1. Introduction

1.1. Research Background

With the continuous popularization of higher education and the global outbreak of the COVID-19, college students in the post epidemic era are facing the huge problem of choosing after graduation. At present, the number of college graduates in China in 2022 is expected to be 10.76 million, an increase of 1.67 million year-on-year, with both scale and increment reaching a record high [1]. Three million graduates in the United States may need to modify their future development path due to the present economic climate, as reported by Consumer News and Business Channel (CNBC) for the 2019-2020 academic year [2]. Meanwhile, the epidemic's effects on the labor market persist. There are still sectors and businesses whose output and operations have not returned to pre-epidemic levels. There is still a lot of doubt about the job market's demand for workers. Some smaller and medium-sized businesses are less able than they were to increase their workforce in recent years. Unfortunately, university recruiting efforts have suffered as a result of the pandemic. There is a coexistence of "difficult employment" and "difficult recruitment", and there are great differences in employment needs among different majors, industries and regions [3-4]. Taking Shanghai as an example, the proportion of college graduates choosing to stay in Shanghai for employment is getting higher and higher every year, and the employment competition in Shanghai is becoming more and more fierce. In addition, according to relevant data, the proportion of college students in Shanghai who choose to go abroad has increased under the influence of the epidemic [5]. According to the comprehensive analysis and judgment, the current employment situation of college graduates is still grim and complex, and the graduates gradually show the characteristics of diversification in their future planning.

At present, the main body of college students is mainly composed of the generation born after the 2000. These young people have been living under the influence of the Internet since childhood. Their information sources are mainly pushed by hot information on social media platforms [6]. Many college students will forward and comment on the hot push related to the post-graduation planning on the social media platform. The reading and comments of these articles on the social media platform are often hundreds of thousands. Many college students also say that these articles have had a great impact on their future planning. Many college students even say that they will choose a job or study abroad institution through the methods suggested in these articles. Weibo has changed the way of information dissemination and realized the real-time sharing of information. It is a new entry to the Internet after portal and search. Among Weibo users, the post-90s and post-00s have become the main force, accounting for nearly 80%; post-80s users account for 18%; The proportion of users after 70 is 3% [7]. College students here refer to undergraduates and junior college students who graduated in June this year.

There have been previous studies between graduates and social media at home and abroad on the use of social media to complete entrepreneurship psychological education for graduates, the impact of social media on the employment of physics graduates who are not physics majors, and the use of social media to conduct virtual job fairs [8-10].

1.2. Research Gap

Some people have previously done analyses on the psychological education of graduates by using social media around the employment of graduates, the impact of the use of social media on Graduates' employment and the impact of social media on virtual job fairs [8-10]. However, no one has done relevant analyses on the relationship between College Students' post-graduation planning (direct employment) and hot content push published on social media under the current epidemic background. At present, this field is still in the blank stage. Therefore, what is worth investigating is: the impact of social media information push on College Students' post-graduation planning?

The purpose of this investigation is to investigate the impact of hot information push on social media on College Students' Graduation planning in the context of the epidemic, so as to better grasp the hot information push on social media platforms, have a certain beneficial impact on College Students' future planning, and enhance their employment willingness. At the same time, it is also expected to bring some inspiration to scholars who want to explore related issues in the future.

1.3. Fill the Gap

According to the analyses background with college students as the analyses object, this paper can draw a hypothesis: the hot information push of social media is closely related to the future planning of college students. The hot information pushed by social media platforms and "studying abroad", "epidemic situation" and "employment" will greatly affect the future planning of college students. In order to test this hypothesis, this paper will apply TPB theory to collect the first-hand data of college students in Hangzhou through questionnaire survey for quantitative data analysis, and analyze the final survey results to get a conclusion to test this hypothesis.

2. Literature Review

2.1. Definition & Development

The planned behavior theory inherits and develops the rational behavior theory (TRA) jointly proposed by Ajzen and Fishbein. It was proposed by Ajzen. When trying to figure out why and how people alter their routines, the theory of planned behavior (TPB) is a useful tool [11]. Human conduct, according to TPB, is the product of a deliberate strategy. Planned behavior theory is a theory that links belief and behavior. This theory is widely used in the analysis of advertising, public relations, advertising activities, health care and other fields. It aims to improve the forecast ability of reasoning behavior theory by including perceptual behavior control. The theory includes five factors: "attitude", "subjective behavior norms", "perceptual behavior control", "behavior intention" and "behavior". The theory holds that among these five factors, all potential influences on behavior have an indirect impact on actual behavior because of the role that behavioral intention plays in its execution. There are three interconnected elements that influence an individual's purpose to behave in a certain way. One is a person's subjective "attitude," or one's "attitude" toward engaging in a given conduct; Second, the external "subjective norms", that is, the "subjective norms" that affect individuals to take specific behaviors; The last one comes from "perceptual behavior control" [1].

The CMM postulates that people are motivated in different ways to pay attention to and process news stories. The quantity of information people take in from the media may be influenced in two ways: (a) through direct exposure to the content and (b) through indirect exposure, which is governed by news elaboration [12].

2.2. Important Results

TPB theory is widely applied in many fields, and many people have applied TPB theory to make relevant analyses and Exploration on some problems. Gibson et al. applied TPB theory to apply this theoretical model in the field of environmental protection. The purpose of this investigation is to determine whether political factions, political ideology and planned behavior variable theory related to water saving forecast the intention to engage in water saving. The analyses results have had an impact on Environmental Exchange, that is, pay attention to strengthening the subjective norms of water conservation [13]. This theory has also been applied to do relevant analyses on entrepreneurship. The purpose of this analyses is to test the influence of attitude, subjective norms and perceived control as antecedents of entrepreneurial intention. The analyses results show that the planned behavior theory (TPB) is applicable and consistent in explaining Sudan's academic entrepreneurship intention [14]. To investigate what elements, contribute to the long-term engagement of social HMS in health, Gu et al. integrated the planned behavior theory (TPB) with the big five theory and the theory of trust. Finally, users' openness to novel experience and perceived risk are also essential elements in determining prolonged health engagement, along with attitude, perceived behavior control, subjective norms, and trust [15].

Attitude has an effect on behavioral intention. Zhang et al. Also selected the planned behavior theory (TPB) and the attitude behavior condition theory (a-b-c) when investigating the influencing factors of residents' willingness to classify household waste (WCI) and waste classification behavior (WCB). Their analysis results show that there are significant differences between residents' WCI and WCB. Attitude, perceived behavior control and classification knowledge are significantly related to residents' WCI, among which attitude has the greatest direct impact [16]. Jinn et al. also based on the theory of planned behavior when analyzing the influencing factors of 233 Fencing Club members' willingness to continue to participate. The results show that attitudes and subjective behavior norms have a positive impact on the willingness to continue to participate [17]. Nikolaos et al. also took the planned behavior theory as a theoretical framework when investigate the savings behavior of pomace families. The results show that the factors that have a direct positive impact on the willingness to save and the final saving behavior are attitude, subjective norms and behavior control, respectively [18].

Subjective norms will also have an impact on Behavior Intention: Zhou et al. once built a measurement scale and theoretical model for the purchase intention of organic agricultural products based on the planned behavior theory, collected 600 effective questionnaires in major cities in China, and applied SPSS and Amos software to analyze the collected data. The empirical results show that the proposed model has good fitting. Product awareness, health and environmental protection awareness, emotional beliefs, behavior attitudes, subjective norms and perceived behavior control were significantly positively correlated with the purchase intention of organic agricultural products. Among them, subjective norms and behavior attitudes have the greatest impact on purchase intention [19]. Li et al. applied to observe the factors influencing Chinese consumers' intention to buy organic food based on the planned behavior theory (TPB) model. The results from the empirical analyses suggest that there is a statistically significant and positive relationship between attitude, subjective norms and perceived behavior control and Chinese consumers' willingness to buy organic food. In this analyses, subjective norms can better explain the difference in Chinese consumers' willingness to buy organic food than the other two forecastable variables [20]. Kamal et al. once discussed high school students' fast food consumption behavior based on the theory of planned behavior. It could be concluded that it is useful to take into account the potentially influential subjective norms (particularly friends) that may impact the high school students' inclination to use fast food. A greater emphasis must be placed on students' sense of agency in their actions [21].

Thorough behavioral control will also have an impact on Behavior Intention: Karimy et al. once anticipated the smoking behavior of Iranian medical college students based on the extended theory of planned behavior theory (TPB). The results also reveal that perceptual behavior control. All of the TPB characteristics were found to be effective predictors of students' smoking habit. A number of factors were identified as predictors of future smoking habit among students [22]. These included students' sense of agency over their actions and their perspective on smoking. Vasumathy et al. have identified the factors that encourage Indian students to engage in export entrepreneurship through empirical application of planned behavior theory. Mtenga et al. have determined the social cognitive factors related to married individuals receiving HCT based on TPB theory. The results showed that nearly half (42%) of the respondents had never suffered from HCT. In the social cognitive structure, the strongest predictor of HCT uptake was attitude, followed by perceptual behavior control [23, 24].

Attention will also affect behavioral intention: when discussing the communication effect of popular science short films, Xin Zeng set the analytical framework as the cognitive mediation model (CMM) of news learning theory. On this basis, combined with the dual coding theory of cognitive psychology, he finally mapped out the path of communication effect. The results show that at the cognitive level, sound and image content will play a positive correlation role in the audience's attention, and only thinking about the relevance of sound content will have a positive impact on the amount of knowledge the audience obtains. There are three aspects that can positively affect the audience's attitude towards science and technology: first, the motivation of monitoring satisfaction before contacting the popular science short film, second, the audience's influence on the information processing methods (attention and relevance thinking) of the short film content including sound and image, and third, the influence on the level of scientific and technological knowledge after the audience watches the popular science short film [25]. When investigate the impact of public knowledge and preventive behavior intention of the H1N1 pandemic in Singapore, who also started from three aspects: motivation, communication and news elaboration, the theoretical model used is also a cognitive intermediary model. The results show that in health situations, motivation (monitoring satisfaction, guidance and cognitive needs) and news attention, elaboration and interpersonal communication can be applied cognitive mediation model and are positively correlated. The aspects that can positively affect the public's understanding of H1N1 influenza are also news attention, elaboration and interpersonal communication. In addition, what has a significant indirect impact is the effect of motivation on behavioral intention, which is partially mediated by communication (media attention and interpersonal communication), elaboration and knowledge [26]. Lee et al. Applied the enhanced cognitive mediation model when analyzing the related factors of breast cancer knowledge among Singaporean women. They used random number dialing to conduct a nationally representative analysis of 802 women aged 30 to 70 [27]. The model's hypothetical course has strong backing. A higher risk perception is associated with increased media consumption [27]. The focus on these three forms of media is tied to interpersonal communication, while newspapers and television receive comparatively more emphasis and depth of coverage. Interpersonal communication is positively related to structural knowledge, while elaboration is related to both factual knowledge and structural knowledge [28]. The investigation found that media attention and knowledge dimensions have different indirect effects through interpersonal communication and elaboration [27].

2.3. Summary

In the past decades, (TPB) rational behavior theory has played an important role in discussing the influencing factors behind some rational behaviors, and has accomplished very good results. In particular, it has played a great role in the analyses on the impact of social media platforms on human behavior. Therefore, this theory can also be applied for reference when analyzing the impact of social media platforms on College Students' post-graduation planning (direct employment). They are all discussing the influencing factors behind a rational behavior, which is completely consistent with the structure of TPB theory. In addition, considering the particularity of the impact of social media platforms, they creatively added the influencing factors of media attention to the TPB model.

3. Method

3.1. Research Design

The quantitative analyses method can systematically measure a large number of samples, has a standardized measurement process, and the results are reliable. It is reliable, accurate and persuasive for exploring the social media for the planning (direct employment) of college students after graduation. It is also inspired by the previous analyses methods of researchers who selected similar theoretical models in this field. For example, Luo et al. applied a quantitative analysis method to analyze the influencing factors of the employment intention of university graduate students in Nanchang, conducted a questionnaire survey, collected data, and referred to the TPB theory, and accomplished good results [28]. After these two aspects of consideration, this paper chooses a quantitative analysis method to analyze the problem. When using the quantitative method, this paper first design the questionnaire. Then enter the pre investigation stage, in which 50 questionnaires are expected to be distributed. After recovery, SPSS software is applied to analyze the reliability and validity of the questionnaire. After analysis and test, the items that do not meet the standard are deleted. After re inspection, the formal investigation stage is officially entered. In the formal investigation stage, 100 questionnaires are expected to be distributed. After recovery, invalid samples are removed for further analysis and calculation of the data.

3.2. Date Collection

Based on Table 1, it can be seen that the scale consists of six parts, namely X1、X2、X3、X4、Y and basic information. In the first part of the basic information, this paper mainly unweather it is a social media platform user, region, grade and gender. Because the main respondents of this questionnaire are social media platform users among college students, samples that do not meet these two questions can be regarded as invalid samples and should be eliminated. The second part of the independent variable X can be divided into four categories, namely, the attitude revealed by X1, the subjective norms revealed by X2, the personal behavior control revealed by X3, and the media attention revealed by X4. The third part of the dependent variable Y revealed the intention of direct employment planning. In designing the contents of these two parts, the dimensions of the questionnaire were considered and the questionnaire items were designed inspired by the scale designed by Liang Jinhui when analyzing the driving factors of teenagers' participation in ice and snow sports in Beijing 2022 Winter Olympic Games [29]. Finally, after the questionnaire was designed, it was distributed through the questionnaire star platform. A total of 120 samples were distributed and 109 were recovered. After review, 9 questionnaires were eliminated and 100 valid questionnaires were finally obtained.

3.3. Date Analysis

Based on Table 2, among the 100 valid questionnaires, there are 34 males, accounting for 34%, and 66 females, accounting for 66%. They are college students, and are widely distributed in most provincial administrative regions of the country. After obtaining the valid questionnaire, analyze the frequency of the data and the reliability and validity of the data. If the test is correct, apply SPSS software to analyze the regression and correlation of the data. Finally, the obtained results are analyzed. First, the fitting with the model is obtained through the analysis of the square value of R. if the fitting is good, the significance of X is analyzed. If x has a significant impact on y, it is necessary to further analyze the specific degree of the impact of X1, X2, X3, X4 on Y, so as to write the formula of the model. And draw a conclusion.

3.4. Results

The results can be divided into four parts: reliability analysis, validity analysis, correlation coefficient and regression analysis.

Validity analysis is mainly applied to analyze the effectiveness of questionnaire design, accuracy, applied to measure whether the topic design is reasonable; The measurement result reflects the degree of investigation. The consistency between the measurement result and the content to be investigated (mainly for the scale). If KMO>0.7, there is a connection between the variables designed in the questionnaire, i.e., Table3. In these analyses, KMO is 0.841 greater than 0.7, indicating that the measurement result is highly consistent with the content to be investigated.

Reliability refers to the consistency or reliability of analyses results. Reliability analysis is mainly applied to measure whether the sample answer results in the questionnaire are reliable and whether they have answered carefully. The higher the test reliability is, the more reliable the results are. (To test the internal consistency between the questions, i.e., whether multiple questions measure the same content and concept. According to Table4, Table5, Table6, Table7 and Table8, the values of the reliability coefficients are greater than 0.7, indicating that the reliability quality of the analyses data is high. In conclusion, the reliability quality of the analyses data is high, and the questions in the scale can reflect the same content and can be applied for further analysis.

Correlation coefficient refers to the correlation coefficient between independent variables and dependent variables. It is applied to measure the correlation between independent variables and control variables. Regression analysis is applied to analyze the relationship between dependent related variables. For example, Table9, if the measured data is at the level of 0.01 (two tails), the correlation is significant. According to the above table, the significance (two tails) of X1, X2, X3, X4 and Y is less than 0.01, so the correlation is significant.

**. At level 0.01 (two-tailed), the correlation is significant.

This paper can further analyze the regression of the data. In the regression analysis results, R revealed the explanatory ability of the independent variable to the dependent variable. The closer R is to the adjusted R, the more stable the data is. For example, Table10explores the explanatory ability of the independent variable to Y, in which the R-square of the independent variable to Y is 0.597, showing a good explanatory ability, and the R-square after two adjustments is close to the R-square before adjustment, indicating that the data is stable. ANOVA is applied to verify the hypothesis. If the significance is less than 0.05, it indicates that the regression equation is meaningful. For example, Table11, the significance is less than 0.001, which indicates that the regression equation is meaningful. The coefficient table finally obtained by regression analysis shows the significance between the independent variable and the dependent variable. If the significance coefficient is less than 0.05, it means that the variable has statistical significance for the model. Otherwise, it means that the variable has no significant impact on the dependent variable. The size of the coefficient revealed the strength of the impact. According to Table12, it can be found that the significance coefficients of X1, X3, X4 and grades for y are less than 0.05, so X1, X3, X4 and grades have statistical significance for Y.

a. Predictors: (constant), X4, gender, grade, region, X1, X2, X3

4. Discussion

After regression analysis, it can be known that X1, X3, X4 and grade have a significant impact on Y, that is, the graduation planning (direct employment) personal behavior control, media attention of college students, Attitude and grade will have a great impact on it. According to the existing analyses results, it can be found that by investing more advertisements about employment on social media, providing more and better employment platforms, and giving college students more positive information about the employment situation, college students' direct employment intention can be improved.

At present, there is no much analyses on the social media platform's impact on College Students' future planning (employment). Most of the analyses on the social media platform focuses on advertising, public health, etc. Liu et al. are analyzing the impact of social media on College Students' Employment - Taking Weibo and Wechat as examples, they don’t apply the TPB model, there is no further analysis and Discussion on the influencing factors behind this problem by using the method of data analysis. However, they explored this issue in the form of questionnaires and interviews. Finally, by analyzing the survey results, they tried to put forward suggestions on the future development of social media in disseminating employment information. When they studied this issue, they were in 2014. At this time, the information and forms that social media platforms can provide for employment are still very limited compared with the current situation, and the development of social media platforms is not yet mature, Therefore, when exploring this issue, this paper did not pay attention to the specific influencing factors behind this phenomenon [30] one of the purposes of this analyses is to confirm whether the social media platform has an impact on the planning (employment) of college students after graduation, and this paper’s data also show that there is a positive relationship. According to the existing results, it can be found that what has a significant impact on College Students' post-graduation planning (direct employment) is the information related to college students' perceived behavior control on the social media platform, as well as their attitude towards direct employment, media attention and grades, which was unexpected before these analyses.

5. Conclusion

With the increasing popularity of higher education and the outbreak of the global epidemic, this article studies the post-graduation planning (direct employment) of "college students", and explores the impact of social media on the post-graduation planning (direct employment) of college students. Through reviewing the "TPB" theory, the questionnaire was distributed to college students. After recovery, the "SPSS" software was applied to analyze and investigate the reliability, validity, regression and correlation of the data. Finally, the social media platform has an impact on the post-graduation planning (direct employment) of college students, verifies the previous hypothesis, and finds that the social media platform has an impact on the post-graduation planning (direct employment) of college students through its impact on the "personal behavior control, attitude, media attention and year" of college students, and this impact shows a positive correlation. Therefore, in order to better solve the current employment problems of college students after graduation, this paper can provide more employment related information and internship employment platforms for college students through social media platforms, and push more popular science articles on improving the employability of college students, so as to enhance their employment willingness and provide more employment opportunities, so as to enhance their positive employment attitude, So as to better solve the employment problem.

This article once again verifies the TPB theory, verifies that the three factors contained in the "TPB" theory have an impact on rational behavior, and further verifies that the "TPB" theory is correct and necessary to supplement the "tar" theory in the "observed behavior control". This article also provides theoretical support for how to better solve the employment problems of college students after graduation, and puts forward suggestions on how to better promote college students to obtain employment with a positive employment mentality.

But at the same time, in the course of the survey, it may be affected by the general environment of the epidemic, and the survey is only conducted for Chinese college students, which can only reveal China's national conditions and is not representative in a broad sense. In addition, there is no good group division in the process of the questionnaire survey, and different grades of college students may have different understanding of employment, this paper should expand the investigation and analyses on the big three and big four groups in order to better grasp the attitude of college students towards direct employment. In the process of investigation, the factors of policy and region are not taken into account. There may be different "talent policies" and employment opportunities in different regions, which will also have an impact on the choice of direct employment of college students. When executing analyses in this area in the future, this paper can take policy and geographical factors into account in the "TPB" model, and increase the proportion of the survey on junior and senior students. In addition, this paper should further expand the scope of the survey to investigate the employment attitudes of university students all over the world.