What Corporate Social Responsibility (CSR) Disclosures Do Chinese Forestry Firms Make on Social Media? Evidence from WeChat

Abstract

Corporate social responsibility (CSR) disclosure serves as a vital bridge for forestry firms to communicate with their stakeholders and obtain legitimacy support. Existing studies focus on forestry firms’ CSR disclosures based on CSR reports but lack consideration of such disclosures on social media. In this study, based on WeChat, the most widely used social media platform in China, we obtained 3311 tweets from 36 WeChat Official Accounts (WOA) of 63 Chinese-listed forestry firms in 2018 and used content analysis to classify the CSR information involved in these tweets based on the stakeholder dimensions. The main analysis results show that the top three CSR dimensions disclosed by Chinese forestry firms in social media are the shareholder (28.21%), customer (26.20%), and employee (23.64%) dimensions, and there are also great differences in the subcontent of disclosure concerns in each stakeholder dimension, e.g., approximately 86% of CSR disclosures for customers are product and service information. Additionally, we conducted a content analysis on the CSR reports of forestry firms using WOA. The results show that firms express different concerns in CSR reports than on social media, and the most mentioned dimensions in their reports are the environment (23.69%), employees (20.91%), and shareholders (20.21%). This indicates that there is a significant difference between the stakeholders that Chinese forestry firms focus on in social media and those that they focus on in CSR reports. This paper is the first study to focus on the CSR disclosure of Chinese forestry firms in social media and provides a reference for scholars to understand the information activities of forestry firms in social media.

1. Introduction

With the launch of the UN 2030 Agenda for Sustainable Development, the United Nations proposed the Global Sustainable Development Goals (SDGs). The 17 SDGs aim to promote better environmental and social performance globally [1]. As the world’s largest emerging country, China is actively responding to global development visions and goals, implementing major strategies for carbon peaking and carbon neutrality, and actively assuming international responsibilities. The role of corporate social responsibility (CSR) in business operations and strategic development is becoming increasingly important [2,3,4]. CSR serves as an important bridge for firms to communicate with stakeholders, help firms to maintain their legitimacy, and obtain key resources. Thus, CSR disclosures are considered by firms to be the most important non-financial information activities [5]. Forests have an irreplaceable integrated value for sustainable economic and ecological development [6,7,8]; although forested areas only account for approximately 1/3 of the total land area, the carbon sink storage in forested vegetation areas accounts for almost half of the total terrestrial carbon pool. Forestry enterprises are microeconomic subjects of forestry industry development and are typical resource-sensitive and environment-sensitive enterprises, with the triple attributes of economic, social, and ecological benefits, which are of great significance for sustainable development [9,10].

According to the China Internet Network Information Center (Source: https://m.thepaper.cn/baijiahao_16888548, accessed on 7 June 2022), in 2021, the number of internet users in China was 1.032 billion, of which 1.029 billion were cellular data users, and the internet penetration rate had reached 73.0%. With previously unheard-of access to information and networks, many companies are finding a new voice in their interactions with consumers and other stakeholders. Among them, social media networks such as WeChat and Weibo are developing rapidly, replacing paper media and traditional online media to a large degree. Social media networks are also being used via an increasing number of firms to communicate for CSR [11,12,13,14], including forestry firms, who also use them as an important tool for CSR communication. Therefore, this paper conducts a study on CSR information disclosure on social media, which can serve as a new reference for the study of the CSR disclosure of forestry enterprises in various countries.

Our study aims to clarify the CSR disclosure of forestry firms on social media and draw on key differences between traditional media and social media. Related studies only focus on forestry firms’ CSR disclosures based on CSR reports but lack consideration of such disclosures on social media. This study is the first to present the viewpoint of the CSR disclosure of Chinese forestry firms in social media and provides a reference to understand the characteristics of the information activities of forestry firms in social media. Potential contributions of this study include: (1) applying a new optic angle to investigate CSR disclosure of forestry firms in China, and extending the point of potential associating and differences between CSR reports and social media; (2) updating and expanding previous study on CSR topic in China, meanwhile, gaining a broader investigation of CSR data from social media; and (3) as the largest emerging market, our study provides a reference understanding of CSR disclosure in other emerging economies for practitioners, managers, and researchers.

The remainder of this paper begins with the literature review and theoretical background about corporate disclosures on social media and CSR disclosure in the forestry industry. Then we turn to the research design, and based on the framework, we describe the variables. The content analysis framework also provides insight into forestry companies activities via a variety of scenarios regarding the level of CSR disclosure. We conclude with a summary of the results with practical implications, limitations, and questions for future research.

2. Literature Review and Theoretical Background

2.1. Corporate Disclosure on Social Media

Before the advent of social media, corporate disclosure on the internet was of interest to researchers [15,16,17]. However, corporate online information disclosure in the presocial media era still had some deficiencies, which are similar to those of traditional paper media; for example, information dissemination required review by authorities and third-party platforms and there was a lack of two-way communication with audiences.

With the birth of social media, internet information dissemination has changed dramatically [18]. (1) The “hierarchy” was broken; for example, while traditional media can be divided into “regional”, “national”, and “international” types, social media does not have such hierarchical properties. (2) The “elite” lost their control, namely, the major “information gatekeepers” (e.g., large newspaper company) lost their control on social media, while non-government organizations (NGOs), consumers, and other weak stakeholders more easily took over the initiative of information spread; (3) social media provides higher dynamism and timeliness, unlike the fixed nature of traditional media; (4) the cost of dissemination of information on social media is extremely low because the use of social media platforms is free or very inexpensive; (5) information users have a stronger sense of trust of social media because the information dissemination therein is more based on personal relationships and trust rather than commercial speech; (6) social media makes communication much more direct, specifically, it provides more direct dialogue between firms and stakeholders than traditional media; (7) there is a potential for higher rates of public response, namely, information dissemination on social media can obtain more public responses; (8) information on social media achieves wider ranges of diffusion, because social media is not limited to specific information-spreading channels but can spread throughout the social network; (9) information dissemination on social media is subject to less institutional control and involvement but wider public opinion scrutiny.

Social media thus has dramatically changed corporate information activities [19]. First, social media helps to promote the efficiency of corporate information activities and reduce the impact of negative information. For example, Blankespoor, et al. [20] and Prokofieva [14] suggested that social media helps the spread of earnings information and then helps to gain more attention from investors. Yang and Liu [21] found that firms can use social media to mitigate the impact of negative financial information in some ways. In addition, social media also helps stakeholders communicate with each other, which in turn reinforces the monitoring for firms. For example, Ang et al. [22] found that social media plays a complementary governance role in the Chinese market and that stakeholder discussions on social media can effectively curb inefficient corporate mergers and acquisitions.

Social media also plays an important role in CSR information activities [23]. First, social media strengthens CSR communication between firms and their stakeholders [24,25,26], significantly improves the perception and recognition of CSR among customers, employees, and other stakeholders [27,28,29,30], and helps transform CSR into intangible values such as improving corporate reputation [29,31]. It is worth noting that compared to traditional communication, CSR information activities on social media are more indicative, which facilitates impression management for firms. She and Michelon [12] found that firms selectively direct their stakeholders on social media to avoid the proliferation of negative news. Pizzi et al. [32] found that CSR disclosures on social media are selective and not always responsive to stakeholder claims. Russo et al. [33] argued that CSR-oriented firms use social media more as a tool to achieve a higher level of legitimacy than as a tool to manage their sustainability strategies and related performance.

2.2. CSR Disclosure in the Forestry Industry

Existing research on CSR disclosure in the forestry industry focuses on traditional reporting vehicles such as CSR reports. First, some studies focus on the information topics in traditional CSR reports for forestry firms. Vidal and Kozak [34] conducted a textual analysis of CSR reports of forestry firms worldwide and found that the proportion of environmental and social dimensions is increasing. Colaço and Simão [35] analyzed CSR disclosures for international forestry firms in the Congo Basin and found that the most frequently disclosed topics are organizational certification and the environment, and that the level of CSR disclosure is higher in firms from Western countries than in those from Asian and African countries. D’Amato et al. [36] analyzed CSR reporting on a global scale involving land use (including forestry firms’ CSR reports) and found that its focus was on the sustainability concepts of the circular, green, and bio-economy. Wang and Juslin [37] studied the CSR reports of Chinese forestry enterprises and argued that the most mentioned topic in these reports is economic responsibility. The driving factors of the CSR disclosures of forestry enterprises have also received great interest from scholars. Hansen, et al. [38], Panwar, et al. [39], and Li, et al. [40] studied the drivers of CSR disclosure in cross-country forestry firms based on globalization and firm-level factors. Wang and Juslin [37] analyzed the driving role of managerial characteristics in CSR disclosures. Li, Toppinen, Tuppura, Puumalainen and Hujala [40], and Lu et al. [41] explored the factors influencing CSR disclosure for Chinese forestry firms, focusing on the firm-level factors that characterize forestry firms’ CSR disclosure. Lu et al. [42] conducted a study on the impact of managers’ risk awareness on CSR disclosure in Chinese forestry firms. In summary, the research on forestry CSR disclosure only focuses on traditional information vehicles such as CSR reports, and no one discusses their CSR disclosure in emerging social media platforms. Specifically, it is worthwhile studying which kind of topics are preferred by forestry firms in the CSR reporting of social media and the difference between CSR disclosure in social media and traditional reports.

3. Research Design

3.1. Data Sources

As the largest developing economy, China has a huge forestry industry. According to a report by the China National Forestry and Grassland Administration (Source: http://www.forestry.gov.cn/main/62/20171221/1086586.html, accessed on 10 June 2022), gross output value of forestry reached 6.49 trillion RMB in 2016 and is expected to reach 9 trillion RMB by 2025 (Source: https://news.cgtn.com/news/2022-02-16/China-s-forestry-industry-output-to-reach-9t-yuan-by-2025-17He4nAk6li/index.html, accessed on 10 June 2022). Thus, Chinese forestry firms have a strong representation.

In China, due to regulatory policies, Facebook and Twitter are not available, and similar mainstream social media platforms are WeChat and Weibo. After consideration, we selected the WeChat platform for study. The reasons are as follows:

First, WeChat allows a higher information capacity for users and allows them to provide more substantive information. Compared with Weibo, which requires users to limit their tweets to 140 characters, WeChat does not have such limitations.

Second, WeChat has the largest user base and extremely high user stickiness in China. Its daily logged-in users reached 900 million in 2017, and the monthly active official account (The WeChat platform provides two types of account services: first, a public account, which allows all users to view the tweets of the account at will; the second category, a personal account, is only used by individuals who only want to allow their friends to view the tweets of their account) reached 3.5 million, with nearly 800 million monthly active followers (As a comparison, Weibo had only 376 million monthly active users in 2017 according to the Q3 2017 Sina earnings report. Source: https://www.sohu.com/a/203437993_667510, accessed on 10 June 2022). Moreover, WeChat provides an electronic payment service that is the most widely used by Chinese people (According to WeChat user report in 2018 year, its payment services get 600 million active users. Source: https://www.sohu.com/a/278496021_506058x, accessed on 10 June 2022), and furthermore, creates a high level of dependency and a long-term habit for its users. As a result, the information shared by enterprises through the WeChat Official Account (WOA) will be received by users more reliably and frequently.

The original sample is 63 listed firms in the China forest industry in 2018, and we excluded 27 samples that did not use the WOA or who had used it for less than 1 year. Therefore, there are 36 sample firms included in the analysis. Appendix B presents a list of the sample firms involved in WeChat. Moreover, 11 of 36 forestry firms used in this study also issued CSR reports in that year, which provided us with an opportunity to compare their CSR disclosures in social media channels and traditional channels. We used Python to obtain the WOAs of forestry firms, totaling 3311 articles. CSR reports are sourced from Cninfo (http://www.cninfo.com.cn). Financial and other firm-level data were obtained from CSMAR.

3.2. Content Analysis

We used the content analysis system created by Lu, Kozak, Toppinen, D Amato, and Wen [41] (Lu, Kozak, Toppinen, D Amato and Wen [39] design a content analysis system for Chinese forestry enterprises based on the guidance of CASS 3.0 and CNFPIA 2.0, which are released by Chinese Academy of Social Sciences and Chinese National Forestry Products Industry Association), which is for Chinese forestry firms, to categorize the tweets and CSR reports.

The system of Lu, Kozak, Toppinen, D Amato, and Wen [41] contains seven level-1 dimensions based on stakeholder theory: environment, customer, employee, supplier, community, government, and shareholder. Each level-1 dimension contains five to thirteen level-2 dimensions; for example, the employee dimension contains a total of seven level-2 dimensions: “Abidance by rule and laws (EM1)”, “Percent of contract signing (EM2)”, “Coverage of social insurance (EM3)”, “Equal employment institution (EM4)”, “Staff development training (EM5)”, “Occupational health and safe producing (EM6)”, “Staff relation management (EM7)” and “Other employee-related (EM0)”. We scored according to the level-2 dimensions of this system; specifically, when the qualitative information of this dimension is provided in the tweet or CSR report, we scored “1”, and when the quantitative information is provided, the score was “2”. Then, we aggregated the scores of all level-2 dimensions to form level-2 variables. Finally, the level-2 dimension variables were aggregated to level-1 dimensions to form level-1 variables. Based on the above steps, we finally made 16 level-1 variables (eight variables for analysis on tweets and eight variables for analysis on CSR reports, variable definitions are reported in

Table 1. Variable definitions.

Variable	Level Annotation	Definition
Content variables for CSR disclosure on social media
WeChat_All	Summary of level-1 variable for social media	Total level of CSR disclosure on social media, equal to the sum of seven level-1 content variable for tweets.
ep_w	Level-1 variable for social media	The level of the social media environment dimension CSR disclosure, equal to the sum disclosure scores of all subdimensions of the environment on social media.
cu_w	Level-1 variable for social media	The level of the social media customer dimension CSR disclosure, equal to the sum disclosure scores of all subdimensions of the customer on social media.
em_w	Level-1 variable for social media	The level of the social media employee dimension CSR disclosure, equal to the sum disclosure scores of all subdimensions of the employee on social media.
su_w	Level-1 variable for social media	The level of the social media supplier dimension CSR disclosure, equal to the sum disclosure scores of all subdimensions of the supplier on social media.
co_w	Level-1 variable for social media	The level of the social media community dimension CSR disclosure, equal to the sum disclosure scores of all subdimensions of the community on social media.
go_w	Level-1 variable for social media	The level of the social media government dimension CSR disclosure, equal to the sum disclosure scores of all subdimensions of the government on social media.
sh_w	Level-1 variable for social media	The level of the social media shareholder dimension CSR disclosure, equal to the sum disclosure scores of all subdimensions of the shareholder on social media.
ep1_w/ep2_w/ep4_w/ep5_w/ep6_w/ep7_w/ep8_w/ep9_w/ep10_w/ep11_w/ep12_w/ep0_w	Level-2 variable for social media	The level of environment disclosure of the level-2 subdimension on social media, equal to the sum of the scores of the information of the level-2 subdimension of the environment. For example, the variable ep1_w is equal to the sum of the qualitative and quantitative scores of the level-2 subdimension indicator EP1 “Environment management system”. The description of the level-2 subdimension indicators are shown in
cu1_w/cu2_w/cu4_w/cu6_w/cu0_w	Level-2 variable for social media	The level of customer disclosure of the level-2 subdimension on social media, equal to the sum of the scores of the information of the level-2 subdimension of the customer. For example, the variable cu1_w is equal to the sum of the qualitative and quantitative scores of the level-2 subdimension indicator CU1 “Product quality management system”. The description of level-2 subdimension indicator are shown in
em1_w/em3_w/ep4_w/em5_w/ep6_w/ep7_w/em0_w	Level-2 variable for social media	The level of employee disclosure of the level-2 subdimension on social media, equal to the sum of the scores of the information of the level-2 subdimension of the employee. For example, the variable em1_w is equal to the sum of the qualitative and quantitative scores of the level-2 subdimension indicator EM1 “Abidance by rule and laws”. The description of level-2 subdimension indicators are shown in
su1_w/su4_w/su5_w/su6_w/su0_w	Level-2 variable for social media	The level of supplier disclosure of the level-2 subdimension on social media, equal to the sum of the scores of the information of the level-2 subdimension of the supplier. For example, the variable su1_w is equal to the sum of the qualitative and quantitative scores of the level-2 subdimension indicator SU1 “Responsibility purchasing system”. The description of level-2 subdimension indicators are shown in
co1_w/co4_w/co6_w/co0_w	Level-2 variable for social media	The level of community disclosure of the level-2 subdimension on social media, equal to the sum of the scores of the information of the level-2 subdimension of the community. For example, the variable co1_w is equal to the sum of the qualitative and quantitative scores of the level-2 subdimension indicator CO1 “The effect of enterprise operation on community”. The description of level-2 subdimension indicators are shown in
go1_w/go2_w/go4_w/go0_w	Level-2 variable for social media	The level of government disclosure of the level-2 subdimension on social media, equal to the sum of the scores of the information of the level-2 subdimension of the government. For example, the variable go1_w is equal to the sum of the qualitative and quantitative scores of the level-2 subdimension indicator GO1 “Enterprise management abided by rule”. The description of level-2 subdimension indicators are shown in
sh1_w/sh2_w/sh3_w/sh4_w/sh0_w	Level-2 variable for social media	The level of shareholder disclosure of the level-2 subdimension on social media, equal to the sum of the scores of the information of the level-2 subdimension of the shareholder. For example, the variable sh1_w is equal to the sum of the qualitative and quantitative scores of the level-2 subdimension indicator SH1 “Investor relation management”. The description of level-2 subdimension indicators are shown in
Content variables for CSR disclosure on CSR report
Report_All	Summary of level-1 variable for the CSR report	Total level of CSR disclosure on the CSR report, equal to the sum of the seven level-1 content variables for the CSR report.
ep_r	Level-1 variable for CSR report	The level of CSR report environment dimension CSR disclosure, equal to the sum disclosure scores of all subdimensions of the environment on social media.
cu_r	Level-1 variable for CSR report	The level of the CSR report customer dimension CSR disclosure, equal to the sum disclosure scores of all subdimensions of the customer on social media.
em_r	Level-1 variable for CSR report	The level of the CSR report employee dimension CSR disclosure, equal to the sum disclosure scores of all subdimensions of the employee on social media.
su_r	Level-1 variable for CSR report	The level of the CSR report supplier dimension CSR disclosure, equal to the sum disclosure scores of all subdimensions of the supplier on social media.
co_r	Level-1 variable for CSR report	The level of the CSR report community dimension CSR disclosure, equal to the sum disclosure scores of all subdimensions of the community on social media.
go_r	Level-1 variable for CSR report	The level of the CSR report government dimension CSR disclosure, equal to the sum disclosure scores of all subdimensions of the government on social media.
sh_r	Level-1 variable for CSR report	The level of the CSR report shareholder dimension CSR disclosure, equal to the sum disclosure scores of all subdimensions of the shareholder on social media.

) and 42 level-2 variables (Since our interest of this paper is not the content of the CSR report, we only use level-2 subdimension variables for CSR disclosure on social media, and no longer use level-2 variables for the CSR report). The scoring index system and examples are reported in Appendix A. The software used for the content analysis is Atlas.ti 8.0.

3.3. Descriptions of Variables

Table 1 reports the variables used in the analysis of this paper:

(1) Content variables for CSR disclosure on social media. These include the total social media disclosure variable WeChat_All, social media environment dimension disclosure variable ep_w, social media customer dimension disclosure variable cu_w, social media employee dimension disclosure variable em_w, social media supplier dimension disclosure variable su_w, social media community dimension disclosure variable co_w, social media government dimension disclosure variable go_w, and social media shareholder dimension disclosure variable sh_w. Moreover, we defined forty-two level-2 subdimension variables, e.g., variable ep1_w stands for the reporting level of the social media environment EP1 subdimension (Some of the level-2 subdimensions indicators of [41], 10 dimensions in all, among them the forest biodiversity conservation dimension EP3 and dispute resolution mechanism dimension CU3, are not addressed in the social media disclosure, so we do not create variables for these dimensions).

(2) Content variables for CSR disclosures in CSR reports. These also include eight variables: the total CSR report disclosure variable Report_All, CSR report environment dimension disclosure variable ep_r, CSR report customer dimension disclosure variable cu_r, CSR report employee dimension disclosure variable em_r, CSR report supplier dimension disclosure variable su_r, CSR report community dimension disclosure variable co_r, CSR report government dimension disclosure variable go_r, and CSR report shareholder dimension disclosure variable sh_r.

4. Results

4.1. Content Analysis of CSR Disclosures on Social Media

4.1.1. Total and Level-1 Dimension Analysis on Social Media

Table 2. Results of total and level-1 dimension analysis on social media.

	Variable	N	Mean	Ratio	Sd	Min	P50	Max
(1)	WeChat_All	36	103.3	100.00%	97.87	2	72.5	470
(2)	ep_w	36	8.14	7.88%	12.14	0	2	50
(3)	cu_w	36	27.06	26.20%	39.61	0	8.5	171
(4)	em_w	36	24.42	23.64%	40.2	0	5	218
(5)	su_w	36	0.64	0.62%	1.33	0	0	6
(6)	co_w	36	5.17	5.00%	6.35	0	3	27
(7)	go_w	36	8.78	8.50%	14.17	0	3	67
(8)	sh_w	36	29.14	28.21%	36.42	0	18.5	141

reports the descriptive statistics of the overall and level-1 dimension variables of social media CSR disclosure for 36 forestry firms. The mean value of the overall disclosure variable WeChat_All is 103.3, and the median value is 72.5. Rows (2) to (8) of Table 2 show the statistics of the environment dimension variable ep_w, customer dimension variable cu_w, employee dimension variable em_w, supplier dimension variable su_w, community dimension variable co_w, government dimension variable go_w, and shareholder dimension variable sh_w, where the top three highest disclosure dimensions are shareholder, customer, and employee, respectively. The mean (median) values of the shareholder dimension variable sh_w, customer dimension variable cu_w, and employee dimension variable em_w are 29.14 (18.5), 27.06 (8.5), and 24.42 (5), respectively. The last four variables are the government dimension variable go_w, environment dimension variable ep_w, community dimension variable co_w, and supplier dimension variable su_w, with mean (median) values of 8.78 (3), 8.14 (2), 5.17 (3), and 0.64 (0), respectively. Column (4) shows the proportion of subdimension disclosure to total disclosure. Among them, the top three dimensions with the highest proportion of disclosure are shareholders, customers and employees, with 28.21%, 26.20%, and 23.64% (The information proportion of level--1 subdimension is equal to the mean value of the subdimension indicator divided by the mean value of the total rating indicator (WeChat_All), for example, the mean value of the environment dimension variable ep_w is 8.14, divided by the mean value of the total rating indicator WeChat_All, 103.3, which equals 7.88%), respectively, which cumulatively account for approximately 80% of the total CSR disclosure on social media. The above results show that the stakeholders that forestry firms care most about in their social media CSR disclosures are shareholders, customers, and employees, and these three stakeholders account for approximately 80% of the disclosures.

4.1.2. Analysis of the Level-2 Environment Subdimension Analysis on Social Media

Table 3. Results of environment level-2 subdimension analysis on social media.

	Variable	N	Mean	Ratio	Sd	Min	P50	Max
(1)	ep_w	36	8.14	100.00%	12.14	0	2	50
(2)	ep1_w	36	0.67	8.23%	1.2	0	0	5
(3)	ep2_w	36	0.08	0.98%	0.5	0	0	3
(4)	ep4_w	36	1.25	15.36%	3.15	0	0	12
(5)	ep5_w	36	0.14	1.72%	0.59	0	0	3
(6)	ep6_w	36	0.19	2.33%	0.52	0	0	2
(7)	ep7_w	36	0.11	1.35%	0.4	0	0	2
(8)	ep8_w	36	2.19	26.90%	3.4	0	1	12
(9)	ep9_w	36	0.72	8.85%	2.01	0	0	11
(10)	ep10_w	36	0.94	11.55%	2.29	0	0	10
(11)	ep11_w	36	0.28	3.44%	0.74	0	0	3
(12)	ep12_w	36	0.42	5.16%	1.13	0	0	6
(13)	ep0_w	36	1.14	14.00%	2.18	0	0	10

reports the results of the environment level-2 subdimension analysis on social media. Row (1) of Table 3 shows the statistical results of environment level-1 variable ep_w, which are also reported in Row (2) of Table 2. Rows (2) to (13) show the statistics for environment level-2 subdimension variables, including ep1_w (see Table 1 for details). Among them, the top four level-2 subdimensions with the highest proportions in social media environment disclosure are the “Research, development, application, and sale of the environment production and devices” variable ep8_w, the “Environmental protection investment” variable ep4_w, the “Other environment-related” variable ep0_w, and the “Reduce pollution and decrease drain” variable ep10_w, with mean values of 2.19, 1.25, 1.14, and 0.94, respectively, and the last four are the “Forest certification” variable ep6_w, the “Sustainable forest management” variable ep5_w, the “Quantity, kind, and risk to human and environment of toxic or exhaust emission” variable ep7_w, and the “Environmental impact assessment of new investment projects” variable ep2_w. Column (4) shows the proportion of level-2 subdimensions to the total disclosure of environment dimensions, e.g., the social media CSR disclosure of the “Environmental management system (EP1)” dimension accounts for 8.23% of the environmental dimension disclosure. Among the twelve subdimensions, the top three ones with the highest percentages are the “Research, development, application, and sale of the environment production and devices dimension (EP8)”, the “Environmental protection investment dimension (EP4)”, and the “Other environment-related dimension (EP0)”, with 26.90%, 15.36%, and 14.00%, respectively. These three types of level-2 subdimensions account for approximately 60% of the total information in the social media environment disclosure.

4.1.3. Analysis of the Level-2 Customer Subdimension Analysis on Social Media

Table 4. Results of the level-2 customer subdimension analysis on social media.

	Variable	N	Mean	Ratio	Sd	Min	P50	Max
(1)	cu_w	36	27.06	100.00%	39.61	0	8.5	171
(2)	cu1_w	36	0.81	2.99%	1.69	0	0	8
(3)	cu2_w	36	0.39	1.44%	1.4	0	0	8
(4)	cu4_w	36	23.22	85.81%	37.3	0	5	169
(5)	cu6_w	36	0.33	1.22%	1.2	0	0	6
(6)	cu0_w	36	2.31	8.54%	3.54	0	0.5	13

reports the results of the level-2 customer subdimension analysis on social media. Similarly, we report the level-1 customer variable cu_w in Row (1) of Table 4. Rows (2) to (6) of Table 4 show the statistics of level-2 customer variables, including cu1_w. The highest is the “Information provision of the product and services” variable cu4_w, with a mean value of 23.22. Column (4) shows the proportion of level-2 subdimensions to the total disclosure of customer dimensions, e.g., the “Product quality management system (CU1)” subdimension accounts for 2.99% of the whole customer dimension. Among all level-2 customer subdimensions, the highest percentage belongs to the “Information provision of the product and services subdimension (CU4)”, whose percentage is 85.81%, accounting for most of the customer dimension disclosures. The above results show that when disclosing customer dimension information on social media, forestry firms prefer to disclose information under the subdimension “Information provision of the product and services (CU4)”.

4.1.4. Analysis of the Level-2 Employee Subdimension on Social Media

Table 5. Results of the analysis of the level-2 employee subdimension on social media.

	Variable	N	Mean	Ratio	Sd	Min	P50	Max
(1)	em_w	36	24.42	100.00%	40.2	0	5	218
(2)	em1_w	36	0.08	0.33%	0.37	0	0	2
(3)	em3_w	36	0.17	0.70%	0.74	0	0	4
(4)	em4_w	36	0.06	0.25%	0.33	0	0	2
(5)	em5_w	36	2.58	10.57%	6.81	0	0	39
(6)	em6_w	36	3.19	13.06%	6.06	0	0	31
(7)	em7_w	36	18.25	74.73%	28.33	0	4.5	148
(8)	em0_w	36	0.08	0.33%	0.37	0	0	2

reports the results of the analysis of the level-2 employee subdimension on social media. Row (1) of Table 5 shows the statistics of the level-1 variable em_w. Rows (2) to (8) of Table 5 show seven employee level-2 variables. Among them, the top three variables with the highest scores are “Staff relation management” variable em7_w, “Occupational health and safe producing” variable em6_w, and “Staff development training” variable em5_w, with mean values of 18.25, 3.19, and 2.58, respectively. Column (4) shows the proportion of subdimensions of level-2 disclosure to the total employee dimension disclosure. Among the seven employee subdimensions, the highest disclosure subdimension is the “Staff relation management dimension (EM7)”, whose disclosure ratios are 74.73%, accounting for approximately three quarters of the employee dimension disclosure. The above results show that when disclosing employee dimension information on social media, forestry firms prefer to disclose information on the “Staff relation management dimension (EM7)”.

4.1.5. Analysis of the Level-2 Supplier Subdimension on Social Media

Table 6. Results of the analysis of the level-2 supplier subdimension on social media.

	Variable	N	Mean	Ratio	Sd	Min	P50	Max
(1)	su_w	36	0.64	100.00%	1.33	0	0	6
(2)	su1_w	36	0.03	4.69%	0.17	0	0	1
(3)	su4_w	36	0.14	21.88%	0.42	0	0	2
(4)	su5_w	36	0.11	17.19%	0.4	0	0	2
(5)	su6_w	36	0.03	4.69%	0.17	0	0	1
(6)	su0_w	36	0.33	51.56%	1.01	0	0	5

reports the results of the analysis of the level-2 supplier subdimension on social media. Row 1 shows the statistics of the level-1 variable su_w. Rows (2) to (6) show the results of level-2 supplier variables, in which the top three highest disclosure positions are the “Other supplier-related” variable su0_w, the “Openness of procurement policy” variable su4_w, and the “Legality of forest product procurement” variable su5_w, with mean (median) values of 0.33 (0), 0.14 (0), and 0.11 (0), respectively. Column (4) shows the percentages of the level-2 supplier dimension. The three highest supplier subdimensions are the “Other supplier-related dimension (SU0)”, the “Openness of procurement policy dimension (SU4)”, and the “Legality of forest product procurement dimension (SU5)”, whose disclosure proportions were 51.56%, 21.88%, and 17.19%, respectively, accounting for approximately 90% of the supplier dimension disclosure. The above results indicate that on social media, forestry firms prefer to disclose information on the “Other supplier-related dimension (SU0)”, the “Openness of procurement policy dimension (SU4)”, and the “Legality of forest product procurement dimension (SU5)”, and that these three subdimensions account for approximately 90% of the supplier dimension cumulatively.

4.1.6. Analysis of the Level-2 Community Subdimension on Social Media

Table 7. Results the analysis of the level-2 community subdimension on social media.

	Variable	N	Mean	Ratio	Sd	Min	P50	Max
(1)	co_w	36	5.17	100.00%	6.35	0	3	27
(2)	co1_w	36	3.31	64.02%	4.08	0	2	14
(3)	co4_w	36	0.56	10.83%	1.83	0	0	10
(4)	co6_w	36	0.67	12.96%	1.41	0	0	6
(5)	co0_w	36	0.64	12.38%	1.64	0	0	9

reports the results of the analysis of the level-2 community subdimension on social media. Row (1) of Table 7 shows the level-1 community dimension variable co_w. Rows (2) to (5) of Table 7 show the statistics for level-2 community variables. Among them, the highest score is the “Effect of enterprise operation on community” variable co1_w, whose mean value is 3.31. Column (4) shows the proportion of level-2 subdimensions to the level-1 community dimension. The “Effect of enterprise operation on community dimension (CO1)” accounts for 64.02% of the disclosure of community dimensions. The above results indicate that forestry firms are more likely to disclose information on the subdimension “Effect of enterprise operation on community (CO1)” on social media.

4.1.7. Analysis of the Level-2 Government Subdimension on Social Media

Table 8. Results of the analysis of the level-2 government subdimension on social media.

	Variable	N	Mean	Ratio	Sd	Min	P50	Max
(1)	go_w	36	8.78	100.00%	14.17	0	3	67
(2)	go1_w	36	1.86	21.18%	3.21	0	0	13
(3)	go2_w	36	0.33	3.76%	0.99	0	0	5
(4)	go4_w	36	0.11	1.25%	0.46	0	0	2
(5)	go0_w	36	6.47	73.69%	11.22	0	1.5	52

reports the results of the analysis of the level-2 government subdimension. Row (1) of Table 8 shows the statistics of the total government dimension disclosure variable go_w. Rows (2) to (5) of Table 8 show all level-2 subdimensions of government, where the top two highest disclosures are the “Other government-related” variable go0_w and the “Enterprise management abided by rule” variable go1_w, and their mean values are 6.47 and 1.86, respectively. Column (4) shows the proportion of the level-2 subdimensions of the government dimensions. The top two highest disclosure percentages are the “Other government-related dimension (GO0)” and the “Enterprise management abided by rule dimension (GO1)”, whose disclosure percentages are 73.69% and 21.18%, respectively, accounting for approximately 95% of the government dimension disclosures.

4.1.8. Analysis of the Level-2 Shareholder Subdimension on Social Media

Table 9. Results of shareholder level-2 subdimension analysis on social media.

	Variable	N	Mean	Ratio	Sd	Min	P50	Max
(1)	sh_w	36	29.14	100.00%	36.42	0	18.5	141
(2)	sh1_w	36	4.56	15.65%	7.58	0	2.5	42
(3)	sh2_w	36	12.89	44.23%	16.59	0	7.5	68
(4)	sh3_w	36	2.86	9.81%	4.52	0	1	20
(5)	sh4_w	36	1.56	5.35%	3.36	0	0	13
(6)	sh0_w	36	7.28	24.98%	10.06	0	4.5	41

reports the results for the shareholder dimension. Row (1) of Table 9 shows the statistical results of the total shareholder dimension variable sh_w. Rows (2) to (6) of Table 9 show the level-2 shareholder variables. Among them, the top three highest disclosures are the “Growth potential” variable sh2_w, the “Other shareholder-related” variable sh0_w, and the “Investor relation management” variable sh1_w, with mean values of 12.89, 7.28, and 4.56, respectively. Column (4) shows the proportion of the level-2 subdimensions. Among the five shareholder subdimensions, the top three highest percentages are the “Growth potential dimension” (SH2), “Other shareholder-related dimension” (SH0), and “Investor relation management dimension” (SH1), whose disclosure percentages are 44.23%, 24.98%, and 15.65%, respectively, accounting for approximately 85% of the shareholder dimension disclosure.

4.1.9. Summary of the Analysis of CSR Disclosures on Social Media

First, forestry firms pay significant attention to the interests of different stakeholders on social media. The three most concerned stakeholders on social media CSR disclosure by forestry firms are shareholders (28.21%), customers (26.20%), and employees (23.64%), and these three categories account for approximately 80% of all disclosures.

Moreover, under the level-1 dimension of each stakeholder, forestry firms also show serious disclosure imbalance for different level-2 subdivisions. Specifically, the three level-2 subdimensions with the highest levels of disclosure of the environmental dimension are the “Research, development, application, and sale of the environment production and devices (EP8)” (26.90%), the “Environmental protection investment dimension (EP4)” (15.36%), and the “Other environment-related dimension (EP0)” (14.00%), which account for approximately 60% of the environmental dimension disclosures. The highest reporting subdimensions of the customer and employee dimensions are the “Information provision of the product and services dimension (CU4) (85.81%) and the “Staff relation management” dimension (EM7) (74.73%). The three highest subdimensions in the supplier dimension are the “Other supplier-related dimension (SU0)” (51.56%), the “Openness of procurement policy dimension (SU4)” (21.88%), and the “Legality of forest product procurement dimension (SU5)” (17.19%), accounting for approximately 90% of the supplier dimension disclosures. The highest community subdimension is the “Effect of enterprise operation on community dimension (CO1)” (64.02%). The highest governmental subdimensions are the “Other government-related dimension (GO0)” (73.69%) and the “Enterprise management abided by rule dimension (GO1)” (21.18%), accounting for approximately 95% of the government dimension. The three highest disclosed subdimensions of the shareholder dimension are the “Growth potential dimension (SH2)” (44.23%), the “Other shareholder-related dimension (SH0)” (24.98%), and the “Investor relation management dimension (SH1)” (15.65%), accounting for approximately 85% of the shareholder dimension.

4.2. Comparative Analysis of CSR Disclosure between Social Media and CSR Reports

4.2.1. Total and Level-1 Dimension Analysis on CSR Reports

Table 10. Results of total and level-1 dimension analysis on CSR reports.

	Variable	N	Mean	Ratio	Sd	Min	P50	Max
(1)	Report_All	11	26.09	100%	7.739	20	23	46
(2)	ep_r	11	6.18	23.69%	3.156	2	5	12
(3)	cu_r	11	2.27	8.71%	1.737	0	2	6
(4)	em_r	11	5.46	20.91%	2.339	3	5	10
(5)	su_r	11	1.27	4.88%	1.104	0	1	3
(6)	co_r	11	2.91	11.15%	1.3	1	4	4
(7)	go_r	11	2.73	10.45%	2.37	0	2	6
(8)	sh_r	11	5.27	20.21%	2.453	2	6	8

reports the descriptive statistics of the total and level-1 dimension variables of the CSR report disclosures for the forestry firms using the WeChat official account (11 of 36 listed forestry firms that use WeChat official account issued the CSR reports in 2018 year). Row (1) of Table 10 shows the statistics of the total disclosure variable Report_All with mean and median values of 26.09 and 23, respectively. Rows (2) to (8) of Table 10 show the statistics of the environment dimension variable ep_r, customer dimension variable cu_r, employee dimension variable em_r, supplier dimension variable su_r, community dimension variable co_r, government dimension variable go_r, and shareholder dimension variable sh_r, where the top three level-1 dimension variables are the environmental variable ep_r, the employee variable em_r, and the supplier variable su_r, with mean (median) values of 6.18 (5), 5.46 (5), and 5.27 (6), respectively. Moreover, their percentages of disclosure are 23.69%, 20.91%, and 20.21%, which in total account for approximately 65% of all disclosures of CSR reports.

4.2.2. Comparison between CSR Disclosure under Social Media and CSR Reports

In Figure 1, we compare forestry firms’ disclosures on social media and CSR reports by stakeholder dimensions.

First, there are differences in the stakeholder dimensions that forestry firms focus on in different channels. The top three dimensions in CSR report channels are the environment (23.69%), employee (20.91%), and shareholder (20.21%) dimensions, while the top three dimensions in social media channels are the customer (32.40%), shareholder (24.70%), and employee (22.00%) dimensions. The dimension with the highest proportion disclosed in the report is the environment dimension (23.69%), but this dimension only ranks 5th on social media (7.24%), with a difference of more than three times, while the customer dimension, which ranks 6th in the CSR report in terms of proportion (8.71%) and 1st on social media (32.40%), has the same difference of more than three times. We suggest that this difference may be due to the characteristics of the information users focused on by firms in different channels: forestry firms are mostly in heavily polluting and environmentally sensitive industries, so in formal reports (CSR reports), they take more into account the needs of professional information users, such as providing more detailed and rich environmental information to analysts, socially responsible investors (SRIs), fund managers, etc. However, in social media, forestry firms are more likely to have to communicate with non-specialist users, e.g., local customers.

Second, there are also commonalities in the stakeholder dimensions that forestry firms focus on under different channels. For example, in both reporting channels, the lowest proportion of disclosure is the supplier dimension (with CSR reports at 4.88% and social media at 0.87%). This is because forestry firms mostly belong to the upstream of the industry chain or self-supply their raw materials, such as paper, wood supply, furniture manufacturing, and other niche industry firms, and suppliers have little influence on them. In addition, the proportion of community dimension and government dimension is also small, and the proportion in the CSR reports and social media is mostly around 10% or lower. This indicates that for Chinese forestry firms, these two types of stakeholders are not the focus of corporate concern.

5. Discussion and Conclusions

This paper studies the CSR disclosures made by Chinese forestry firms on social media. Based on the framework of Lu, Kozak, Toppinen, D Amato, and Wen [41], we used content analysis to analyze the social media posted by Chinese-listed forestry firms on WeChat in 2018 and found that (1) there are differences for Chinese forestry firms that care about stakeholders in social media and traditional CSR report channels; for example, the top three dimensions present in the social media channel are shareholder, customer, and employee dimensions, while the top three dimensions present in the traditional CSR report channel are environment, employee, and shareholder dimensions; (2) the proportion of disclosed subcontent for each stakeholder on social media also shows a great imbalance; for example, the most concerning subdimension in the disclosure for the shareholder dimension is the “Growth potential” subdimension, which accounted for 44.23% of the shareholder’s information. The results of our analysis suggest that the characteristics of social media users lead to distinctive features of CSR disclosures made by forestry firms on social media, with significant differences in the main points of attention from disclosures made through traditional channels (CSR reports). We argue that the main reasons for the differences in the focus of information disclosure between forestry firms in social media and traditional CSR reports are as follows: the target information users of CSR reports are professional users, e.g., analysts, institutional investors, and financial media; however, the general public (e.g., individual consumers, grassroots employees, and community residents) constitute the largest proportion of stakeholders but are less capable of using information, specifically, they lack intuitive judgment of the financial indicators [43] and industrial technical indicators recorded in traditional reports, and are more concerned with intuitive information such as protection of consumer rights and demonstration of corporate strengths, so forestry firms are more interested in displaying such intuitive information in social media than in CSR reports.

The following questions of this paper need future research: first, the WeChat platform cannot fully display all the comments because of its program design (According to the rules of the WeChat platform, users’ comments on WOA need to be reviewed by the account owner before they are made public. As a result, negative comments about the firm are filtered out), which leads us to be unable to analyze the interactive communication between firms and information users; second, information activities conducted by firms on the WeChat platform use a large number of pictures and videos, and Kassinis and Panayiotou [44] found that visualization tools play an important role in CSR impression management by firms, which is limited by technical means, making it difficult for us to include these images and videos in our analysis; third, some non-listed forestry firms also use social media to communicate CSR messages, but we are unable to include them in our analysis due to the difficulty of obtaining their firm-level data. We look forward to obtaining additional information through questionnaire analyses, field research, and case studies in the future to provide further answers to the above questions. Furthermore, we will investigate the following in future studies: (1) the motivation and the factors affected on the forestry firms’ disclosure differences between social media and CSR report; (2) Social media CSR disclosure of other industries related to carbon emission reduction (e.g., energy industry [45]).