5.3 Method

Dutch companies (defined in this chapter as companies that are incorporated and have a business in the Netherlands) were selected by means of purposive sampling.4 The researcher aimed for maximum variation in the selected cases in order to make a broad and meaningful comparison possible between different approaches implemented by companies of varying size and with varying business structures, numbers of employees, sectors in which they operated, listed or non-listed character, and multinational operations. All the companies investigated for this chapter have a two-tier board structure with a management and supervisory board.

Nine companies were approached, four of which agreed to participate in the research. All of the companies in the sample are large Dutch companies according to Dutch corporate law criteria, i.e., meeting at least two of the following criteria: (1) the value of the assets according to the balance sheet exceeds €20 million; (2) the net-turnover for the financial year exceeds €40 million; and (3) the company has on average, more than 250 employees. The investigated companies met all three requirements. In fact, they all have 1600 plus employees working in the Netherlands, a net turnover of at least €300 million and a net profit of €5 million or more. These inclusion criteria were used because the former gender target figure that applied to large companies used the same criteria for application of the law. Owing to their size, these companies had been facing pressure from the Dutch government to increase the number of women on their corporate boards and were expected to have implemented interventions to increase the number of women at their corporate top.

According to the findings in Section 5.4.1, three of the four investigated companies had interventions in place and one company did not. Nevertheless, it was still illuminating to investigate the latter company because it complied with the law without having implemented interventions. Therefore, including this deviant case in the sample served the purpose of investigating why that company had succeeded in appointing women to its boards.

Data was collected from multiple evidentiary sources (triangulation) in order to increase the reliability and validity of the study.5 Data was collected from interviews, policy documents and other external documents, such as annual reports. Due to Covid-19 restrictions the companies could not be visited on-site and, therefore, no field notes or observations could be used for this study. A total of 29 semi-structured interviews were conducted but only 28 interviews were used due to one participant’s refusal to sign the ‘informed consent form’.

A minimum of three interviews was conducted with each company. People from different organizational layers were interviewed, including HR managers (Chief Human Resources Officer), other people working in the HR department and management and supervisory board members. In some companies, CEOs, members of the works council, the diversity manager, senior management, a headhunter and people working in the business were interviewed. By selecting respondents from different organizational layers, plurality of responses and variety in the representation of various angles and perceptions was achieved. The interviewee respondents were a mix of women and men.

The first seven interviews were conducted in person and the remainder was conducted through video calling due to Covid-19 restrictions. Interviews took place between February 2020 and May 2021 and were conducted in Dutch or English. All interviews were recorded and transcribed by professional transcribers. Each interview lasted approximately 30 to 105 minutes.

All interview transcripts and documents were coded with descriptive coding6 using NVIVO computer-assisted qualitative data analysis software.

Template analysis was used to analyze the data.7 With template analysis, themes are identified and organized in such a way that patterns in the data are revealed.8 A coding template is developed with codes and themes, usually based on a subset of the data,9 that ‘summarises themes identified by the researcher(s) as important in a data set and organises them in a meaningful and useful manner’.10 Accordingly, several a priori themes were defined in advance of the coding of the data.11 Preliminary codes were used for a subset of the data (five interviews), resulting in an initial template. The remaining interview transcripts and documents were subsequently coded using this initial template.12 The initial template was modified several times, mainly to include more themes and sub-themes within a theme, resulting in a final template (Figure 5.1). Subsequently, the data interpretation took place by searching for patterns in themes and across cases.13