Huang et al.

(2023)
Agustina et al. (2021)
Xie et al. (2019)
Metilelu et al. (2023)
Nunkoo and Ramkissoon (2011)

INTRODUCTION

This document presents a synthesized analysis of five research papers that apply quantitative

methodologies to tourism research. The papers are organized not only for easy reference but also to

illustrate a clear evolutionary path from basic linear models to advanced structural equation modeling.

Each analysis follows a consistent structure to facilitate understanding, while a continuous thread of

comparative insight highlights the strengths, trade-offs, and appropriate contexts for each method

across different tourism scenarios.

PAPER 1: LINEAR REGRESSION [Huang et al. (2023)] – A CLEAR MODEL FOR A

SPECIFIC CONTEXT

Overview

This study investigates factors influencing young tourists' (20-29 years old) intention to use GoShare

electric motorcycle-sharing services for tourism activities in Taiwan. The research addresses the

intersection of sustainable tourism and the sharing economy.

Methodology

The study employed a quantitative cross-sectional design using convenience sampling of 271
respondents. Data was collected via online questionnaires and analyzed using descriptive statistics, t-

tests, ANOVA, and multiple linear regression in SPSS.

Key Findings

 Tourist characteristics "Freestyle travel" and "Enthusiasm" significantly predicted use intention.

 "Product Attribute" had a stronger impact than "Environmental Efficiency."

 Rental factors (β = 0.655) demonstrated a much stronger influence than tourist characteristics (β =

0.155).

 The primary barrier to use was the perception that "Renting is not easy."

Conclusions and Implications

Intention to use is significantly influenced by travel style but is dominated by service quality

perceptions. The study provides actionable insights for service providers to improve product attributes

and for policymakers to support sustainable mobility.

Critical Evaluation & Methodological Trade-off

Strengths: Clear focus and robust, appropriate statistical methods.

Limitations: Narrow demographic focus, convenience sampling, and dated data (2019).

Trade-off: The choice of a simple linear model enhances practical clarity and actionability for a

specific business problem but sacrifices generalizability and depth, offering a snapshot rather than a

deep exploration of behavior.

Personal Reflection

This paper demonstrates a clear disconnect between stated environmental attitudes and actual
behavioral drivers, which are dominated by practicality. It raises a critical question about how these

drivers might vary across cultures, suggesting a potential for a mixed-methods follow-up study.

PAPER 2: LOGISTIC REGRESSION [Agustina et al. (2021)] – INFORMING STRATEGY

WITH BINARY OUTCOMES

Overview

This research identifies determinants of "quality" foreign tourists defined as those spending above

average to support Indonesia's shift from a quantity-based to a value-based tourism strategy.

Methodology

Using secondary data from a national survey (n=2,504), the study employed binary logistic regression

in R, classifying tourists based on an average expenditure threshold of US$134.82.

 Strongest predictors: hotel accommodation (20.61x higher odds) and stays >8 days (8.24x higher

odds).

 ASEAN nationals, business travelers, and middle-aged entrepreneurs were more likely to be high

spenders.

 Gender was not a significant predictor.

 Shopping was the largest expenditure category (41.11%).

Conclusions and Implications

Tourist "quality" is systematically associated with specific, targetable characteristics. Findings support

refined marketing and policy development to attract higher-value tourists.

Critical Evaluation & Methodological Trade-off

Strengths: Large, official dataset and a statistically appropriate method for the research question.

Limitations: Exclusion of air/sea arrivals and correlation vs. causation issues.

Trade-off: The dichotomization of a continuous variable (spending) creates a clear, policy-friendly

binary outcome but loses nuanced financial data and may obscure more complex spending patterns.

Personal Reflection

This paper showcases how logistic regression can directly inform national economic strategy. The

power of hotel accommodation as a behavioral indicator is fascinating and worthy of further qualitative

investigation into the decision-making process behind accommodation choices.

DIAGNOSTICS

Overview

This study identifies drivers of China's domestic tourism income using time-series data (1994-2014)

while directly addressing statistical issues like non-stationarity.

Methodology

A quantitative longitudinal study using secondary data and multiple linear regression with cointegration

theory in EViews, including unit root and cointegration tests.

Key Findings

 All variables were second-order integrated with a confirmed cointegration relationship.

 High explanatory power (R² = 0.9988) with GDP and domestic tourists showing significant

positive effects.

 Unexpectedly, urban and rural travel rates showed negative coefficients.

 Diagnostic issues included possible autocorrelation and suspected multicollinearity.

Conclusions and Implications

Domestic tourism income is driven by GDP and tourist numbers. The study demonstrates the

importance of cointegration analysis for time-series data.

Critical Evaluation & Methodological Trade-off

Strengths: Application of advanced econometric techniques (cointegration) often overlooked in tourism

research.

Limitations: Small sample size (n=21 years) and unaddressed multicollinearity/autocorrelation.

Trade-off: While the model accounts for non-stationarity, the failure to fully diagnose and correct for
multicollinearity undermines confidence in the individual coefficient estimates, leading to potentially

misleading interpretations (e.g., the negative travel rate coefficients).

Personal Reflection

This paper is a crucial reminder that a high R² is not a hallmark of validity. It emphasizes that

counterintuitive results should be seen as signals of potential methodological issues, a lesson that

reinforces the importance of rigorous model testing.

PAPER 4: NON-LINEAR REGRESSION [Metilelu et al. (2023)] – ALIGNING METHOD WITH

DATA REALITY

Overview

This study introduces a novel non-linear normal-power model to analyze the tourism-economic growth

relationship in West African economies, challenging the inadequacy of linear models for bimodal data.

Methodology

Using longitudinal panel data from four ECOWAS countries, the study employed normality tests and

non-linear modeling in R, with a comprehensive model comparison using log-likelihood, AIC, BIC,

MSE, R², and LRT.

Key Findings

 GDP per capita data was confirmed as bimodal and non-normal.

 The non-linear model significantly outperformed the linear model on all fit indices.

 Tourist arrivals explained a large variation in GDP (e.g., Ghana 88.5%, Nigeria 77.2%).

 The Likelihood Ratio Test (LRT) overwhelmingly favored the non-linear model.
Conclusions and Implications

The tourism-growth relationship in emerging economies is inherently non-linear. This urges strategic

investment and more accurate forecasting using appropriate models.

Critical Evaluation & Methodological Trade-off

Strengths: Addresses a significant methodological gap with a novel approach and a robust, multi-

faceted comparison.

Limitations: Limited generalizability (four countries) and oversimplification with a single predictor.

Trade-off: The innovative model offers a superior fit for the specific, non-normal data but at the cost of

simplicity and ease of interpretation compared to a standard linear model.

Personal Reflection

This paper powerfully demonstrates that methodological misalignment can mask complex realities. It

underscores the necessity of testing model assumptions rather than defaulting to standard methods, a

fundamental principle for rigorous research.

PAPER 5: STRUCTURAL EQUATION MODELING (SEM) [Nunkoo & Ramkissoon (2011)] –

A FRAMEWORK FOR THEORETICAL RIGOR

Overview

This conceptual paper provides a primer on Structural Equation Modeling (SEM) for tourism

researchers, explaining its theory-driven nature and advantages over basic regression.

Methodology

A methodological review that synthesizes existing statistical literature, outlining SEM components,

processes, and evaluation criteria without primary data analysis.

Key Findings
 SEM process involves theory development, measurement model testing (CFA), and structural

model evaluation.

 Advantages over regression: modeling measurement error, simultaneous testing of relationships,

and theory development through respecification.

 SEM is complementary to regression, not a replacement.

Conclusions and Implications

SEM is a powerful second-generation tool superior for testing complex multivariate theories but

requires strong theoretical grounding.

Critical Evaluation & Methodological Trade-off

Strengths: Fills an accessibility gap with a structured, non-technical overview.

Limitations: Lack of empirical demonstration and remaining conceptual density for beginners.

Trade-off: The paper excels at theoretical explanation but fails to provide the practical, applied

demonstration that would bridge the gap for practitioners.

Personal Reflection

This paper highlights how accounting for measurement error can fundamentally alter findings. It forces

a crucial consideration of the balance between methodological sophistication and accessibility in

research.

SYNTHESIZED COMPARATIVE ANALYSIS

The Evolutionary Arc of Methodology

The papers collectively chart an evolution in analytical sophistication and purpose:

1. Basic Associations: Paper 1 (Linear Regression) for direct hypothesis testing of continuous

outcomes.
2. Classification: Paper 2 (Logistic Regression) for binary outcomes and policy-friendly

categorization.

3. Temporal Dynamics: Paper 3 (Time-Series) for longitudinal data, introducing essential

econometric corrections.

4. Complex Realities: Paper 4 (Non-Linear) for non-standard data distributions, proving that

misapplied linear models can be invalid.

5. Theoretical Systems: Paper 5 (SEM) as a capstone, providing a framework for modeling latent

constructs and complex networks of relationships.

Cross-Cutting Themes

Economic Focus: All empirical papers center on economic outcomes, from individual spending to

national GDP.

Policy Relevance: A consistent drive to provide actionable insights for industry and policymakers.

The Primacy of Fit: The recurring, fundamental lesson is that methodological choice must be dictated

by the research question and data characteristics, not by convention or simplicity.

Identified Research Gaps

Methodological Isolation: A pronounced lack of integration between quantitative findings and

qualitative insights.

Cultural Generalizability: Limited cross-cultural comparisons in tourist behavior studies.

Contemporary Context: A need for updated studies reflecting post-pandemic tourism dynamics and the

integration of more comprehensive predictor variables.

These five papers form a coherent narrative about the power and pitfalls of quantitative methods in

tourism research. The key synthesis is not merely a list of techniques, but a fundamental principle: the

methodological tool must be fit for purpose.

Huang et al. (2023) and Agustina et al. (2021) show that well-applied basic models yield clear,

actionable results for specific, well-defined problems. Xie et al. (2019) and Metilelu et al. (2023)

demonstrate that methodological negligence can invalidate findings, while methodological innovation

can uncover hidden truths that simpler models miss. Finally, Nunkoo & Ramkissoon (2011) provides

the theoretical framework for building more robust, comprehensive models that acknowledge the

complexity of human behavior.

The ultimate lesson is that methodological rigor, grounded in a deep understanding of both technique

and context, is what transforms raw data into meaningful evidence. It is this rigor that enables truly

evidence-based decision-making in the complex and multifaceted world of tourism.

 Agustina, N., Puspita, C. D., Arifatin, D., & Yordani, R. (2021). Application of logistic regression

to determine the quality of foreign tourists to Indonesia. Journal of Physics: Conference Series,

1863(1), 012029. [Link]

 Huang, Y.-J., Jhuang, S.-Y., Lin, Y.-S., & Chan, H.-Y. (2023). Applying linear regression analysis

to identify willingness of using environment-friendly electric motorcycle-sharing for tourism

activities: A case study of GoShare. Engineering Proceedings, 38(1),

57. [Link]

 Metilelu, O., Ekum, M., & Toki, E. (2023). Modelling the non-linear effects of tourism

development on emerging market economies using normal-power model. Scientific African, 20,

e01731. [Link]

 Nunkoo, R., & Ramkissoon, H. (2011). Structural equation modelling and regression analysis in

tourism research. Current Issues in Tourism. [Link]

 Xie, S., Zhou, F., & Xu, X. (2019). Analysis of factors affecting domestic tourism income based on

multiple linear regression. In Proceedings of the 2019 9th International Conference on Education

and Social Science (ICESS 2019) (pp. 1044-1047). Francis Academic Press.