Since the study uses two different questionnaires, the sampling designing process included two different processes. The specification of the target population for the first survey was as follows:
Elements: Experts in the tea industry
Sampling units: Individuals
Extent: Sri Lanka
Time: 1999
The sample was selected purposively based on the knowledge of the researcher. Respondents were selected mainly from government organisations and from tea brokers. Experts from the firms were excluded from this survey because the open-ended questions in the second questionnaire provide their views on the tea industry. Experts who are involved in the tea industry were selected from the following government organisations: Sri Lanka Tea Board, Tea Promotion Bureau, Tea Research Institute and Sri Lanka Export Development Board. Forbes and Walker Tea Brokers (Pvt) Ltd. was selected as one of the brokering firms.
The specification of the target population for the second survey was as follows:
Elements: Value-added tea producers
Sampling units: Firms
Extent: Sri Lanka Time: 1999
5: Firm
The target population should have included all finns involved in producing V AT during
1999, but a lack of complete and up-to-date lists of VAT producers in the country created problems in obtaining the total population. Therefore, this study was based on
all the finns (90 in total) registered in 1999 under the category tea at the Ceylon
Chamber of Commerce28, Sri Lanka. Prior to the survey, all the executives of the 90
finns were contacted over the telephone and asked to advise whether or not they had
been involved in the production of VAT during 1999. Fifty-eight finns indicated that
they were involved in the production of VAT during the year 1999. Therefore, all the 58
finns (64 per cent of the total finns registered at the Ceylon Chamber of Commerce, Sri
Lanka) were taken as the initial sampling frame29 for the study (Appendix 1). Table 5.1
shows details of the finns that were registered at the Ceylon Chamber of Commerce
under the category tea. Seventeen finns indicated that they were not involved in tea
during 1 999. The reasons for maintaining the membership varied from finn to finn, but
the common ones were future expectations of entering into the industry and non participation due to either lack of orders or low profitability compared with that of other businesses.
Table 5.1
Details of the Firms Registered at the Ceylon Chamber of Commerce (n=90)
of Finn Number of Finns
Finns in V AT 58
Finns in bulk tea and tea related activities 1 5
Finns not in tea 1999 1 7
Although the target population was specified as VAT producers i n Sri Lanka during
1 999, only 58 firms who confirmed that they were involved in VAT production were considered as the sampling frame for the study. This could have led to a sampling frame
28 Colombo tea auctions are held at the Ceylon Chamber of Commerce, Colombo. The registered members of the Chamber can bring any number of participants to the auction but non members have to obtain a letter from an auction broker if they want to participate in the auction.
29 A sampling frame is a representation of the elements of the target population (Malhotra, 1999, p. 330).
5: Firm
errorO, if there had been any VAT producers who had not been registered at the Ceylon
Chamber of Commerce during 1 999. Malhotra ( 1999), pointed out that it is essential to
recognise the existence of any sampling frame error and to know the treatment required to overcome the problem. Further, he pointed out that sampling frame error can be treated in three ways, and these are as follows:
1 . by redefining the population i n terms of the sampling frame,
2. by screening the respondents in the data collection phase to ensure that they satisfy the criteria for the target population, and
3. by adjusting the data collected by a weighing scheme to counter-balance the sampling frame error.
Before the data collection, respondents were screened by contacting the nominated executives and inappropriate elements were eliminated from the sampling frame. However, the main weakness of the sampling frame is that there can be a non representation due to non-membership at the Ceylon Chamber of Commerce. This was overcome by redefining the population in terms of the sampling frame. Although this is a simple approach it does prevent the researcher from being misled about the actual population being investigated (Malhotra, 1999, p. 33 1 ). After redefining, the target population can be represented as follows:
Elements: Value-added tea producers registered at the Ceylon Chamber of
Commerce
Sampling units: Firms
Extent: Sri Lanka
Time: 1999
The other components of the sampling design process, that is, sampling technique, sample size and executing the sampling process, were not adopted due to the limited
30 Sampling frame error may be defined as the variation between the population defined by the researcher and the population implied by the sampling frame (list) used (Malhotra, 1 999, p. 1 00).
5: Firm
number of firms engaged in VAT production in 1999. All the 58 firms included in the sampling frame were selected for empirical exploration.