A Concentric Model of Private Firm Leadership in China: Insights from Founding Entrepreneurs

Abstract

As discussed in Chap. 2, there has been little empirical research focusing specifically on executive leadership in Chinese private enterprises. The existing literature is composed mainly of large-sample survey studies of executive leadership, with no distinction between firms with different ownerships. Given an extremely dynamic, complex and inhospitable environment, both institutional and social, we argue that the emergence and success of the vibrant private economy can be ascribed primarily to the exceptional leadership of the founders of private enterprises as well as the other leaders who have worked with them. To further develop this insight, we interviewed 13 carefully selected and extraordinarily successful Chinese leaders of private firms to gain an in-depth understanding of their unique backgrounds, management practices and thoughts on leadership. Chaps. 3 to 15 contain the interview transcripts. In this chapter, we present the results of the content analysis of these 13 interviews.

Appendix 16.1

Data analysis: Attribute content analysis and thematic content analysis

We carried out qualitative analysis in two parts. The first is analysis based on the four leadership attributes described in Chap. 2, identified from the published large-sample research on Chinese firms, including both private and state-owned firms. The four attributes (4 Ds) are Determination, Divinity belief, Discipline and Duality focus. The confirmatory content analysis method was adopted in coding the transcribed data, following a similar step employed by Tsui, Wang and Xin (2006b). The purpose was to confirm the presence of these four attributes and to identify potentially new attributes that had not emerged during the qualitative analysis of the published sample studies.

The second is an open coding process of the interview data with no previously established framework, in order to identify the principal elements in the interviewees’ management philosophy, the development process of their thought, and those elements that might have driven the superior performance of their enterprises. The thematic content analysis method was used, following the step established by Boyatzis’ (1998) data-driven coding process. The purpose of this phase of analysis was to identify the generic pattern of leadership’s management philosophy and principle in managing business in China’s dynamic and changing environment.

Attribute content analysis

One author of this book first gave training on the coding to two early-stage researchers with a Chinese cultural background, providing definitions of the four attributes and suggested keywords or key phrases to identify each attribute (see Table 16.2). The principal unit of analysis was the paragraph. The coding instruction consisted of the following procedure. The two coders read Chap. 2 to gain a very clear understanding of the meaning of ‘the four Ds.’ They then reviewed the keywords associated with each attribute. After gaining some familiarity with the definitions of the four attributes, they both read an interview from beginning to end twice to become very familiar with it. They then performed the coding, paragraph by paragraph. For each paragraph, they had to decide if its content captured any of the attributes. Sometimes, a leader might digress, but it is important to focus on the main idea and avoid being distracted by the digression. If the digression captures an attribute, then the coder can assign a second attribute. In other words, the coder can assign each paragraph up to two, but no more than two, attributes. If a new concept is warranted, it would be coded as a separate attribute. Once the coding was finished, the coding results were entered into an Excel sheet, and the results of the two coders’ work were compared. Differences were discussed and resolved if possible. Finally, we calculated the proportion (percentage) of each attribute mentioned against the total number of paragraphs in the interview text.

One interview was selected for training and practice coding. After the two coders had completed the coding, one author of this book discussed each paragraph with them. After two rounds of coding, comparing, discussing the differences and recoding, the agreement between the third round of recoding was greater than 80 % on all the attributes.

After the second round of coding the first interview, a new attribute, ‘reflective thinking,’ was identified. The coding of the remaining 12 interviews was based on the five attributes, the four Ds plus reflective thinking.

For the other 12 interviews, the two coders carried out their coding independently. After completing the coding of three interviews, the two coders met with the researcher to discuss any ambiguous content and whether they had identified any new concepts. Most of the new concepts appeared in one interview only (e.g. paternalism) and did not appear in the others. If possible, we incorporated such unique concepts into existing attributes. For example, paternalism was integrated into divinity belief when it referred to benevolence or morality, and to discipline when it referred to authoritativeness. The agreement between the two coders for the five attributes ranged from 81 % to 94 %, indicating a high level of agreement.

Thematic content analysis

Another author of this book carried out an independent thematic content analysis based on a data-driven method (Boyatzis 1998), without bearing any previous theoretical framework in mind during the coding process. Following Table 16.1, the 13 companies and leaders were divided into four industries: banking, insurance and financial investment; technology and e-commerce; construction and real estate; and consumer goods and retailing. Codes and themes were identified, again based on the paragraph as the unit of analysis.

The coding process included three phases: the first was theme identification within each company; the second phase was to compare the identified themes across companies within the same industry. Third, the salient themes of each industry were applied to the companies in the other three industries to make a cross-industry comparison. A second co-author of this book participated in the third phase of coding to enhance reliability. An initial agreement of cross-industry theme comparison reached 78.6 %, higher than the required level of 75 % (Boyatzis 1998). Reasoning and discussion on these differences took place between the two authors, with a final agreement reached of 100 %.