household occupants or checking bank accounts, nor is it market investigation. The ‘hard nuts to crack’ which were concerned with people, as emphasized by XX document, commonly existed in all communities and were the key points in our investigation.
4
Sections and question types in questionnaire design. NPC: National People’s Congress.
The new ‘assisting the work of the street’ section had a total of 11 questions, which corresponded to the specific tasks of each functional department; that is to say, to let residents evaluate the work performances of the 2 committees of their own communities. Why did they add the new section of ‘assisting the work of the street’? Deputy Director General S, a leader in charge of polls, stated at a feasibility analysis meeting: The polls of the previous two years, I tell the truth, at first everyone (street departments) held a wait-and-see attitude. As a result, the effect was very good, and the general reflection was consistent with the actual impression. For example, annual evaluation of many departments used the ranking by the polls … Other functional departments (of our street) hope to share the data. By taking this opportunity, we can also make the polls more relevant to the actual work of the street. I think this is a good idea.
5
The major revisions on the sections include the following: first, the grassroots democracy that originally belonged to the self-governance section was shifted to the section of ‘elections by the National People’s Congress and communities’, which is the section of democratic supervision. Second, the original party building section which was focused on the evaluation of the project ‘Map of Party Building Resources’ was included into the section of self-governance and shared governance. Third, the original section of the People’s Well-being was included into the sub-section ‘TXJY’ featured projects in the section on self-governance and shared governance activities. ‘People’s Congress and other elections in the community’ became an independent section in 2016, for which one possible explanation is that the time of the poll coincided with the grassroots People’s Congress election every five years. When I was investigating the XHMJ community, I heard that some residents came to fill the questionnaires right after finishing the voting activity of grassroots People's Congress deputies. The Map of Party Building Resources and TXJY featured projects are the projects on government governance innovation that C street promotes. The street replaced the original section with them, aiming at testing the effects of promotion in each community and supporting the key points regarding the current work of the district and street. At the poll project proposal meeting, the street Secretary of Party Committee J who presided over that meeting said: Highlighting the characteristics of C street and reflecting the mores of C street are the important targets of our polls. Whenever we talk about party building, we would think it is very insubstantial. Whenever we talk about people's well-being, we would think it is very complicated. But what are the contents of community work? … Personally speaking, the Map of Party Building Resources and ‘TXJY’ (featured projects) are just the party building and people’s well-being of C street.
6
The direction of context condensing: how to ask questions?
After setting the framework of context samples, the following step is to decide how to define contexts as questions. The birth of a question means that a context is defined from a certain perspective. Taking the representative context, the election of the People’s Congress and property owners organized by communities as an example, there are many ways for residents to be associated with the context, and there are various possibilities for defining it, which can be directed to their target questions respectively. For example, consider how the following questions are covered in the final questionnaire:
‘Do you think the property owners’ organization can solve the contradiction between property owners and the property management company?’ ‘Have you attended the voting in the general election by the neighborhood committee of your community?’ ‘Do you think the grassroots NPC deputy election is fair?’ ‘Are you familiar with the candidates in the election by the property owners’ organization?’
In the final version, ‘How do you think about the organization of the election of NPC deputies and election of property owners’ organization in the community?’, the street has the power of combining the context of grassroots democracy with the evaluation of the organizing ability of the neighborhood committee in such activities. In evaluating the grassroots democratic construction in a community, the item does not ask if the NPC deputies have fulfilled their duties or if the election process is just, nor does it ask the reason for people’s indifference to elections at present, but asks only about the organizing ability of the neighborhood committee. Apparently, the organizing ability of the neighborhood committee cannot hold up the subject of democratic supervision. However, it is categorized within this subject and has become the criterion for calculating grassroots democracy indexes.
Why did the street select questions in this way? Does it conform to the scientific principle? The target of polls is to reflect residents’ evaluations of the community. When the street related polls to performance evaluation, the polls became residents’ evaluations of their neighborhood committees. The NPC members’ performance and justice in the election process are not within the scope of the work of the two committees. Moreover, the indifference to elections is a sensitive political topic, and the influence of the two committees on the election work lies only in organizing this activity, so questions can only be settled in this direction. The validity and reliability in questionnaire design require explicit targets and clear wordings of questions, and the answers to the questions should be stable and insusceptible to time in similar questions (Rasinski, 2008: 64). In the feasibility random test (which will be described later) in the early stage, we found that the questionnaire did not have the problem of ‘no opinion’ as mentioned by Kronick et al. (2002), and the interference factors of ambiguity in understanding and context effect were very few. The design of the questionnaire basically fulfilled the principle of validity and reliability.
Examples like Figure 1 can be widely found. In the aspect of public health, the questionnaire asked about the popularity of health knowledge lectures among people, rather than the conflicts between doctors and patients, as well as difficulties and high costs of getting medical service; in the aspect of migrant population management, the questionnaire asked about the problems of group renting rather than the problems of urban management law enforcement and settlement difficulties for migrant people; the questionnaire asked about the problems of domestic violence intermediation and legal publicity, rather than the issues of petitioning, house demolition, right safeguarding by house owners, and environmental protests. Apparently, the hot questions that concern the academic circle and the society having to do with grassroots governance have not been included in the questionnaire, but the questions in the questionnaire were related to them, to support the topic behind them. There are multiple directions in simplifying social contexts into questions, while the manner of simplification by the society and the academic circle is totally different from that used by the grassroots government. At the poll mobilization meeting in 2015, Secretary J said: ‘The number of questions in our questionnaire should be controlled; we should ask questions about people's vital interests, and the real questions confronted in our work.’
7
In the work of street and neighborhood committees, the ‘vital’ and ‘real’ problems may not be contradictions between doctors and patients or law enforcement by urban management, because they are not within the scope of the committees’ work. Group-oriented leasing governance is within the scope of the neighborhood committee’s responsibilities. In the ‘Operation Manual for Building TXJY Star Level Neighborhood Committee of the District’ as the performance appraisal standard in 2016, ‘actively eliminating the “6 kinds of chaos” (disorganized stalls, messy construction, disorderly hanging up, random scribbling and stacking everywhere, effectively regulating and controlling group-oriented leasing’ and ‘implementing population service and management well and knowing the basic information and dynamic situation, achieve actual effects in public defense and public treatment’ took up 3 points respectively (100-mark system). When the right to operate a poll is in the hands of the grassroots government, its jurisdiction and hierarchical relationship will be embedded in the process of simplifying social contexts into questions.
One of the representative questions in the questionnaire.
Question quantification: the power behind the scales
The calculation standard for public opinions is quantification, and therefore, the questions on the questionnaire should be converted into standard scales; that is, setting items to total up individual opinions. There are two methods of setting scales: sequencing and counting. The former scales the selection domain of the questions, simulates it into a number axis, and divides it into n equal parts, such as the five options in the famous Likert scale. The latter calculates the number of occurrences of an event; for example, asks the number of occurrences or directly sets them as multiple choice questions. However, it is difficult to apply the principle of option quantification to the real lives of communities: how can we simplify a complicated social fact into a point on a standardized scale?
Taking the community ‘TXJY’ self-governance project as an example, all communities are asked about the implementation situation of the featured TXJY project in the district. In principle, each community should have its own features in self-governance activity innovation; for example, the ‘sharing refrigerator’ of PX community, the ‘YWYS’ council meeting of WY community, and the WeChat Official Account ‘MFLShow’ of MFL community. However, not all communities could push such distinctive ‘brands’ in a short period. Over half of the communities reported daily activities such as offering help or subsidies to needy families, fire drills, knowledge lectures, sewer dredging, and solving power cut troubles with title of TXJY to the street. The street was well informed about what communities had real TXJY featured projects. As a matter of fact, in order to coordinate with the featured TXJY project promoted by the district and the star-level neighborhood committee appraisal activities, the projects in some communities were directly guided and implemented by the street. There might be various reasons for the communities that do not have ‘genuinely’ characteristic projects. Some communities were still soliciting topics and working out ideas and some ‘did not want to make any trouble’ (said by a middle-aged community secretary); however, one point that should not be overlooked is the interference of the street. If the questionnaire asks residents’ true evaluation of their community’s TXJY project, then the factor of the street could not be ignored, and some communities might score zero. Meanwhile, other than the featured TXJY project, there were many activities or practical tasks involved with people’s well-being to be completed by communities. A resident said in an interview: Our secretary, Y, is not fond of the limelight like others, we don’t understand that stuff … I know that every Wednesday afternoon, he took some people to do cleaning up and went on duty on roads. We live in old communities with complicated conditions. You can see all kinds of people. There are non-natives who are street vendors and tenants, you would hear quarrelling, throwing shit at each other next door. What a mess! Things are difficult for us. No easy task.
8
Meanwhile, the right to set the questions was delegated to each community. As far as the above-mentioned single-option question was concerned, since it could neither exclude the interference factor of they themselves convincing the public, nor undertake the risk of selecting true TXJY projects from false ones, the street let the communities themselves report the featured TXJY projects on their own. As for multi-choice questions, the street was unfamiliar with the situation of activity implementation in the communities. Although communities would report to the street before they carried out activities, the format of their report was like that of a newsletter, the quality of which is difficult to check. Two communities had their ‘self-governance council’; however, their actual operation effects may have varied greatly, and the report format and information were similar. In addition, the activity numbers and frequencies reported by communities varied greatly, with many influencing factors. The communities with better conditions or ambitious secretaries may have had more activities with higher quality. The secretary of community JX once mentioned that [w]e are now a small community, and working staff are only a few.
9
Two years ago, our community was relocated, and a lot of residents moved away. In Spring Festival and other festivals, some old neighbors missed each other and came over. How would they come here for publicity of medical treatment and public health?
10
It seems that a community’s power in setting questions can determine what kind of activities can be included into questions; as a matter of fact, it can affect the result by controlling options. Taking the multi-choice question ‘Please choose the community-organized activities that you know or have attended’ as an example, the measurement method of this question is calculating the percentage of ticked items. Since the economic status and personnel allocation of each community were different, the number of options was not specified by the street. It only required respondents to tick at least three options. For an indefinite number of options, the communities applied some tricks in setting the options: in cases in which other conditions were fixed, the fewer the option settings were, the higher the scores of the question would be. It is estimated that the average number of options the community set in the questions was 5.14, with a standard deviation of 1.60. The correlation coefficient between the score and the number of options is −0.801, and the significance of the two-tailed test is p < 0.001. The number of options set by each community directly affected the community’s score on the question; that is to say, the communities setting fewer options got higher scores. As a result, a paradox emerges: if the street stipulated the number of options as a standard, then the standard of more options would be better for communities with favorable conditions and worse for those with poor conditions; if the street did not stipulate the number of options, the best method for them was to set fewer options, but it could not screen out the communities whose performances were outstanding in the TXJY featured project.
From ‘questions’ to ‘data’
The randomness of samples is one of the most important indexes for the scientificity of polls, and thus the street takes it seriously. As far as methodology is concerned, simple random sampling can be supported by the law of large numbers and the central limit theorem of mathematical statistics, then conform to the assumption of independent and identical distribution. In actual practice, social surveys are often based on simple random sampling, taking into account region, age, work, and other structural elements that affect sample distribution; that is, using stratified sampling to reduce sampling costs and improve efficiency (Kalton, 1983: 8–47). In terms of survey methods, face-to-face surveys, telephone surveys, self-administered questionnaires, and online surveys can serve as alternatives (Donsbach and Traugott, 2008: 240–282; Groves and Lyberg, 2010).
Filling in test papers: the reason for filling in questionnaires intensively
The final sampling method used in this poll was stratified sampling based on the categorization of people in each community. The survey method was gathering people to fill in the questionnaire intensively. The manner in which the samples were gathered to fill out the questionnaires is hardly ever seen in polls in the western world. As it is often called ‘answering papers’ by residents, filling out a questionnaire is like taking a test.
11
As a matter of fact, alternative survey methods such as in-home interviews, street interviews, and telephone interviews (Tourangeau and Smith, 1996) were discussed at the feasibility analysis meeting. In addition to the advantages and disadvantages brought about by characteristics of these techniques, a more important factor is the restriction of poll projects. During a nationwide, face-to-face survey in the US in the 1980s, each respondent cost around $1000 US dollars (Groves, 1989, cited in Donsbach and Traugott, 2008: 228). Although the scale of this poll was far smaller than a nationwide poll, it still consumed a lot of money, which was paid by the street’s project fund. In response to this problem, Deputy Director S said: You said that we should knock on every door in the 35 communities; that would cause disgust, so there is no need to do so, and we don’t have enough time, energy and manpower … Every community has a meeting room to gather residents together, the ballot box, the meeting room or something, anyway, they are ready-made, the (people) of the communities are familiar, and it is more formal.12,13 After resident representatives were seated at the venue, the community secretary started to make a year-end report. Supervised and guided by technical staff, the respondents filled in the questionnaire and submitted the questionnaires. Some resident representatives (generally 8–10 people) took part in the structural interviews organized by the street and gave scores and evaluations to the working staff of the 2 committees of the community. Working staff of the street had interviews with the two committees of the community and gave them scores.
During the public opinion polls of all communities, related working staff of the street, technical staff, community cadres, and residents were all present. The procedures were clear and performed efficiently. Everything was finished intensively in half a day, and the community fulfilled its mission. Interestingly, when I asked residents how they felt about the public opinion poll being organized by the community, one resident stated: ‘We have several activities each year which are organized in the same way.’ 14 For residents, activities are frequently organized in communities; compared with the sodality of ‘Moon appreciation and the Mid-autumn Festival’ and health knowledge lectures for senior citizens, the public opinion poll differed only slightly in form.
Can randomness, then, the index that Deputy Director General S cares most about, be ensured? Technical explanation is also convincing. Filling in questionnaires intensively can ensure the same influence by survey environment, survey time, and working staff in the survey as well as control the common factors, such as context, interviewer, and season. In addition, the street also conducted some random tests on small samples (10–15 people) of the communities to test the validity and reliability of the questionnaires through street interview, and took this as a robustness test for the method of filling in the questionnaire intensively.
Control power on number of residents: politics with stratified sampling
How to classify people in sampling? The lower limit of sample number in each community was uniformly defined as 50 people. Where did these people come from? It is generally difficult to get a roll that includes all members (Kalton, 1983: 6–7), and the street could not get the real-time dynamic information of the populations of the communities, and there was no high-quality population for sampling. Even if the household data were available, the population mobility caused by demolition of houses, work changes, and further studies would have caused a large number of non-response errors (Weisberg, 2005: 159–204).
The scheme of stratified sampling in the polls was that the overall information reported by the communities to the street was used as a sampling frame for each community, and samples were determined by the street. In order to ensure randomness, the street specified the total populations of each community: first, the minimum number of people that each community reported should be 200 to 1000 in proportion to the actual number of permanent residents; second, the number of residents–-that is, the number of people from each community who finally filled out the questionnaires–-should be no less than 50; third, the street stipulated the categories of residents, which should include community party members (including in-service party members who serve both their work units and communities), building team leaders, volunteers, members of mass activity teams, property management personnel, co-construction unit personnel (including shop owners along the street), property owners’ organization members, Party representatives/NPC members/Chinese People’s Political Consultative Conference (CPPCC) members, community policemen, and other residents; finally, the name, gender, telephone number, and population category of each resident to be sampled must be complete. After communities submitted the sampled population as required, they would send it to the street for stratified sampling.
There might be some problems in delegating the power of determining sampling populations to neighborhood committees, since the neighborhood committees would choose residents for sampling based on the cost principle. As a matter of fact, it was very difficult for many neighborhood committees to acquire the materials for 200–1000 residents. Generally speaking, the materials for residents who were registered by the communities were available, and thus, residents were screened by the index ‘having association with the neighborhood committee’. As long as we investigate the individual character variables of the questionnaire, such as the age of the residents, educational background, type of occupation, etc.—for example, by making a scatter diagram—we can find skewness to varying degrees. For example, among the 1773 samples in all residential areas, the proportion of retirees reached 74.84%, and the proportion of samples living in the community for more than 10 years was as high as 81.22%. The situation of elderly retirees constituting the majority participating in community activities had not changed (Xiong, 2012). However, the basis for stratification was 11 categories of people. As long as the proportion of the 11 categories of people could meet the requirement, it could be called ‘good sampling quality’ technically.
The problem lies in the sampling population and the method of population division, but there were reasons for setting both of them. The problem with the sampling population was that the resources that the neighborhood committee could mobilize were biased. In response to this problem, the secretary of HD community said in an interview: ‘The young people are busy working, and we can hardly meet them, so how can we invite them to come here?’ A working staff member beside him added: ‘They may not come even if they have time.’ The secretary continued: The young people at present don’t know anything about (the work) of communities. The people we had frequent contact with knew well the difficulty of our work and our situation. When we invited them to fill in the papers, they basically agreed.
15
The control power of neighborhood committee was also embodied in the number of samples. Based on the principle of sample size setting, the number of residents required to fill in questionnaires should not be less than 50. Considering that some sampled residents might be absent, the street determined the sample size of stratified sampling to be 75 respondents for each community. Judging from the actual attendance at the poll in 2016, the average number of residents filling out questionnaires in each community was 51.14. Among the communities, 15 had just 50 attendees, only 5 communities had less than 50 attendees, and 3 of the 5 communities had 49 attendees. The street did not stipulate the upper limit of attendees and urged the communities to encourage more than 60 people to come, but the actual attendance in each community closely matched the minimum number of people required by the street. Moreover, the phenomenon emerged in 2015. What kind of control power did the community neighborhood committee have over the residents who filled out the questionnaires? Had the sample been ‘contaminated’ or was there impersonation? To this end, at the suggestion of the technical staff, in 2016, the street offered a resident signature book randomly in some communities to check the authenticity of the samples. At the site, I was curious about the name of a very young person on the signature book and wanted to interview him. After asking the working staff of the neighborhood committee, I found that the young man was busy and his mother had come over to fill out the questionnaire instead of her son. In this regard, street staff member F, who was responsible for the survey said: (To make the namelist match the requirement) is very difficult. It is no easy task to match the signature book (and sampling brochure). We don’t know whether there is a bottom line or standard for the error rate until now. Anyway, we should make them match … His mother should also be considered a resident in the community.
16
One explanation is that the surplus of sample population and the loopholes in population categorization offered the neighborhood committees leeway. The sample size of 75 people offered leeway of 25 people for the community cadres, so they could select residents who were more likely to ‘support work’. According to the criterion of the 11 categories of people, type overlapping of residents in the sample population is very common. A person can be a volunteer, a member of the mass activity team, and a party member. Community cadres can dispose residents with multiple identities to meet street requirements, showing larger room of operation.
From data to data totaling
After assigning values for all scales, the street could add up scale values according to all the totaling principles; then different index data, ranking, and grades of corresponding communities were obtained; and, finally, the public opinion map could be drawn. Constructing an index system is the last stage of a public opinion poll and the most important part for the street and communities. Each index was determined based on corresponding segments; for example, party building resource index, self-governance, and shared governance index, index of democracy at the grassroots level, TXJY index, and index of supporting functional offices and departments. The final calculated indexes were to be used by all functional offices and departments, and should be submitted to street leaders and even reported to the district. Moreover, they would be edited to make a ‘Diagnosis book for XX community’ to be distributed to each community.
Rating and weighting
Score proportion of each department and office in the poll.
The first technique for index calculation: calculation method
As an action to find out the situation of all communities of the street, since 2015,
17
a progress index was set in each public opinion poll, and the calculation was based on the scores of the communities in the polls over the years. As a matter of fact, there were great differences in topics, section division, and score weight for the public opinion polls, and thus, the benchmark year standard for comparing one year with the previous year makes little sense. Seen from the results, in 2015, the communities achieved an average score in polls of 89.95, and 86.93 in 2016, marking a decline in score. If the benchmark year is used to calculate the progress index, most communities regressed in 2016. At the achievement meeting, Deputy Director General S stated: We need the progress index; otherwise, our leader will easily notice the absence of this part. The progress index is the direct reflection of the working performance of our street C, it is obvious at first sight … Are there any other methods?
18
The second technique for index calculation: materials for the calculation
According to the design of the poll, each index was formed by corresponding questions and should not be changed. However, regarding the TXJY index, Deputy Director General S held a skeptical view. The ranking of PX community was far behind the expected value, and he thought that should not have happened. After calculation and error elimination for the second time, the results remained the same. After thinking for a long time at the achievement meeting, Deputy Director General S started to explain the whole thing and suggested: I think we should not blindly believe statistical approaches. The situation is like this. Originally, this mobile refrigerator (shared refrigerator) was also strongly recommended by the street TXJY but was put in the PX to let them manage the project. Now that PX is so low, the secretary (of the community) C would have a say. Let's see if we can think of a method. Of course, we should not go against the scientific method, and this is the premise.
21
The third technique for index calculation: grading of communities
Distribution of indexes after grading.
Residents’ utterance and silence
In the analysis above, the residents were in a state of aphasia, and the street needed some residents to fill out the questionnaires, while the neighborhood committees needed to organize such residents. Residents filled questionnaires out in a given environment, answered given questions, and made choices from some given options. Is it right, then, that the residents could not get rid of these restrictions? This was not exactly the case. The final part of the questionnaire included some open-ended questions, which were about ‘highlights of the community’ and ‘problems of the community’. Although many residents did not answer the 2 subjective questions, there were 899 samples with answers to ‘highlights of the community’, making up 50.71% of all samples and 556 samples with answers to ‘problems of the community’, making up 31.36% of all samples. Seen from the data obtained by on-site survey and individual characteristics, the respondents were mainly elderly people who had retired, so they acted slowly in reading and filling out the questionnaire; that is to say, their time cost in writing down answers to the subjective questions was higher. The enthusiasm of residents in writing down answers to such questions was more or less reflected in the quality of their answers. I estimated that roughly 40% of samples had long-sentence answers to subjective questions. The long sentences were mainly about specific problems; for example: The unashamed land developer occupied the public garages of homeowners; the parking fee is too high. What is even more hateful is that, when we were striving for justice and defending our rights, some of us were sued by them and arrested by the police. They threw their weight around and trampled on the law and justice!
22
The street dealt with subjective answers in the following two ways: first, the answer was written into the poll report by calculating the frequency of keywords—for example, the answer above fell within the category of ‘property management’, the word frequency of which was counted; second, it was written in the Community Diagnosis Manual, copies of which were to be distributed to the communities. Short and more informative sentences were used in the manual—for example, ‘It is notable that homeowners blocked the roads.’ Actually, the operation mode for polls is to make constant calculations after information has been condensed. The social contexts would inevitably be condensed into keywords in calculation. However, as stated above, there is far more than one method of dimensionality determination, context condensation, and definition, so the keywords of the case could also be ‘bullying by the property developer’ (price upsurge of garages); ‘law violation by residents’ (blocking roads); ‘social security crisis and social protest’ (arresting residents). For the street, it was inappropriate to write such keywords into the poll report: the former two words were inappropriate in wording, while the last two were confessions of their faults, indicating misconduct by the street and the community. Another answer that falls into the ‘property management’ category was ‘The security guard of the No.1 Door of the neighborhood always picks on people.’ The phrase ‘property management’ not only eliminated the social contexts behind the case, it also diminished the severity of the response. After the street condensed long sentences into short ones according to its own standard and categorized them into keywords, the subjective answers which originally could reflect the heterogeneous information to some extent became obscure after tailoring and polishing. What was the reason for this handling? To this end, an elderly teacher W in the street polling group
23
said: I know about the situation of XH, which did make a stir that time. That is a new neighborhood, and the neighborhood committee managed that district just for a short time, you know. This happened, to say nothing of them, we (the street) have no power to manage it.
24
Another index can show residents’ attitudes indirectly, namely, the non-response rate. It can be regarded as ‘the weak’s weapon’. Question abandonment means refusing to write answers. There are many factors influencing this phenomenon; for example, that the respondents cannot understand or are impatient regarding the question stem, are not satisfied with the question or setup of the options, or even do not want to cooperate with the poll, etc. Which communities’ residents, then, were most likely to have such attitudes? The questionnaire of the poll in 2016 had 44 questions, and, after calculating the per capita non-response rate, it was found that SD community had the highest abandonment rate, which was 20.50%; YFC community had the lowest abandonment rate, which was 4.15%, marking a difference of nearly five times; and the standard deviation of the communities was 4.32%. If we suppose that the resident samples of all communities were uniform and the answering environment was standardized, the per capita abandonment reflects at least two possibilities: the residents were not satisfied with the public opinion poll organized by the street, or they were not satisfied with the working performance of the community. The street set the questionnaire quality index based on the non-response rate and offered an explanation for it, as a kind of residents’ soft attitude, that is to say, to indirectly reflect residential areas’ satisfaction degree on the working team of the ‘two committees’ by using the attitude presented by invisible behaviors, which agrees with the poll principle from top to bottom.
25
Conclusions and discussion
In terms of funding, personnel, and duration, the scale of the poll conducted by C street was expanded and its survey process displayed greater scientificity. The street aimed to ‘make it a brand of the district’. 26 Its demonstration effect was amazing. As early as the beginning of 2016, when the poll project was praised by the leaders of the municipal Party committee, other streets in the district had sensitively realized the advantages of this mode and come for consultations with technical staff. What is the fate of the complex public opinion? We cannot account for it clearly. It is a process of information losing from context screening, question definition, and making a questionnaire, from setting options, implementing sampling, and establishing an index system. How the information is lost is determined by the space of power operations. In this process, the authorities at all levels positively or negatively set their own regulations, but they hold to the idea that ‘everything should be justified and every step should be implemented scientifically’. Under the guise of science, the tracks of power can hardly be identified. Finally, the informative social contexts disappeared and became visible indexes and the poll map. Through processing at different steps by the neighborhood committees and the street, the public opinions reflected by the poll could not be more familiar to them. Focusing on the polling process, the structure among the street, neighborhood committees, and residents became more stable. Existing problems such as community dualization, owners’ protests, the infiltration of state will into the daily lives of communities, and the political indifference of the middle class will continue to exist. In contrast, the street obtained legitimacy in some respects. The public opinion poll in communities which had lasted for three or more years is truly an innovation project on social governance, and the results of the poll would be taken as the reference for year-end evaluation of all communities for the functional departments and offices, and as the standard for community cadre appraisal and personnel setting, even as the reference in decision making by the street. For neighborhood committees, the issuing of the ‘Diagnosis Book for XX Community’ in the name of the residents of that community by the street set the direction of future work. Policy legitimacy, officials’ achievement, operation parameters of the bureaucratic machine, and work orientation of the community are all advantages for main bodies in grassroots governance brought about by public opinion polls. The only party that does not use poll data is the residents.
Is there, then, any cheating in the poll? According to my observation of the whole process, there was no cheating. However, polling is originally a statistics process, so how did it become a political process of power invasion? With the double security of polling methodology, how can public opinions be fabricated? The above-mentioned questions bring us to the primary question: can technical governance fulfill its mission? If the technique of the public opinion poll is still not so precise, then would governance by big data (Zhang, 2017) work? Would governance by smart government (Hu and Wang, 2017) work? Would governance by complex system paradigm (Fan, 2015) work? 27 Viewed from the perspective of our analysis of the whole process of the public opinion poll, the key to the question seems not to be the level and precision of technical governance, but the logic of it. When the polling technique changes ‘potentially’ complicated public opinions into ‘actual’ data and indexes, public opinions are gradually simplified and compressed. During this process, no matter how strict the technique is, it can only ensure the ‘legitimacy’ in simplification, while the direction of simplification is determined by the operator.
Theoretically, in a survey, public opinions can be compressed in any direction and finally form a database. In reflecting public opinions, since researchers work out public opinions, can that which is worked out by the government not be called ‘public opinions’? Or do public opinions as a constructive substance rather than an entity lack the status that they should have? Different goals of polls urge different operators, such as researchers engaged in social sciences, media, or governments, to select their topics of interest respectively; considerations of different perspectives, funds, or personnel allow them to choose different ways of surveying and sampling. The factors above are not stipulated by the polling technology, but focus polling in different directions of simplification and acquire totally different public opinions. In the same way, setting up a government affairs account on WeChat/Weibo could not solve the problem of pushing similar news; grid governance could not solve the problem that can be easily solved by submitting photos by the grid managers; a public opinion monitoring platform could not realize identification, categorization, and qualitative measurement of events, just like the taxation techniques in the Qing dynasty could not avoid the blackmail of common people by government runners, scribes, and servants. The technique is not distorted, the reason is the futility of technique on operators’ will.
The paradox of technical governance thus begins to take shape. Since the mission of techniques is to bring things out by simplifying the possibilities, it surely gives way to the power of the operators to decide the direction of simplification. Since governance means subjects’ ‘tending’ and ‘guidance’ of objects, the process of the state ‘tending’ society
28
is simplifying and guiding society in a certain direction. In measuring society through technical governance, the state also molds a society in its eyes by implanting its genes therein. Therefore, the paradox of technical governance is that in observing social views through the eyes of techniques, what the state obtains might be its own reflection. Governance determines the direction of measurement; techniques determine the process of measurement. Technical governance views society from a specific perspective; in restricting the image formation of society, the possibility for the state might also be restricted.
29
When local governments are competing for systematic innovation and governance innovation, when we acknowledge the advantages of technology in simplifying social complexity and pursue better technology, the advice by Foucault (1972: 229) might not be inappropriate: That a particular discourse cannot be resolved by a prior system of significations; that we should not imagine that the world presents us with a legible face, leaving us merely to decipher it … there is no prediscursive fate disposing the word in our favour.
Footnotes
Acknowledgement
This article benefited from the follow-up survey conducted by Professor Bo Peng on the ‘Evaluation on Public Service Capability of X City Neighborhood’ from 2015 to 2017. (For research ethics, this paper anonymizes the project, the place names, and people’s names and positions.) As the head of our research team, I would like to extent my gratitude to Yong Fang, Xiao Han, Yijia Chen, Wei Xie, Jianfei Zhu, Bowen Zhou, Weiqiang Huang, Xiaobing Zhang, Bin Shen, and other team members for their excellent and hard work of interviews, and data collection and sorting.
This article was presented at the ‘Governance Transformation and Social Organization Development in Contemporary China’ conference at Shanghai University. Thanks for the suggestions and encouragement of teachers, including Xiaochun Huang, Han Zhang, Xiulin Sun, Yanhua Deng, Zhiming Sheng, and Yingying Ji, who inspired me to adjust my writing direction. I am also greatly indebted to the anonymous reviewers for giving timely review comments. My gratitude also goes to Dr. Haozhou Lin for his ongoing discussion with me in the process of writing the paper. The author takes sole responsibility for the article.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.