The psychosocial cognition of college students' on COVID-19 is positively correlated with active coping while negatively correlated with negative coping. The college students with high cognition are likely to take active coping measures (13). This notion indicates that improving the cognition on the disease can aid individuals to respond positively. Approximately 84.0% of college students believe that the government has taken positive and effective measures, recognized the efforts made by the authorities and have confidence in the mandated epidemic prevention and control efforts (14). Purportedly, 95.8% of college students have collected information about COVID-19 via WeChat and microblog. Still, 83.0% of college students have collected information about COVID-19 epidemic situation via TV and radio. The emerging media, represented by WeChat and microblog, are characterized by instant communication, popularity of the main body and visualization of pictures and texts. Such platforms have broadened the learning channels of college students and meet the psycho-socio needs of college students. Thus, new media have become the main channel for college students to acquire epidemic information.

The psychological mood of college students is greatly affected due to rising pressures from diverse sources. For example, the implementation of strict home quarantine causes many college students to stay at home. Often characterized as in the psychological development stage, these students are likely to develop negative psychological emotions, such as depression, anxiety, and stress, due to undue pressures both from the epidemic and long hours of study. In serious cases, these emotions can even affect the students' health. Depending on the environmental circumstances, the prevalence rates of PTSD (Post-Traumatic Stress Disorder) and depression during the outbreak range from 2.7 to 9.0% (15). If college students are in their fourth year (the year of graduation), and live in seriously affected areas, then extreme fear is the key influencing risk factor causing psychological distress, followed by short sleep time.

Odriozola-Gonzz et al. (16) investigated the psychological wave of Spanish University of Valladolid during the outbreak. About 50.43% of the interviewees said that they were moderately affected, some even severely. Students majoring in the arts and humanities, social sciences and law had higher scores in anxiety, depression, and stress, whilst those majoring in engineering and architecture have lower scores. The clinical depression cases increased by 25–30 times due to the closure measures taken during the outbreak, and the number of Greek college students with suicidal thoughts increased by eight times (17). One third of people accepted the theory that COVID-19 was conspiracy. The study results especially emphasized the need to take specific intervention measures for the mental health of disadvantaged groups.

Put simply, investigating the knowledge of college students on COVID-19 and analyzing their attitudes toward epidemic prevention and control measures will be key to unlock and improve our understanding of how best to prevent and hinder the spread of COVID-19 among college students, as well as how best to plan on beneficial and impactful educational interventions for them (18). The COVID-19 had a greater impact on the psychology and sleep of college students (19, 20). Virtual mentorship pairing was the highest rated educational intervention suggested by first- and second-year medical students. The third- and fourth-year medical students had frequently cited virtual surgical skills workshops (21). The prevalence rates of anxiety and depression were 7.7 and 12.2%, respectively (22). The study results of (23) showed that the incidence of somatic symptoms amongst college students was 34.85 (mild, 26.26%; moderate, 8.59%). The incidence of somatic symptoms in primary school students was 2.39% (all mild).

Amongst the entire cohort, concern regarding COVID-19 was positively correlated with the occurrence of somatic symptoms. Somatic symptoms were likely amongst college students expressing greater concern regarding the threat to life and health posed by COVID-19 and the efficacy of prevention and control measures. Approximately 70% of learners reported that they were involved in e-learning during the lockdown. Most learners used android mobile in e-learning environments (24). Students have been facing various problems related to depression anxiety, poor internet connectivity and unfavorable study environment at home. Importantly, students who face enormous challenges in their studies during this pandemic come mainly from remote areas and marginalized sections of society.

The outbreak of the epidemic increased the loneliness of overseas students and affected their career development (25). With the outbreak of the epidemic, a stable family income and living with parents could effectively reduce anxiety (26). Economic effect aside, the impact on daily life and academic delay are also positively correlated with anxiety, whilst social support is negatively correlated with anxiety (27). Savitsky et al. (28) studied the influence of COVID-19 on the psychology of nursing students. Their results showed that female gender, lack of personal protective equipment (PPE) at work and being a parent were significantly associated with high anxiety scores. Zhang et al. (29) studied the influence of COVID-19 on the psychology of international students in Medical Bachelor, Bachelor of Surgery of Zhejiang University. Even though students' knowledge on Traditional Chinese Medicine (TCM) and their discussion and consultation behaviors have been significantly improved, and students' understanding on the necessity of TCM has been significantly enhanced based on the online Chinese medicine courses, most students still preferred face-to-face classroom learning compared to online learning.

In general, researchers believe that major IPO is the product of the comprehensive action of various social factors, and they can explore the causes from various aspects, including the subject and object of public opinion, media, social environment, contingency factors and the manifestation of public opinion (30).

Briefly, some scholars analyzed the causes of IPO from four dimensions, including political and economic reasons, social environment and hot events, false and bad information dissemination, the strong media influence of new media era (31, 32). Also, some scholars analyzed the configuration of the influencing factors on IPO and found that maintaining high sensitivity of information source, high participation of Internet users and/or high media activity and high government intervention can stimulate the generation of a heat index of high public opinion (33, 34). In the study reported herein, we investigated the causes of generating a major IPO from specific fields.

The first cause includes objective and realistic factors. The IPO is an inevitable product of the frequent occurrence of social public events. The sudden public events will cause great psychological influence on the public (35, 36). The second comprises government intervention factors. If the timeliness of government intervention is worse, and the intervention level is high, then the heat of IPO is high. The heat of public opinion events responded by the government via press conference and social media is high (37, 38). The third involves the psychological factors of public opinion subjects. The optimistic mood of netizen and the pessimism of the government can bring favorable results to the development of public opinion, and the influence of the emotional state and intensity of the government is great (39).

Adult attachment (AA) is a sustained and long-term emotional connection between adult individual and current peers and a stable interpersonal interaction style of the individual (42). Hence, it can influentially affect the consciousness of one's social participation in a positive manner.

Different scholars have conceptualized how social motivation (SM) can influence or affect people differently. While Hilvert-Bruce et al. (43) believed that SM is the basic force driving people's social behavior, and Hernandez et al. (44) believed that SM is an internal motivation that promotes individual activities to achieve a certain goal. Based on the dictionary of psychology (45), we purport that SM is the psychological tendency of individuals to engage in certain activities to meet their own social needs, which is directly related to their social needs, but not to their physiological needs. Hence, it can surely influence the consciousness of one's social participation.

Information affective disposition (IAD) refers to the college students' inner emotional attitude toward the text when they are reading the major epidemic information. Therefore, it too can certainly influence the consciousness of one's social participation as well as one's emotion, which is linked invariably to H4 below.

Emotional motivation (EM) refers to the condition wherein people release public crisis information to express emotions. After the outbreak of a major epidemic, the public's psychology, emotion, and belief will be affected to different extents. Individuals in such crisis situations are prone to have panic, depression, anxiety, and other negative emotions. The public can reduce anxiety and vent discontent based on the information transmission via the Internet (46, 47). The public will often express their feelings to relatives, friends, and even strangers via real-time dissemination of information to seek comfort, understanding and spiritual support to alleviate the inner uneasiness in the crisis situation or release the dissatisfaction with the crisis management stakeholders (48). Hence, as noted in the discussion for H3, IAD will certainly also influence one's emotional motivation.

Emotional catharsis is a form of public right of discourse and participation, which is conducive to the self-regulation of personal emotions in crisis situations, seeking self-comfort and psychological balance (49). Not surprisingly, we can speculate that AA, which projects one's attachment to others' emotions and related feelings, to influence one's EM significantly.

Risk relates to the consideration of the possible loss or injury caused by objective facts (50, 51). EM has a significant influence on individual participation in social activities (52). Put together, risk perception, especially that is related to the risk that the dissemination of error messages will possibly lead to crime, will generally be heighten for the public when the loss or injury caused may be serious. Accordingly, such risk perception will assert an influence also on one's EM.

As indicated, SM or one's psychological tendency to engage in certain activities to meet one's social needs can play a critical role in the dissemination of information. Since the 1970s, emotion study has blossomed into a new discipline. Exploring key influencing factors on microblog users' participation in interaction, Stuart and Martin (53) found that perceived usefulness and expectation confirmation significantly affect users' willingness to use.

Weingart et al. (54) analyzed the intention of dissemination vis-à-vis the theory of SM and pointed out that personal needs, interest promotion, self-expression and self-actualization are the main motivations to influence the intention of dissemination.

As well, the theory of EM proposed by Auguste Comte, a famous French philosopher and the founder of sociology and positivism, tells us that emotion, behavior, and talent comprise humanity. Emotion is a decisive factor; as such, human emotion drives human reason and activity goal. Put simply, human behavior is mainly driven by EM. Hence, it is expected that EM will critically influence one's willingness to disseminate information.

Cao et al. (55) analyzed the influencing factors of users' willingness to disseminate enterprise microblog information from the perspective of enterprise microblog marketing. Their results showed that the benefits and reputation increase positively affected their intention of dissemination.

Some scholars also analyzed users' motivation from the perspective of a microblog generation content (56). The research conclusions from Hsu and Lin (57) pointed out that the entertainment role, altruism and promotion of honor had a positive influence on the generation behavior of bloggers. The promotion of status positions and the sense of belonging had a significant influence on the continuous generation and sharing of a user content. Park et al. (58) indicated that the interest, usefulness, altruism, and identity have a significant influence on the generation behavior of video users.

The social participation mechanism, a participatory democratic process and approach, achieves the growth and development of society via active participation into social development activities moderated by the concern of the public over their own interests and awareness respecting social public interests and public affairs (59). This approach is a code of conduct and norm for the entire society to discuss, consult and make decisions on social affairs (60, 61).

With the frequent occurrence of crisis and the development of democracy, the consciousness of crisis, political consciousness and the sense of participation are gradually strengthening. In public emergencies, public participation has become a new norm (62). The motivation of crisis information dissemination in a mobile Internet environment includes egoistic (emotional and social) motivation and altruistic (informational) motivation (63).

In this study, the questionnaire survey method was used to conduct empirical research and data collection. Chinese college students served as the objects investigated.

The questionnaire, designed according to the unified standards and requirements, was distributed online via WWW.WJX.CN, an online research data collection platform located in Changsha, China. This platform executes the questionnaire administration process and provides powerful data collection, storage, and analytical tools to deeply explore the value of data being collected. Presently, reported statistics on WWW.JX.CN indicated that about 83.15 million users have responded to over 6.55 billion questionnaires.

In this study, a total of 438 questionnaires were distributed, and 319 valid questionnaires were collected with an effective respond rate of 72.83%. Descriptive statistics indicated that, amongst the valid samples, 48.6% were male, and 51.4% were female.

According to data shown in Table 1, the distribution of the objects investigated was as follows:

With respect to the “Grade” variable, first-year students (freshmen) accounted for 28.8%, second-year students (sophomore) for 35.4%, third-year students (junior) for 24.8% and forth-year students (senior) for 11%.

In term of specialized subject, liberal arts students accounted for 36.7%, those with science major for 34.8%, those with engineering major for 26.3%, and students of arts and sports for 2.2%.

As for time spent online: 4.1% of college students surfed the Internet for 1–2 h daily, 28.5% surfed the Internet for 3–4 h daily, 37.9% surfed for 5–6 h daily, and 29.5% of the students surfed for more than 6 h daily.

Regarding the frequency of paying attention to COVID-19 public opinion: 28.5% of college students paid attention to it daily, 54.5% of college students often paid attention to it, 16.3% of the students paid attention to it occasionally, and 0.6% of them paid less attention to it.

The questionnaire adapts multiple items from previously well-developed scales. It was also designed in strict accordance with the translation–back translation procedure. Altogether, the questionnaire has been appropriately adjusted through pretesting relevant questions, recognizing the actual situation of Chinese college students.

The comprehensibility of problem description was ascertained via interviews. The pre-survey questionnaire was issued according to the results. The questionnaire was further modified vis-à-vis the feedback of the pre-survey. The formal questionnaire was finally formed. Besides the investigation on the basic information of college students, the Likert 7-scale was used to record the attitude of objects investigated toward various questions provided in the questionnaire. “1” represents strongly disagreed; “7” indicates strongly agreed.

The Affective Disposition (AD) Scale was adopted from the (64) questionnaire with the scale comprising two question items. The average score of all items in the scale was regarded as the AD measurement score. The AD was significant when the score was high. The Dissemination Willingness Scale was adopted from the (52) questionnaire with its scale having three question items. The average score of all items in the scale was regarded as the measurement score. The dissemination willingness of the objects investigated was strong when the score was high. The Emotional Motivation (EM) Scale was derived from the (65) questionnaire. This scale contained two question items. The average score of all items in the scale was regarded as the EM measurement score. The EM of individuals was strong when the score was high.

The Adult Attachment (AA) Scale was derived from the (66) questionnaire; accordingly, this scale has three question items. The average score of all items in the scale was regarded as the AA measurement score. The AA of individuals was strong when the score was high. The Social Motivation (SM) Scale was derived from the (44) questionnaire, which contained two question items. The average score of all items in the scale was regarded as the SM measurement score. The SM of individuals was strong when the score was high. The Social Participation Consciousness Scale was drawn from the questionnaire designed by (67). This scale contained three question items. The average score of all items in the scale was regarded as the Social Participation Consciousness measurement score. This characteristic of individuals was strong when the score was high. In the Risk Perception Scale, the questionnaire designed by (68) was adopted. This scale contained two question items. The average score of all items in the scale was regarded as the risk perception measurement score. The risk perception of individuals was strong when the score was high. The final measurement items of seven constructs may be found in the Appendix.

The reliability and validity of risk perception, dissemination willingness, AA, EM, and SM were analyzed by SPSS24.0 with results shown in Table 2. This table illustrates that the Composite Reliability (CR) of all potential variables to be more than 0.8, and Cronbach's α coefficients of all variables are all higher than the recognized minimum level (0.6). This indicates that all the scales have good reliability.

Exploratory factor analysis was used to test the construct validity of the scale. The factor load of each item corresponding to all variables was greater than the threshold value of 0.7, thereby indicating a good construct validity of the scale (69).

The average variance extracted (AVE) of each variable is more than 0.65, thereby indicating that the convergence validity of the table is good (70). Table 3 illustrates that the square root of AVE of each variable to be more than the correlation coefficient of this variable vis-à-vis other variables. A good discriminant validity exists amongst the variables (71). In summary, the scales used in this study have good validity.

Path coefficient represents the strength of the relationship between independent and dependent variable. The analytical results of the path coefficient for the research model being validated are shown in Figure 1 and Table 4.

From the analysis results, among the first ten (10) primary hypotheses advanced on the structural model test, eight (8) were supported. R² represents the variance variability of the dependent variable explained by the independent variable. In this study, bootstrapping repeated sampling method is used to calculate the t-value of the significance test based on the selection of 2,000 samples. The degree of interpretation of consciousness of social participation, EM and dissemination willingness are 0.426, 0.427, and 0.545 respectively, thereby indicating that the model has a good interpretation effect (71).

As quantified in Table 4, bootstrapping method has been used to test the significance of the path coefficient of the structural model. As depicted in the distribution plot of hypothesis testing shown in Figure 2, there is a significant correlation between the seven variables in the hypothesis test.

First, AA (adult attachment) has a significantly positive influence on the consciousness of social participation (β = 0.157, t = 2.438), affirming Hypothesis H1. SM (social motivation) also has shown a significantly positive influence on the consciousness of social participation (β = 0.320, t = 6.323), confirming Hypothesis H2. Conversely, IAD (information affective disposition) has no significant positive effect on the consciousness of social participation (β = 0.047, t = 0.774), inferring that Hypothesis H3 is not supported.

Even so, for hypotheses H4 through H6, AD has shown to have a significant positive influence on EM (emotional motivation) (β = 0.249, t = 5.114); hence, Hypothesis H4 is supported. AA also has a significant positive influence on EM (β = 0.238, t = 4.182), confirming Hypothesis H5. As well, risk perception has a significant positive influence on EM (β = 0.215, t = 3.678), thereby confirming Hypothesis H6.

In term of hypotheses H7 through H10, results of the data analysis showed that SM also has a significant positive influence on the dissemination willingness (β = 0.204, t = 3.636). Therefore, Hypothesis H7 is supported. As well, the consciousness of social participation has a significant positive influence on dissemination willingness (β = 0.235, t = 3.890). Hence, Hypothesis H8 is also supported. Additionally, EM has a significant positive influence on dissemination willingness (β = 0.160, t = 2.967), thereby confirming Hypothesis H9. However, risk perception has no significant influence on dissemination willingness (β = 0.023, t = 0.409). As such, Hypothesis H10 is not supported.

In this study, we conducted a mediating effect hypothesis test by using a three-step mediating regression analysis as proposed by (72). In the first step, AD is used to regress dissemination willingness; next, AD is used to regress EM. Finally, in the third step, AD and EM were used together to regress the dependent variable (dissemination willingness). The regression coefficient must be significant in the first and second steps. In the third step, if the regression coefficient of EM is significant, and the regression coefficient of the independent variable AD is insignificant, then EM plays a complete mediating role. If the regression coefficients of AD and EM are significant, but the regression coefficient of AD becomes weak, then EM plays a partial mediating role.

This study analyses the mediating role of EM in the relationship between AD and dissemination willingness via the bootstrapping method on the basis of the suggestions of (73). Here, 5,000 times of repeated sampling is adopted in the bootstrap analysis to construct a 95% confidence interval for bias correction.

Table 5 demonstrates that the indirect effect of EM in the relationship between AD and dissemination willingness is 0.051 (CI = [0.014,0.088]), and its bootstrap 95% confidence interval does not include zero. This result indicates that the mediating effect of EM between AD and dissemination willingness is significant. Therefore, hypothesis H11 is supported. The indirect effect of the consciousness of social participation in the relationship between SM and dissemination willingness is 0.070 (CI = [0.038, 0.127]), and its bootstrap 95% confidence interval does not include zero. This result indicates also that the mediating effect of the consciousness of social participation in the relationship between social motivation and dissemination willingness is significant. Hypothesis H12 is confirmed.

First, the centralization treatment is conducted for the independent and moderating variables to test the regulatory effect of SM on the relationship between risk perception and dissemination willingness. Then, the product terms of SM, risk perception and dissemination willingness are constructed to conduct a multilevel regression analysis. SM has a significant moderating effect on the relationship between risk perception and dissemination willingness.

In the study, risk perception is divided into high, medium, and low situations to clearly show the role of the moderating variables. Excel is used to draw the influence of SM on dissemination willingness in high, medium, and low risk perception. The main effect is 0.87; the moderating variable effect is 0.724; and the moderating effect is −0.121. The moderating effects are shown in Figure 3. Our results show that the influence of SM on dissemination willingness is reduced when the risk perception is high, confirming Hypothesis H13.

This study unveils key influencing factors on IPO dissemination in major epidemic prevention and control from two primary dimensions, namely, SM and EM. The conclusion of this study compensates for the current studies on group factors and the mechanism of social participation consciousness. This study enriches IPO research with several theoretical insights, expanding the research scope of students' information behavior theory in the process of spreading public opinion on the epidemic.

Studies have theoretically verified that EM plays a mediating role in the relationship between AD and dissemination willingness; consciousness of social participation plays a mediating role in the relationship between SM and dissemination willingness; and SM moderates the relationship between risk perception and dissemination willingness (77). The research model offers insight and analysis on the governance mechanism of IPO in the prevention and control of major epidemic to supplement the understanding of current studies on the role of governance mechanism.

The governance effect of IPO determines the guidance of public opinion in the context of sudden events, which is directly related to social stability. Effective measures for IPO governance must be determined.

The outcome of this effort is of great relevance to the governance of IPO in the prevention and control of major epidemic and of applicable significance to advancing the government's information governance ability. We should strengthen the guidance of public opinion in major epidemic prevention and control and fully build and make good use of all types of information transmission paths. The spread of IPO is a group activity, rather than an individual action. This situation has a strong external effect.

In the context of public emergencies, IPO has an inciting force to the group. When dealing with the IPO in major epidemic prevention and control, it also tests the ability of prevention and response to group events. We must set up a network warning in advance, implement information communication channels in time and ensure the real-time interaction of information. These steps can reduce the possibility of Internet users to believe and spread rumors to a certain extent, thereby promoting the public to form self-awareness and moral obligations and maintain a harmonious network environment.