Research Design and Methodology
5.5 Sampling and data collection
5.5.1 Justification of the sampling method
The first person encountered was interviewed and subsequently after completing each interview, the next person. A refusal was overlooked until the next person accepted. Very few people refused to participate. There are limitations in a convenience sample of this type and there is a risk of response bias. In order to overcome potential bias, the sampling was conducted in different locations and a large sample was collected.
5.5.2 Data collection procedure
Prior to commencing data collection, the number of tourist arrivals to Samui Island was assessed using statistics provided by the Tourism Authority of Thailand (TAT), which is the government agency responsible for tourism promotion. Arrival statistics for 2004 and 2005 show that Australian tourist arrivals on Samui Island during January and February were 62.3% (in 2004) and 68.5% (in 2005) of the total number of Thai tourists that arrived during the same period. This would suggest that the Australian sample size should be between 62.3% and 68.3% of the Thai sample size in order to ensure that the sample size for the two groups is in the right proportion. However, according to Zikmund (2003), sample size should be based on the size of the population. Based on this prescription a suitable sample size for a population between 10,000 and 20,000 would be approximately 297 to 302 respondents. Since the total population of Australian and Thai tourist arrivals in January and February was 15,452 and 10,584 respectively, a total of 300 respondents for each nationality is
a suitable sample size. Leading quantitative academics have asserted that a total of 300 respondents is a sufficient sample size for most quantitative studies, and a suitable number for various types of statistical analysis such as descriptive techniques, correlation and factor analysis (Comrey & Lee 1992; Field 2005; Kass & Tinsley 1979; Tabachnick & Fidell 2001). This consideration is especially important for the conduct of factor analysis where it has been found that the parameters tend to be stable regardless of changes to the number of respondents per variable, when the sample size is equal to or larger than 300 respondents (Kass & Tinsley 1979).
In order to achieve the target number of respondents, the data collection process was conducted in several places including in resorts, at the airport, at the ferry pier and in other places on Samui Island where tourists congregate. In the first instance resort guests were accessed by agreement with resort management. Prior to going to Samui Island the researcher had made arrangements through the Tourism Authority of Thailand (TAT), to have the questionnaires given to arriving guests at resort hotels on Samui Island. In order to ensure that the number of required Australian tourists could be identified most efficiently; the various resort hotels were listed. The sales manager of Qantas Holidays, one of the premier travel wholesalers for the Australian tourism market, was contacted to help in identifying the most popular resorts for Australian tourists on Samui Island. Following this discussion, 16 resort hotels were identified. A list of these 16 resort hotels was sent to the TAT to negotiate an agreement to survey at these properties. These were identified as the most appropriate resorts and agreement was reached with various resort managers. However, it was anticipated that resort guests might be reluctant to complete questionnaires while on holiday, so several other data collection points were identified for the survey. These were local
attractions where resort guests spend most of their time during their holiday (e.g. beaches) as well as at the airport and the ferry pier where tourists depart from the island. Resort guests were intercepted during their stay or just prior to departure. A full explanation of how responses were achieved is provided below. It was anticipated that the hotel survey method would include a higher level of incomplete surveys. However, the two combined methods provide a large and representative sample.
5.5.2.1 Resorts
Management in several resorts agreed to distribute the questionnaires. These included Muang Samui resort, Chaweng Regent resort, Nora Beach Resort and Spa, Anatara Resort and Spa, Amari Palm Reef resort, Imperial Boat House resort, Imperial Samui resort, Impiana resort, Nova Samui and Banana Fan Sea resort. A total of 1,000 questionnaires (500 for Australian respondents and 500 for Thai respondents) were distributed to each resort depending on the agreement reached between the resort managers and the researcher. Due to management privacy, the researcher was not allowed to make direct contact with guests. The questionnaires were left at the front desk at most resorts for receptionists to distribute to guests when they were checking in. The house keeping staff were to pick up the completed questionnaires following check out. At the end of the data collection period, 87 completed questionnaires (66 Australian and 25 Thai) were returned to the researcher.
Since the numbers of completed questionnaires were insufficient, some additional resorts were added for the study. Those resorts were Chalie’s Hut, Montien House, Chaba Samui resort, Long Beach resort, Chaweng Villas and Chawengburi resort. A
total of 86 completed questionnaires (70 Australians and 16 Thais) were returned from these properties.
At the end of this stage, 177 completed questionnaires (136 Australians and 41 Thais) had been collected from the resort hotels. This number was insufficient for data analysis purposes. Consequently, the data collection procedure was expanded to local attractions on the island.
5.5.2.2 Local attractions
These attractions included the Bungy Jump, Toureast travel agency and Chaweng beach. At the end of this stage, the researcher collected a total of 189 completed questionnaires (148 Australians and 16 Thais), which was still insufficient. Consequently, the data collection procedure was extended to the airport and ferry pier in order to collect data from tourists prior departure.
5.5.2.3 Airport and Ferry pier
Since the peak periods vary (the ferry pier was busy in the morning while the airport was busy in the afternoon), the researcher was able to collect at both places during the day. A total of 68 completed questionnaires (all Thai) were collected from the ferry pier at the island capital, also known as Na Thon. Most of the respondents were tourists who were travelling by car, arriving at the ferry pier to cross back to the mainland. Data was collected while respondents were queuing in their cars waiting to board the ferry. A total of 344 completed questionnaires (152 Australian and 191 Thai) were collected at the airport. Respondents were approached after they had checked in their baggage and were waiting for the final boarding call. The response
rate was highest at the airport, because there were few interesting activities at the airport and the waiting period ranged from one to three hours. As a result, most respondents were happy to fill out the questionnaires. At the end of this stage, a total of 600 completed questionnaires (300 Australians and 300 Thais) had been received.