Based on our sampling frame, stratified on location, sector, and size, firms were randomly drawn from local firm registers. In a first step, for each city and industry a sampling pool was established by drawing every nth company from complete local firm registers. After confirming the eligibility of the firms, our local partner approached local entities such as the Bureau of Industry and Commerce, party organizations, and other institutions such as industry guilds for their support and assistance in securing entrepreneurs’ participation in our survey. Once local support was secured, the sampled entrepreneurs were invited to participate.
Eligible respondents were the companies’ chief executive officers (CEOs). If the CEO could not participate in the interview personally, the interview was rescheduled. This increased comparability of responses across all companies. Moreover, the strategy helped secure more detailed and reliable information on specific processes that fall naturally into the domain of the firm manager. In a majority of cases, the respondent also held ownership shares in the company (83 percent of interviewees) or was even one of the company founders (74 percent of interviewees).
Firm sample by city, sector and firm size, 2006
| Industry | Size* | Nanjing | Nantong | Changzhou | Wenzhou | Ningbo | Hangzhou | Shanghai | Total |
| Textile | L | 2 | 5 | 3 | 2 | 0 | 4 | 4 | 20 |
| M | 0 | 13 | 5 | 14 | 10 | 11 | 4 | 57 | |
| S | 10 | 9 | 14 | 10 | 21 | 23 | 11 | 98 | |
| Auto parts | L | 1 | 5 | 1 | 1 | 4 | 1 | 5 | 18 |
| M | 6 | 3 | 2 | 7 | 5 | 0 | 8 | 31 | |
| S | 32 | 12 | 12 | 15 | 11 | 7 | 37 | 126 | |
| Machinery | L | 2 | 5 | 0 | 1 | 0 | 3 | 1 | 12 |
| M | 1 | 5 | 1 | 5 | 5 | 8 | 8 | 33 | |
| S | 24 | 20 | 14 | 27 | 24 | 18 | 3 | 130 | |
| Pharma-ceutical | L | 0 | 2 | 0 | 1 | 0 | 7 | 1 | 11 |
| M | 0 | 5 | 6 | 0 | 0 | 1 | 4 | 16 | |
| S | 7 | 5 | 8 | 2 | 9 | 9 | 10 | 60 | |
| Electronics | L | 2 | 2 | 0 | 3 | 0 | 2 | 0 | 9 |
| M | 2 | 6 | 7 | 4 | 5 | 3 | 1 | 28 | |
| S | 13 | 4 | 20 | 9 | 7 | 6 | 3 | 62 | |
| Total | 102 | 101 | 103 | 101 | 101 | 103 | 100 | 711 | |
* L = more than 300 workers; M = 101 to 300 workers; S = 10 to 100 workers in 2003.
Overall, with 711 completed interviews in 2,842 contacted firms, the response rate for our 2006 survey wave was 25 percent. This rate meets standards widely used by the professional survey research industry, which regards overall response rates of firm surveys in the range of 25 percent as satisfactory. Baruch and Holtom (2008) analysed 1607 studies published in the years 2000 and 2005 in 17 refereed academic journals, and identified an average response rate of 35.7 percent with a standard deviation of 18.8 for studies that utilized data collected from organizations. Studies in the Asian context typically yield lower response rates.
In our survey nonresponses often resulted from strict adherence to the rule only to interview the firm’s CEO. Extended business trips— partly outside China— or participation in national and international trade fairs made personal visits with CEOs at times simply impossible. If after three contact attempts it was still not possible to line up a personal appointment for a face- to- face interview with the firm manager, we replaced the firm by another randomly sampled participant firm drawn from the sampling pool.
