The concept of impression management was introduced by Goffman (1959), who argued that “impression management is like theater” and focused on the role of self-representation in daily life. Jones and Pittman (1982) extended impression management from the sociological realm to the psychological realm and suggested that self-presentation also involves one’s attempts to control others’ impressions of his/her personal characteristics. Social psychologists Leary and Kowalski (1990) suggested that impression management consists of two distinct processes, namely, impression motivation, which refers to an individual’s desire to control the perceptions and impressions that others form of him or her, and impression construction, i.e., “choosing the type of impression to be created” and putting it into practice. On this basis, Tetlock proposed a three-component model adding “impression monitoring”, indicating that individuals consciously seek to make a certain impression on others and keep a close eye on others’ impressions of themselves.

With the development of impression management theories, more and more scholars have applied these theories to the analysis at the organizational level, and took enterprises as the main objects of analysis to explore the drivers and impacts of the impression management behavior of enterprises (Jin, 2018). Some researchers have already proved the applicability of impression management at the macro-organizational level. However, there are seldom studies that have investigated the impression management behavior of governments as study objects, and even among the very few that do exist, they mainly adopted case studies to analyze the improvement of government image in cyberspace (Landtsheer et al., 2008; Geng, 2014). Allen and Caillouet proposed different impression management behaviors for different stakeholders, where the improvement of impression management behaviors are designed to improve images while protective impression management behaviors to reduce damage to images (Mohamed et al., 1999). This study focused on the impact of the improvement of impression management behaviors.

As a kind of trust relationship, government trust is the psychological state and attitude that the public holds toward the government at the level of expectations and perceptions (Hetherington and Husser, 2012). What impression a government presents and how the impression is constructed influence the credibility and governing power of the government (Geng, 2014). A country’s foreign policy often affects the impression of the people of other countries (Hadaś-Dyduch, 2019; Minardi and Lestari, 2019). Citizen’s government trust can be enhanced through improving forms of information dissemination and impression management is exactly the strategic demonstration of information (Sallot, 2002). The impression management behavior of government enhances public trust in local government by influencing the public’s perception of governments, making them think that their governments fit well with current social expectations and social norms (Dowling and Pfeffer, 1975). Zhang and Ip (2019) found that the relationship between city government’s performance in pibulic service and residents’ trust in government is of great significance to public management of Chinese cities. Furthermore, evaluation of government performance interacts with perception of corruotion to negatively influence political trust (Wang, 2016). Therefore, the following hypothesis was proposed:

Cohesion the degree of mutual attraction, identification and support, which leads to emotional resonance or behavioral agreement based on common values, common goals, and common interests (Back, 1984; Li Y. N., 1998). No society can sustain and develop without public trust, support and commitment which are means that reduce the cost of political governance and effectively promote political solidarity and social integration. Political identification or trust as important factors influencing social cohesion (Li Y. S., 1998). Social cohesion is the degree of dependence, cooperation and unity among members of a society. It is particularly important for countries in transition (Jenson, 1998). The existence of groups which have grievancFes threatens social cohesion (Sapsford et al., 2019). The evidence suggests that governments that are not responsive to the expressed needs and wishes of their people do eventually fail (Acemoglu and Robinson, 2012). Huang directly regarded government trust as a dynamic indicator for indirectly interpreting changes in social cohesion. Some industrial management studies have confirmed the relationship between enterprise management and enterprise cohesion (Kovacova et al., 2020; Kovacova and Lewis, 2021). As a result, the following hypothesis was proposed:

Studies have been conducted to analyze the factors influencing government trust, primarily from three aspects, namely, performance factors, cultural factors, and individual factors. Among them, individual factors mainly include individual characteristics and cognitive ability levels, specifically gender, age, education level, place of origin, and personality of individuals. Based on data from a large-scale questionnaire survey of Chinese urban citizens, Zhao employed linear regression and found that citizens with low education levels had higher government trust compared to those with high education levels (Zhao and Wang, 2019). Zhong and Chen (2002) suggested that better educated people were more rational and more skeptical of the information presented by governments and they are inclined to rationally analyze governments’ impression management behavior, thus their government trust tended to have smaller increases. Thus, the following hypothesis was proposed:

In this study, we used a questionnaire survey analysis approach. The official questionnaire was finalized and the formal research kicked off in November 2018. The questionnaires were distributed online to students enrolled in the Songjiang University Town in Shanghai, and 390 questionnaires were collected in total, of which 336 were valid.

The questionnaire consisted of two parts. The first part contained questions for demographic data, including the gender, age, and education level. The second part contained questions for the main variables of the study, including governments’ impression management behavior, government trust, and social cohesion.

The independent variable was “governments’ impression management”, which was based on the scale developed by Goffman (1959) and Pan et al. (2017), and adjusts it according to government scenarios, forming “Government departments will convey care to the public” “The publicity issued by government departments to the society is vivid and cordial” and other 5-item scale measurements (α = 0.85); the dependent variable was “social cohesion”, which was based on the scale developed by Jenson (1998), Wang (2013) and Zhu (2007), forming “I think that members of society are willing to help each other” “I love our motherland very much” and other 5-item scale measurements (α = 0.74); the mediating variable was “government trust”, which was based on the scale developed by Gao (2014), forming “I think the government is willing to listen to the opinions of the people” “I think the government handles things fairly and appropriately” and other 5-item scale measurements (α = 0.90). All scales were evaluated on a 7-point Likert scale, ranging from 1 = “completely disagree” to 7 = “strongly agree”.

The moderating variable in this study was citizens’ education level, comprising four categories of the below-university level, bachelor’s level, master’s level, and doctoral level.

Moreover, factors such as citizens’ gender, age, and household income have been proved to affect citizens’ trust in government and social cohesion, and were thus included in the model as control variables in this study (Zhu, 2007; Wang, 2013; Wu et al., 2016).

In order to understand the general characteristics of the samples and the overall performance of the main variables, the study first conducted a descriptive statistical analysis of the samples as well as the key variables. Based on the statistics, 52% (n = 187) of the samples studied were female, and 48% (n = 173) were male, representing a basically reasonable gender ratio in line with reality. Among them, the largest age group was the group of 16 to 27 years old, of whom the group of 19 years old took the majority, totaling 84, followed by the group of 18 years old, totaling 67, then the group of 20 years old, totaling 50, finally the group of 16 years old, with 1 person only. As to household income, the majority was concentrated in the range of 100,000–200,000 RMB, totaling 142 persons, followed by a total of 132 persons in the range of less than 100,000 RMB, 54 persons in the range of 200,000–300,000 RMB, 24 persons in the range of 300,000–400,000 RMB, 3 persons in the range of 400,000–500,000 RMB, and 6 persons in the range of more than 500,000 RMB. Table 1 shows the means, standard deviations, and Pearson correlation coefficients of all variables in this study.

Due to the complex structure of the model, and the mediating and moderating relationships between the constructs, the hypothesis testing in this study was performed using the PLS-SEM method. The R square of the dependent variable represented the predictability of the theoretical mode (Chin, 1998). The analysis revealed that the R² of social cohesion was 17% and the R² of government trust was 8%.

According to Figure 2, the results of data analysis showed that governments’ impression management positively affected the public trust in government (β = 0.22, p < 0.001), public trust in government positively affected their social cohesion (β = 0.35, p < 0.001), and citizens’ education levels positively moderated the relationship between governments’ impression management and government trust (β = 0.15, p < 0.05). Thus, H1 and H2 were verified, while H3 was not, with the moderating effect graph shown in Figure 1 below. According to Figure 1, the effect of governments’ impression management on citizens’ trust in government increased with citizens’ education. One possible explanation is that individuals with higher education levels enjoy more social capital and are more willing to participate in public affairs, which also reflects a higher level of trust in government among this group.

Preacher and Hayes (2004) deemed the Bootstrap method as a more robust and effective way to test mediated effects with moderation. Therefore, this paper performed a robustness test on the data using the PROCESS program developed by Hayes. The “Bootstrap” analysis in this study was performed on 5000 replicate samples with a confidence interval of 95%. If the upper and lower limits of the confidence interval included zero, the effect tested was considered insignificant. The mediating effect was first tested, with the results of the data analysis shown in Table 2. It can be seen that the confidence intervals did not include 0, no matter with low, median or high government trust, indicating that the mediating effect of government trust between governments’ impression management and social cohesion was significant.

According to Table 3, the judgment indicator of the existence of moderating effect of education on the indirect relationship of governments’ impression management affecting social cohesion was 0.062 [confidence interval (0.003, 0.139)], with no 0 and with significance, thus confirming mediating effect with moderation. Therefore, the government trust of people with higher education is more likely to be affected by the behavior of government impression management; The political trust of people with low education level is not easy to be affected by the behavior of government impression management. This may be because people with low education level do not pay more attention to government behavior than those with high education level, and they do not understand the complex reasons behind government behavior.