Choose an application

• Reference Manager
• EndNote
• RefWorks (Direct export to RefWorks)

The global increase of income and leisure time, the improved educational levels, the progressive decrease of travel costs and the consequent 'democratization' of travels have encouraged more and more cities to enter the tourism market. However, while opportunities to attract tourists seem to be increasing, time remains a limited resource and competition to attract tourists stays fierce. Understanding what determines tourism satisfaction can be particularly useful for both policy makers and destination marketers. Satisfaction is indeed a necessary condition to ensure repeat visitation or, at least, positive word-of-mouth. Getting a better insight on these topics can therefore be extremely useful to understand whether cities are positioning themselves competitively in the tourism market. In the last 20 years, tourism has become an integral part of the so called 'experience economy': for the holidaymaker, the tourism experience is of high personal value and is accompanied with satisfying and pleasurable emotions. The destination's 'basic' offerings (e.g. accommodation, information services) are not sufficient anymore at fully satisfying tourists. Emotional reactions to the tourism experience are fundamental determinants of satisfaction and post-consumption behaviors such as willingness to come back and intention to recommend. Apart from some exceptions, however, research on the experiential dimensions of the tourism offerings remains largely underexplored. Academic authors have mainly focused on satisfaction with 'physical' destination attributes, with little evidence being available about 'immaterial' factors affecting tourists' satisfaction. This thesis wants address the following research questions: what does determine visitor satisfaction in a tourism destination? Is the emotional dimension of a tourism destination a signification determinant of visitor satisfaction? The city of Mons was chosen to empirically answer these questions. Mons - a medium size city, counting around 95,000 inhabitants and located in a former industrial region in Belgium ('Le Centre' - is the European Capital of Culture (ECoC) 2015. As part of Mons 2015, a new tourism strategy was devised aiming at boosting both cultural and business tourism. This strategy wants to build on the potential echo of this mega-event to attract at least 500,000 visitors by 2015, compared to 250,000 in 2011. The modelling strategy used in this thesis draws on Structural Equation Modelling (SEM) with the objective to explore the relationships between four constructs, namely: perceived emotional quality, perceived quality of the destination's offerings, overall satisfaction and post-consumption behavior. The model was tested on a sample of 297 visitors selected randomly at major tourist locations in Mons from January until June 2015. The results confirm that both the perceived quality of the destination's offerings and perceived quality of the emotional experience are significant predictors of visitor satisfaction, with a stronger effect of 'emotional quality'. The parameter estimate for the relationship between visitor satisfaction and post consumption behavior is also strong and positive. However, we have to reject the hypothesis according to which the perceived emotional and functional quality would impact on destination loyalty.

Choose an application

• Reference Manager
• EndNote
• RefWorks (Direct export to RefWorks)

In a single-case experimental study (SSED), one or a few cases are repeatedly observed under several conditions, for instance in a baseline condition and a treatment condition. One such design is the AB phase design where there are two phases, the baseline or no treatment phase followed by the intervention or treatment phase. In group-comparison studies (GCED) groups of people are usually assigned to two or more groups which are then compared to one another. One such design is where there is a control or no-treatment group compared with a treatment or intervention group. These designs are quite similar to one another in that they compare the impact of treatment against its absence. Effect sizes are statistics that reflect the magnitude and direction of the treatment effects. Often, different studies yield different sizes and magnitudes of treatment effects. These effect sizes from different studies can then be combined together by technique called Meta-Analysis to yield more reliable estimates of the true effect size of a treatment. Standardized Mean Difference (SMD) is an effect size that can be used to measure the magnitude and direction of effects of GCED and SSED. Hedges' g, the effect size we used for GCEDs as used in Hedges, 1981, is the difference between the mean of the experimental or treatment group and the control group, divided by the pooled standard deviation. For SSED AB designs, we used the standardized mean difference Gingerich (1984), the difference between the mean treatment and mean baseline phase, divided by the pooled standard deviation. These effect sizes are similar but they are inherently different from one another as the GCED residual variance is a result of both, between participant variation, as well as unreliability on the individual's level (if the measurement were taken at a different time point the result would have been different) and SSED does not have between subject variances because each study often comprises a single individual. Van den Noortgate and Onghena (2008) showed how in theory SMD metric for SSED can be made comparable to SMD for GCED, and how results of studies with both kinds of designs can be combined in one meta-analysis by means of hierarchical linear models. Simulations, however, show that the converted effect sizes are biased, because the correction factor that is used depends on variance estimates that may be over estimated in small data sets. So, the purpose of the thesis is to investigate whether the use of fully Bayesian estimation for the correction factor yeilds better results. The results of this study show that the estimates for SSEDs using the Bayesian approach are less biased after the transformation compared to the maximum likelihood approach and the bias is now tolerable. The estimates using Bayesian approach are stable across different simulated conditions and the SSED effect sizes are indeed comparable to GCED effect sizes, barring the small bias. However, the estimates from maximum likelihood approach have a smaller mean squared error (MSE) despite the bias.

Choose an application

• Reference Manager
• EndNote
• RefWorks (Direct export to RefWorks)

Meta-analysis can be seen as a special case of multi-level analysis because meta-analytic data are clustered in a hierarchical way (with study participants nested in studies, but additional groupings are possible). In this sense, two-level and three-level meta-analyses represent very often used models nowadays. However, the possibility of a hierarchical model with the addition of an extra level has been barely used in practice as well as described in terms of its possible advantages and performance. In this sense, the aim of this study was to discuss the four-level analysis and its performance in a realistic but challenging situation where datasets are small and unbalanced. This is done by comparing the three-level meta-analysis with the four-level meta-analysis through simulations. Datasets were simulated assuming a balanced and an unbalanced nature of the data under a four-level hierarchy. These datasets where analyzed with a four-level and a three-level meta-analysis and the comparison was done in terms of bias, mean squared errors (MSE), standard errors (SE), Type I error rate and power for the fixed effect estimates. In addition, variance estimates are compared and the interpretation of the variances is discussed. As it was expected, the rate of Type I errors as well as the power were higher for the three-level model than for the four-level model. The SE's were also different between the two models, and the bias and MSE were significantly depending on the unbalanced of balanced nature of the simulations. In conclusion, the four-level meta-analysis showed a better performance than the three-level model in terms of controlled Type I error rate, higher power and accurate Standard Errors estimation.

Choose an application

• Reference Manager
• EndNote
• RefWorks (Direct export to RefWorks)

From the moment a stock is part of a portfolio, there is risk involved as the price path of the stock is not known in advance. Some risk is involved with the price change over a longer time period, but there is always the risk of a sudden jump (up or down) in the stock price. The risk involved can be managed by including derivatives of the stock. The most well known and publicly traded of these derivatives are "vanilla" put and call options. These options, when held in a portfolio, give the right (but not the obligation) to buy or sell the stock at a predetermined price - the strike price - and date - the expiration date. Buying this right requires a market-based price to be paid to the issuer of the option. This price takes into account the risks, both long term and short term, associated with the stock price. In practice, this risk is often measured by implied volatility, the only parameter in the Black-Scholes pricing model. A popular measure of this implied volatility for the S&P500 index is the CBOE VIX index. The VIX is a continuously updated time series calibrated from option prices and therefore based on the market-implied risk. The Black-Scholes model does not account for risk of a sudden jump in the stock price, but only for risk accumulated over time. There are other models available which do account for this jump risk such as the three-parameter Variance Gamma (VG) model. There have been several suggestions in literature for risk parameters which take jump risk into account. They compare the historical parameters (calibrated from the historical stock price) to the risk-neutral parameters (calibrated from option prices). These risk parameters have names such as up premium, down premium, market price of jumps, the market price of large moves or the market price of asymmetry, each with their distinctive interpretation. For example the market price of jumps (or the level premium) is a measure of the overall jump arrival rate in the risk-neutral measure relative to the historical jump arrival rate. Just like the VIX index, it could be informative (for stock traders) to have a continuously updated time series of these risk parameters. This could help to understand the risks inherent in the market at any given time. The goal of this thesis is to interpret the model and risk parameters, collect the market data (stock and option prices for the SPY ETF), perform the parameter calibrations (historical and risk-neutral), plot and finally analyze the corresponding risk parameter time series. From the results, it follows that the risk parameters react differently to the stock price evolution. For example, a sudden up move in the stock price makes the risk-neutral pricing temporarily more conservative by pricing drops more. The level premium typically moves in correspondence with the VIX index, but the market price of large moves and asymmetry shed more light on what actual change in risk the option premia charge for. The collected data also allows comparing intraday change versus overnight change in risk. For the considered time period, only the market price of jumps was found to be higher intraday than overnight. For the other risk parameters no significant conclusion could be made.