Sample and respondents selection
For any survey based research, the sample and its size are of utmost importance and required to be designed before the data collection (De Vaus, 2002)., It is ideal to have a larger sample size to reduce the likely errors in generalising to the population (Flynn et al., 1990; Miller and Salkind, 2002). Nevertheless, keeping in mind the time and cost constraints in collecting and analysing the data (Corbetta, 2003), the technique of probability sampling (e.g., stratified random sampling) is universally used in survey-based research strategies (Rungtusanatham et al., 2003; Saunders et al., 2007; Karlsson, 2009). As suggested by Saunders (2007, p214), this technique is commonly executed in four stages: (1) identification of a suitable sampling frame on the basis of research questions; (2) determination of a suitable sample size; (3) selection of the sample, based on appropriate sample; and (4) checking for the sample representativeness.
For conducting this research, the data collection was conducted in NCR Delhi region of India. The final sample was extracted from ten Indoo-Japanese joint ventures which were already discussed in Section 4.2.3. The sample frame is the population at sites and the desired unit of analysis is the individual employee, the unit from which the information is acquired and analysed (De Vaus, 2002; Babbie, 2004).
On the basis of previous studies (Corbetta, 2003; Saunders et al., 2007), the current study used the following formula to determine the sample size for this current research.
n = p% × q% × ( )²
n: the minimum sample size required
P%: is the proportion belonging to the specified category
q%: is the proportion not belonging to the specified category
z: is the z value corresponding to the level of confidence required
e%: is the margin of error desired
As recommended by Kwaw-Mensah (2008), Ary et al. (2010), and Kalton (1983), a conservative value of 50% (0.50) was set for both p and q to establish the maximum possible categories, with the level of confidence assigned at 95% certain. The associated z was found to be 1.96 and the desired margin of error e was assigned at 5% (0.05). Hence, the calculated minimum sample size of the current study was 385 (384.16).
n = 0.50 × 0.50 × ( )²
n ? 0.25 × 1536.64
n ? 384.16
As long as the respondents are concerned, ideally the questionnaire should be completed only by those employees who have knowledge and experience of implementing shop floor management and continuous improvement, so that, it reflect the research objectives as well as minimisies individual response bias. However, from a practical viewpoint, it was difficult to ensure that all respondents had shop floor management and continuous improvement experience. Hence, the respondents were randomly selected shop floor employees which also included the shop floor supervisors and managers.
For, Based on recommendations of previous studies (Forza, 2002; Babbie, 2004; Saunders et al., 2007), the sample representativeness was determined by the characteristics (e.g., salary grade, gender, length of service, structure, place of work, etc.) of the respondents and compared with the characteristics of the population. The organisational structure was also used to check the representativeness between the respondents and the population from the joint ventures. The results were shown in the following Section 6.2.1.
The data collection and screening
The data were collected via a self-administered method (Fowler, 1995; Corbetta, 2003; Saunders et al., 2007). The researcher personally visited each one of the selected companies to distribute the questionnaires and gave a brief introduction, outlined the objectives and explained the procedure to complete the questionnaire (Table 6.5). 100 hard copies of the questionnaire were distributed to respondents in the 10 companies. Putting all together, 1000 questionnaires were distributed, of which 527 samples were returned, giving a response rate of 52.7%.