• Tidak ada hasil yang ditemukan

Data and Results

Dalam dokumen Quantitative Methods in Tourism Economics (Halaman 53-56)

The empirical study was carried out by means of a questionnaire distributed throughout 2009, which was presented to a stratified, random sample of British and German tourists at the end of their vacations, with the central aim of determining their reasons for their intention to return to a specific destination. This survey considered the push and pull motives based on the literature review of destination choice studies and the socio-demographic profile of the respondents, in addition to some trip characteristics.

The questionnaire included three sections. The first one rates a set of 15 push motives and the second includes 19 pull motives (destination attributes). The set of push and pull motives are the most quoted in the literature (Correia et al.2007b), and measured using a five-point likert scale ranging from 1 (not important) to 5 (extremely important) the importance of each factor to their intention to return.

The third includes a dichotomous variable related with the intention to return or not, ranging from 0 (no return) to 1 (return). The latter section considers the socio-demographic factors (age, gender, education, travel companion, marital status, and frequency of travelling abroad and length of stay in Algarve on the present visit).

The definition of the sample was based on the number of passengers departing from the Algarve airport of Faro. Budgetary restrictions and the limited time available caused a decision to collect data from 2,000 respondents from the United Kingdom and Germany departing from Faro airport. These nationalities represent 56.3 % of the European tourists in Algarve. The sample was randomly stratified by the number of tourists by nationality. Hence data about British tourists account for 81.3 % whereas German account for 18.7 % of total sample, accordingly to the number of overnights registered to those nationals in Algarve.

The returned questionnaires totalled 1,500, from which 1,212 completed questionnaires were retained for the present analysis, representing a response rate of 61 % of the sample chosen. This corresponds to a sampling error of 2.8 % with a confidence interval of 95 % and this was considered an acceptable sample of respondents (Dillman 1978). The questionnaires received but not considered for the present research were discarded due to missing fields and other problems.

The extensive examination of the survey validity, reliability and generalisability leads to the inference that there is nothing in the evaluation to suggest that it is either invalid or unreliable. The characterisation of the respondents is displayed in Table3.1.

The general characteristics of these respondents indicated that they were British or German tourists travelling as couples, with an average age of 48 years and a standard level of education with an average 13.3 years of schooling. They are frequent flyers since they travel more than twice a year and spend more than 9 days on average at Algarve on the present visit. This profile leads to an overall definition of the responding tourist as early middle-aged couples and experienced tourists, motivated to return to the Algarve by its sun and beach attributes.

A mixed logit model was fitted, with the observed dichotomous variable (return) used to classify the individuals who declared that they will (¼1), or not return (¼0) as the dependent variable for British and German tourists, and the variables listed in Table3.2above as explanatory variables. From the set of push and pull motives considered in the questionnaire, some were discarded from the analysis because the model performs well without these variables. The Mixed logit is considered the most promising state of the art discrete choice model currently available to analyze Table 3.1 Socio-demographic and tripographic characterisation of the sample

Frequency (%) Average Standard deviation Gender

Male 63.7

Female 36.3

Average age 48 years old 13.6

Education (qualification level) 13 years 3.2

Marital status

Single 31.0

Married 69.0

Travel companion

Husband/wife 45.4

Friends 6.0

Family 34.2

Alone 10.4

Group 4.0

Travel experience

It is my first trip 0.3

1 time per year (once) 10.3 2 times per year (twice) 25.6 More than 3 times per year 63.8

Length of stay 9.1 days 10.3

Motivations for return (climate and beach)

Scale: 1¼not important to 5¼extremely important

4.5 0.7

Table 3.2 Characterisation of the variables

Variable Description Mina Maxb Mean Std. Devc

Dependent variable

Return-British British tourists declare that they will return

0 1 0.24 0.43

Return-German German tourists declare that they will return

0 1 0.20 0.29

Socio-demographic characteristics hypothesis

Age The respondent’s age 18 85 43 14

Education Education (number of years of education) (4-primary school, 24–Ph.D.)

4 24 13.28 3.66

Genre The gender (0¼male, 1¼female)

0 1 0.42 0.49

Marital status Marital status (1¼single;

2¼married; 3¼with children)

0 1 1.75 0.47

Frequency hypothesis

Frequency Number of times you travel abroad for holidays during a year? 1¼once, 2¼twice, 3¼three times a year, 4¼more than three times a year

1 4 3.07 0.94

Duration hypothesis

Duration What was your length of stay?

(number of days)

2 90 8.81 8.09

Push and Pull hypothesis

How important to you are the following factors if you decide to return? (1¼not important, 5¼very important)

Push 1 Do different things 1 5 3.47 1.38

Push 2 Have an adventure 1 5 2.56 1.48

Push 3 Talk about the vacations 1 5 2.81 1.47

Push 4 Release stress 1 5 3.36 1.36

Push 5 Escape from routine 1 5 3.93 1.31

Pull 1 Landscape 1 5 4.39 0.75

Pull 2 Nature 1 5 4.33 0.79

Pull 3 Cultural attractions 1 5 4.00 0.93

Pull 4 Nightlife 1 5 3.77 0.95

Pull 5 Transport 1 5 3.85 0.93

Pull 6 Shops 1 5 4.31 0.80

Pull 7 Lodging 1 5 4.35 0.90

Pull 8 Climate 1 5 4.47 0.74

Pull 9 Beach 1 5 4.09 0.92

Pull 10 Gastronomy 1 5 4.13 0.85

Pull 11 Relaxing environment 1 5 4.05 0.89

aMin¼minimum

bMax¼maximum

cStandard deviation

questionnaire data, Hensher and Greene (2003). The advantage of the Mixed Logit over alternative models is based in two improvements. First, it allows the error term to combine different statistical distributions, which is an improvement relative to alternative specifications that rely in one specific distribution. Second, it allows for random taste variation parameters (parameters that describe characteristics not linked to observed characteristics, whenever the traditional logit allows for taste variation related to observe characteristic).

Focusing on the adequacy of the standard logit model, the RESET test is implemented. The result gives evidence of mis-specification of the logit model.

This may be due to the presence of unobserved heterogeneity, depending on the aforementioned explanatory variables. A HAL test of Chesher and Santos-Silva (2002) was applied, using likelihood ratio procedures to test against this type of heterogeneity. Firstly, it was considered that it was dependent on all of the explan- atory variables. Next, a classic selection procedure is applied. The findings display evidence of unobserved heterogeneity depending on age and education, inducing the estimation of a mixed logit model with random coefficients for these variables.

Table3.3displays the results.

Model parametersbrelate changes in the explanatory variablesXito changes in the response probability. While the parameter signs indicate the direction of the relationship, they are not directly interpreted as marginal changes to the mean value of the dependent variable. This is because of the non-linear form of the distribution function. Based on the global model, using the transformation 100(eb1), it can be said that each additional visit of the individual British tourist yields a decrease of about 44.95 % in the likelihood of returning to the Algarve, while for a German tourist it is about 17.87 %.

Dalam dokumen Quantitative Methods in Tourism Economics (Halaman 53-56)