INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,

MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)

ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XIV, Issue XI, November 2025

Exploring How Green Marketing Influences Shipping Tourist

Destination Choice: Evidence from R Programming-Based Logistic

Regression and Roc Analysis

¹Dr. M. A. Shakila Banu, ²Pazila Sara. R. S

¹Associate Professor, PG& Research Department of Commerce, Jamal Mohamed College (Autonomous)

Trichy-20

²BE- Naval Architecture and Offshore Engineering, AMET University, Chennai

DOI : https://doi.org/10.51583/IJ L T EMAS.2025.1411000067

Received: 24 November 2025; Accepted: 01 December 2025; Published: 10 December 2025

ABSTRACT

This study explores how green marketing influences the destination choice of shipping tourists, a segment where

traveller’s increasingly seek sustainable experiences, but empirical evidence remains limited. This quantitative

research employed a structured questionnaire and used R-based statistical techniques, including logistic

regression, ROC analysis, and mediation analysis, on a sample of 325 shipping tourists to predict destination

choice. The analysis confirmed excellent instrument reliability (alpha = 0.90) and found strong positive

correlations between Green Marketing Awareness and Destination Choice (r=0.832); however, the logistic

regression model found no statistically significant individual effects for the predictors, and the Causal Mediation

Analysis failed to detect any significant direct or indirect effects. While the model showed fair overall

discrimination (AUC = 0.7845) and an adequate fit (Hosmer-Lemeshow p > 0.05), its practical utility is severely

limited by poor sensitivity (only 0.92% correctly classified true positives); therefore, future research must re-

evaluate the binary measurement of Destination Choice and incorporate unobserved factors to enhance the

predictive power of the model and generalizability of findings.

Keywords:

Green Marketing, Shipping Tourism, Destination Choice, Environmental Attitude, Logistic

Regression

INTRODUCTION

Shipping tourism, including cruises and river/boat travel, has become increasingly popular in India and

worldwide as tourists seek both leisure and sustainable experiences. In recent years, tourists have become more

environmentally conscious and prefer destinations that follow eco-friendly practices (Akram & Lavuri, 2023).

Green marketing, which includes promoting sustainability, eco-certifications, and environmental responsibility,

plays a vital role in shaping tourists’ perceptions of destinations (Chang et al., 2025). Studies show that eco-

friendly messaging increases tourists’ trust and positively influences their decision-making, making it an

essential strategy for shipping tourism operators (Vicente et al., 2024).

However, there is limited research on how green marketing affects shipping tourist destination choice

specifically. Understanding this influence can help cruise and boat-tourism operators develop strategies to attract

environmentally conscious travellers while supporting sustainable tourism practices (Setiadi et al., 2024; Kumar

& Kulshreshtha, 2023). This study uses R-based logistic regression to predict the probability of tourists choosing

green-marketed shipping destinations and applies ROC analysis to measure the accuracy of the model. The

findings aim to provide practical insights into how green marketing can enhance destination selection and

promote eco-friendly behaviour among shipping tourists.

www.ijltemas.in

Page 738

INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,

MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)

ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XIV, Issue XI, November 2025

Review

• SuneelꢀKumar & NishaꢀDevi — AIMT Journal of Management, Vol.ꢀ12, Issueꢀ1–2, 2023, The study examines

the effect of green marketing tools (eco-labels, eco brands, eco advertisements) on green consumer behaviour

among tourists in Himachal Pradesh, India. They find that increased use of green branding and eco-labels

strongly correlates with tourists’ willingness to pay and choose eco friendly tourism options. Journalskart

• Mona Karina, Rinto Rivanto, Hasan Basri & Endang Iryani — Branding: Jurnal Manajemen dan Bisnis, Vol.ꢀ4,

Issueꢀ1, 2025, This qualitative-literature study argues that green marketing can transform conventional tourism

into sustainable tourism. The authors emphasise that just green messages are not enough — business actors must

deliver them authentically and consistently, and work with government and local communities to succeed.

Journal UINSgd

• Chien Chung Yu & Chun Chu Liu — International Journal of Religion, Vol.ꢀ5, Issueꢀ4, 2024, pp.ꢀ224–230,

This paper builds a green marketing model for the tourism and amusement industry using fuzzy analytic

hierarchy process. It suggests that “brand loyalty” and “corporate image” are key indicators for green marketing

in tourism. The authors argue that sustainable management demands that companies integrate green marketing

into their core strategy. Ijor+1

• Sanjaya, D., Arief, M., Juli Setiadi, N. & Heriyati, P. — Journal of Eastern European and Central Asian

Research (JEECAR), Vol.ꢀ11, Issueꢀ3, 2024, pp.ꢀ553–572, The study explores how digital green marketing

campaigns (online eco ads) and environmental beliefs shape tourist behavior and revisit intentions in eco-

tourism. It finds that well-crafted digital campaigns significantly influence tourists’ environmental beliefs, which

in turn increase their commitment to revisit sustainable destinations. IEECA

• SuneelꢀKumar (Prof.) & NishaꢀDevi — AIMT Journal of Management, Vol.ꢀ12, Issueꢀ1–2, 2023 (same as #1,

but additional insight), Beyond demographics, they also analyse how green marketing tools influence pro-

environmental behaviour and find that older or more educated tourists respond more strongly to eco brands and

ads, implying that green marketing effectiveness may vary across tourist segments. (Derived from same paper)

Journalskart

• Jumadi, Ascasaputra Aditya, Arya Saputra &Aldi Irsyad Burhani — Proceedings of the 2nd UPY International

Conference on Education and Social Science, Atlantis Press, 2023, pp.ꢀ471 477, They use a regression model

(via SPSS) to examine the effect of green marketing (product, place, promotion) on green tourism. They find

that “product” (eco friendly offering) and “place” (green destination) significantly influence tourists’ perception

of green tourism, but “promotion” has weaker influence, highlighting that just advertising green is not enough.

Atlantis Press

• Sri Rahayu, Hikmatul Aliyah & Sudarwati — International Journal of Economics, Business and Accounting

Research (IJEBAR), Vol.ꢀ6, Issueꢀ1, 2024, This study looks at green marketing components (green product, green

promotion, green price) and their effect on intention to visit eco-tourism sites. They also test whether

environmental knowledge mediates these relations. They find that green price and green promotion significantly

influence visit intention, but green product does not, and environmental knowledge does not mediate the effect

of green product. jurnal.stie-aas.ac.id

• Mdpi – Green and Environmental Marketing Strategies and Ethical Consumption — Sustainability, Vol.ꢀ15,

Issueꢀ16, 2023, This paper explores the role of green environmental strategy, green marketing, and psychological

factors (perceived behavioral control, subjective norms) on ethical consumption in tourism. Using TPB (Theory

of Planned Behavior), they find that green marketing strongly influences ethical consumption and that

psychological control plays a moderating role.

Overview of Reliability Test (Cronbach’s Alpha):

Reliability testing is used to check the consistency of a set of questionnaire items that measure the same concept.

Cronbach’s Alpha (α) is the most common reliability coefficient. It indicates how closely related a set of items

www.ijltemas.in

Page 739

INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,

MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)

ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XIV, Issue XI, November 2025

are as a group. The value of α ranges from 0 to 1, where a higher value (usually ≥ 0.7) shows good internal

consistency. The formula for Cronbach’s Alpha is:

r Cronbach’s Alpha is:

ꢀ

_푖=1푋 휎_푖²

휎_푇²

∑

퐾

훼 =

(1 −

)

퐾 − 1

------------>1

Where,

k = Number of items in the scale,

휎_푖²= Variance of each individual item,

휎_푇²= Variance of the total score of all items.

Overview of Correlation Analysis

Correlation analysis is used to measure the strength and direction of the relationship between two variables. It

helps researchers understand whether an increase in one variable is associated with an increase or decrease in

another variable. The most common measure is the Pearson correlation coefficient (r), which ranges from -1 to

+1. A positive value indicates a direct relationship, a negative value indicates an inverse relationship, and 0

indicates no linear relationship. The formula for Pearson correlation is:

ˉ

∑(푋_푖− 푋)(푌 − 푌)

푖

푟 =

2

ˉ

√

∑(푋_푖− 푋) ∑(푌 − 푌)

푖

------------>2

Where,

X_iand Y_iare individual observations,

X ˉand Y ˉare the means of X and Y, respectively.

Overview of Logistic Regression

Logistic Regression is a statistical method used to predict the probability of a binary outcome (like Yes/No, 0/1)

based on one or more independent variables. It is widely used in research to study how predictors influence

categorical outcomes, such as whether a tourist chooses a green-marketed destination. The logistic regression

model transforms the linear combination of predictors using the logit function:

푝

logit(푝) = ln⁡(

) = 훽₀+ 훽₁푋₁+ 훽₂푋₂+ ⋯ + 훽_푘푋_푘

1 − 푝

------------>3

Where,

p= Probability of the outcome occurring (e.g., choosing a green destination),

β_0= Intercept,

β_1,β_2,…,β_k= Coefficients of independent variables X_1,X_2,…,X_k.

www.ijltemas.in

Page 740

INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,

MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)

ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XIV, Issue XI, November 2025

Overview of Model Fit & Diagnostics:

In regression analysis, model fit and diagnostics help us check whether the model explains the data well and

whether assumptions are valid. A well-fitted model accurately predicts the dependent variable and shows

minimal error. Common measures include R-squared (for linear regression), Deviance (for logistic regression),

AIC/BIC, and Residual analysis. Diagnostics help identify outliers, influential points, and multicollinearity. In

R, the car package provides useful tools like vif() (Variance Inflation Factor) to check multicollinearity and

outlierTest() to detect extreme points.

For logistic regression, model fit can be measured using Deviance (D):

퐷 = −2 × [ln⁡(퐿_full) − ln⁡(퐿_sat)]

------------>4

Where,



퐿_full= Likelihood of the fitted model,

퐿_sat= Likelihood of the saturated model (perfect fit).

Overview of Hosmer–Lemeshow Test:

The Hosmer–Lemeshow test is used to check the goodness-of-fit of a logistic regression model, that is, how well

the predicted probabilities match the observed outcomes. In this test, data is divided into g groups (usually

deciles) based on predicted probabilities, and observed and expected events in each group are compared. The

test statistic follows a chi-square distribution and is calculated as:

퐺

(푂_푔− 퐸_푔)²

퐶 = ∑ 푋

퐸_푔(1 − 퐸_푔/푛_푔)

푔=1

------------>5

Where,



푂_푔= Observed number of events in group g,

퐸_푔= Expected number of events in group g,

푛_푔= Number of observations in group g,

ꢁ= Total number of groups.

A p-value > 0.05 indicates a good fit, meaning the logistic regression model predicts the outcomes adequately.

This test is widely used by Indian researchers to verify if their logistic models are reliable and suitable for

prediction.

Overview of Roc Curve and AUC Analysis:

The ROC (Receiver Operating Characteristic) curve is used to evaluate the performance of a binary classification

model, such as predicting whether a shipping tourist chooses a green-marketed destination or not. The curve

plots True Positive Rate (Sensitivity) on the Y-axis against False Positive Rate (1 – Specificity) on the X-axis at

different probability thresholds. The Area Under the Curve (AUC) quantifies the overall ability of the model to

discriminate between the two classes. Its value ranges from 0 to 1, where a higher AUC (closer to 1) indicates

better model performance. The True Positive Rate and False Positive Rate are calculated as:

www.ijltemas.in

Page 741

INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,

MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)

ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XIV, Issue XI, November 2025

ꢂ푃

퐹푃

TPR (Sensitivity) =

,FPR =

ꢂ푃 + 퐹푁

퐹푃 + ꢂ푁

------------>6

Where,



TP = True Positives, FN = False Negatives,

FP = False Positives, TN = True Negatives.

Overview of Odds Ratio (OR) Interpretation:

Odds Ratio is used in logistic regression to measure how a one-unit change in an independent variable affects

the odds of an event occurring (e.g., choosing a green-marketed shipping destination). It is calculated by taking

the exponential of the logistic regression coefficient (β). The equation is:

푂푅 = 푒^ꢃ

------------>7

Where βis the estimated coefficient from the logistic regression. An OR > 1 indicates that as the predictor

increases, the likelihood of the event occurring increases; an OR < 1 indicates that the likelihood decreases; and

OR = 1 means no effect.

Overview of Mediation Analysis:

Mediation analysis is used to understand how an independent variable (IV) affects a dependent variable (DV)

through a mediator (M). It helps researchers find indirect effects along with direct effects. The basic model

involves three equations:

1.

Path a: Effect of IV on mediator

푀 = 훼 + 푎푋 + 푒₁

Path b and c′: Effect of mediator and IV on DV

----------->8

2.

⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡푌 = 훽 + 푐^′푋 + 푏푀 + 푒₂

------------>9

3.

Total effect (c):

푐 = 푐^′+ (푎 × 푏)

------------>10

Where,



푋= Independent variable (e.g., AGM),

푀= Mediator (e.g., Environmental Attitude),

푌= Dependent variable (e.g., Destination Choice),

푎= Effect of IV on mediator,

푏= Effect of mediator on DV,

푐^′= Direct effect of IV on DV,

www.ijltemas.in

Page 742

INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,

MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)

ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XIV, Issue XI, November 2025



푎 × 푏= Indirect effect of IV on DV via mediator.

Research Questions

 How reliable and consistent are the questionnaire items measuring green marketing, environmental

attitude, and destination choice among shipping tourists?



What is the strength and direction of the relationship between green marketing practices, environmental

attitude, and shipping tourists’ destination choice?

How do green marketing, environmental attitude, and purchase intention predict the likelihood of

shipping tourists choosing eco-friendly destinations?

How well does the logistic regression model fit the data, and does it meet goodness-of-fit criteria for

predicting shipping tourists’ destination choice?

Does environmental attitude mediate the relationship between green marketing and shipping tourists’

destination choice, and what is the magnitude of this indirect effect?

Objectives of The Study

• To assess the reliability and consistency of the questionnaire items measuring green marketing, environmental

attitude, and destination choice among shipping tourists.

• To examine the strength and direction of the relationships between green marketing practices, environmental

attitude, and shipping tourists’ destination choice.

• To determine how green marketing, environmental attitude, and purchase intention predict the likelihood of

shipping tourists choosing eco-friendly destinations.

• To evaluate the fit and adequacy of the logistic regression model and test whether it meets the required

goodness-of-fit criteria for predicting destination choice.

• To investigate whether environmental attitude mediates the relationship between green marketing and shipping

tourists’ destination choice and quantify the magnitude of this indirect effect.

Rational of The Study

Shipping tourism is growing rapidly in India as travellers increasingly prefer eco-friendly and responsible

tourism options, yet very limited research explains how green marketing actually influences their destination

choices. Existing studies highlight the importance of eco-labels, green branding, digital green promotions, and

authentic sustainability practices in shaping tourist behaviour, but their applicability to the shipping tourism

sector remains under-explored. Understanding this gap is important because cruise and boat-tourism operators

need clear, data-driven insights to attract environmentally conscious tourists and support sustainable tourism

development. Therefore, this study uses R-based logistic regression, model diagnostics, ROC analysis, and

mediation analysis to examine how green marketing, environmental attitude, and purchase intention influence

the probability of tourists choosing green-marketed shipping destinations. By providing scientific evidence on

the strength of these relationships, the study aims to help tourism operators design effective green marketing

strategies that build trust, improve environmental responsibility, and encourage sustainable destination choices

among shipping tourists.

Scope of The Study

The present study “Exploring How Green Marketing Influences Shipping Tourist Destination Choice: Evidence

from R-Based Logistic Regression and ROC Analysis “ focuses on understanding how green marketing

www.ijltemas.in

Page 743

INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,

MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)

ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XIV, Issue XI, November 2025

influences the destination choice of shipping tourists by using R-based statistical techniques such as logistic

regression and ROC analysis. The study covers key variables like green marketing practices, environmental

attitude, purchase intention, and eco-friendly destination choice among cruise, boat, and river-tourism travellers.

It includes an assessment of the reliability of the questionnaire items, the relationships among the major

constructs, and the predictive ability of green marketing factors in shaping tourist decisions. The scope also

extends to testing the model’s accuracy and goodness-of-fit using tools such as the Hosmer–Lemeshow test,

odds ratio, and AUC values. While the study is limited to shipping tourism and depends on self-reported data, it

provides practical insights for tourism operators, policymakers, and marketers who aim to promote sustainable

and eco-friendly shipping destinations.

Research Gap

Although previous studies have examined how green marketing influences tourist behaviour in general eco-

tourism settings, there is still very limited empirical evidence on how these practices specifically affect shipping

tourists’ destination choice, particularly in the context of cruises, river tourism, and boat-based travel. Existing

research mainly focuses on eco-labels, green promotions, environmental beliefs, or sustainable tourism

intentions in land-based tourism, but does not analyse how these factors operate in the unique environment of

shipping tourism, where safety, mobility, and destination ecosystems differ significantly. Moreover, prior studies

have rarely applied advanced statistical techniques such as R-based logistic regression, ROC curve analysis, and

mediation modelling to predict and validate tourist decision-making in green-marketed maritime destinations.

The literature also lacks an understanding of how environmental attitude acts as a mediator between green

marketing cues and actual destination choices among shipping tourists. Therefore, a clear gap exists in

integrating green marketing variables, psychological mediators, and predictive statistical modelling to explain

how tourists choose eco-friendly shipping destinations—an area this study aims to address.

RESEARCH METHODOLOGY

This study, titled “Exploring How Green Marketing Influences Shipping Tourist Destination Choice: Evidence

from R-Based Logistic Regression and ROC Analysis,” follows a quantitative research design using a structured

questionnaire to collect data from shipping tourists in India. The questionnaire includes items on green marketing

practices, environmental attitude, purchase intention, and destination choice. First, the reliability and internal

consistency of all scales are tested using Cronbach’s Alpha in R. Correlation analysis is then used to examine

the strength and direction of relationships among the key variables. To predict the probability of tourists choosing

eco-friendly shipping destinations, an R-based logistic regression model is applied, supported by odds ratio

interpretation, multicollinearity checks, and model diagnostics such as AIC, Deviance, VIF, and the Hosmer–

Lemeshow goodness-of-fit test. ROC curve and AUC analysis are used to evaluate the classification accuracy

of the model. Further, mediation analysis is conducted to test whether environmental attitude acts as a mediator

between green marketing and destination choice. All statistical procedures, including reliability testing,

correlation, logistic regression, mediation analysis, and ROC evaluation, are performed using R programming

to ensure accuracy, transparency, and replicability of the research findings.

Research Limitations

1. The study relies on self-reported data, which may contain response bias or socially desirable answers from

environmentally conscious tourists.

2. The sample size is limited to selected shipping tourism locations, reducing the generalisability of the findings

to all maritime tourism segments.

3. Logistic regression results depend on the selected variables, and unobserved factors such as income, travel

motivation, or cultural differences were not included in the model.

4. The analysis uses cross-sectional data, which restricts the ability to establish long-term behavioural changes

or causal relationships.

www.ijltemas.in

Page 744

INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,

MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)

ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XIV, Issue XI, November 2025

5. ROC and model diagnostics evaluate prediction accuracy but cannot fully capture real-world complexities of

tourist decision-making in dynamic environmental contexts.

Analysis and Interpretation

>library(readxl)

>Book1 <- read_excel("D:/Research Paper/4/Book1.xlsx")

>View(Book1)

Reliability Test (Alpha)

> df <- Book1 # rename for safety

> alpha(df[, c("AGM", "GMP", "EA", "PTI", "DC")])

Reliability analysis

>Call: alpha(x = df[, c("AGM", "GMP", "EA", "PTI", "DC")])

Table 1: Reliability Test (ALPHA)

Metric

raw.alpha

std.alpha

Value

Description

0.90

The standard, unstandardized Cronbach's Alpha coefficient.

The standardized Cronbach's Alpha coefficient (calculated if all

items were standardized to have a variance of 1).

0.93

Guttman's Lambda 6, which uses the squared multiple correlation

(SMC) to estimate item communalities; often considered a sharper

lower bound for true reliability than alpha.

G6(smc)

0.65

9.4

The average inter-item correlation.

average_r

S/N

The Signal-to-Noise Ratio (ratio of reliable variance to error

variance).

0.0093

3.8

The Asymptotic Standard Error for Cronbach's Alpha.

The average score across all items.

ase

mean

sd

0.73

0.68

The standard deviation of the item scores.

The median inter-item correlation.

median_r

95% confidence boundaries

Method

Lower Confidence

Boundary

Alpha (α)

Upper Confidence Boundary

0.88

0.90

0.92

Feldt

Duhachek

www.ijltemas.in

Page 745

INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,

MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)

ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XIV, Issue XI, November 2025

Reliability if an item is dropped

ID

raw_alp

ha

std.alp

ha

G6(smc)

averag

e_r

S/N

alpha se

var.r

med.r

0.87

0.90

0.87

0.85

0.88

0.90

0.91

0.87

0.85

0.89

0.90

0.93

0.89

0.90

0.64

0.69

0.70

0.63

0.60

7.1

8.8

9.6

6.9

5.9

0.0120

0.0092

0.0097

0.0125

0.0147

0.034

0.016

0.022

0.033

0.038

0.68

0.71

0.73

0.68

0.55

AGM

GMP

EA

PTI

DC

Item statistics

r.drop

0.78

n

raw.r

0.87

0.81

0.77

0.87

0.93

std.r

0.87

0.80

0.77

0.88

0.93

r.cor

0.85

0.78

0.71

0.86

0.92

mean

sd

325

3.7

3.8

3.7

3.8

0.89

0.96

0.84

0.77

0.86

AGM

GMP

EA

0.68

0.65

0.81

PTI

0.88

DC

The reliability analysis shows that the overall Cronbach’s Alpha is 0.90, indicating excellent internal consistency

among the items measuring green marketing (AGM, GMP), environmental attitude (EA), purchase intention

(PTI), and destination choice (DC). The 95% confidence interval (0.88–0.92) further confirms the stability of

this reliability. Item-wise results show that all items have strong item-total correlations (raw.r = 0.77 to 0.93),

meaning each variable contributes meaningfully to the scale. The “alpha if item dropped” values (0.85–0.90)

indicate that removing any item does not improve overall reliability, proving that all items are important for the

construct. Overall, the instrument used in this study is statistically reliable, consistent, and suitable for further

correlation and regression analysis.

Correlation Analysis

> cor(Book1[, c("AGM","GMP","EA","PTI","DC")])

Table 2: Correlation

AGM

GMP

EA

PTI

DC

1.0000000

0.4780675

0.7771078

0.5946515

0.4780675

1.0000000

0.3540343

0.8682065

0.7771078

0.3540343

1.0000000

0.4982750

0.5946515

0.8682065

0.4982750

1.0000000

0.8323478

0.6989881

0.6595951

0.7569288

AGM

GMP

EA

PTI

www.ijltemas.in

Page 746

INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,

MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)

ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XIV, Issue XI, November 2025

0.8323478

0.6989881

0.6595951

0.7569288

1.0000000

DC

The correlation analysis shows the strength and direction of relationships among the study variables: green

marketing (AGM, GMP), environmental attitude (EA), purchase intention (PTI), and destination choice (DC).

AGM has a strong positive correlation with DC (r = 0.832) and EA (r = 0.777), indicating that higher awareness

of green marketing is associated with stronger environmental attitudes and higher likelihood of choosing eco-

friendly destinations. GMP is moderately correlated with DC (r = 0.699) and highly correlated with PTI (r =

0.868), suggesting that green marketing practices influence tourists’ purchase intentions and destination

selection. EA shows a moderate positive correlation with DC (r = 0.660), highlighting that environmentally

conscious tourists are more likely to select green destinations. PTI has a strong positive correlation with DC (r

= 0.757), reflecting that tourists’ intention to purchase or engage with eco-friendly offerings significantly affects

their destination choice. Overall, all correlations are positive, indicating that green marketing, environmental

attitude, and purchase intention positively influence shipping tourists’ destination choice, supporting the

rationale for further logistic regression analysis.

Logistic Regression

> model1 <- glm(DC ~ AGM + GMP + EA + PTI,

+

data = Book1,

family = binomial)

> summary(model1)

Call:

glm(formula = DC ~ AGM + GMP + EA + PTI, family = binomial, data = Book1)

Table 3: Logistic Regression

Coefficients

Estimate

Std. Error

z value

P r(>|z|)

-1.40017

0.67731

0.23255

0.23805

0.23422

0.34926

-2.067

0.0387 *

(Intercept)

0.22717

-0.05385

0.16883

-0.16254

0.977

-0.226

0.721

-0.465

0.3286

0.8210

0.4710

0.6417

AGM

GMP

EA

PTI

414.65 on 324 degrees of freedom

409.49 on 320 degrees of freedom

419.49

Null deviance

Residual deviance

AIC

4

Number of Fisher Scoring iterations

The logistic regression model examines the effect of green marketing awareness (AGM), green marketing

practices (GMP), environmental attitude (EA), and purchase intention (PTI) on shipping tourists’ destination

choice (DC). The intercept is significant (Estimate = -1.400, p = 0.0387), indicating the baseline log-odds of

www.ijltemas.in

Page 747

INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,

MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)

ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XIV, Issue XI, November 2025

choosing a green destination when all predictors are zero. None of the independent variables (AGM, GMP, EA,

PTI) are statistically significant at the 5% level (p > 0.05), suggesting that individually, they do not have a strong

predictive effect on destination choice in this model. The signs of the coefficients indicate positive relationships

for AGM (0.227) and EA (0.169) with DC, meaning higher awareness and stronger environmental attitude may

increase the likelihood of choosing a green destination, while GMP (-0.054) and PTI (-0.163) show negative,

though non-significant, effects. The reduction in deviance from the null model (Null deviance = 414.65; Residual

deviance = 409.49) indicates a slight improvement in model fit, but the change is modest. The AIC value of

419.49 suggests the model is reasonably parsimonious but may require additional predictors or larger sample

size for stronger predictive power. Overall, the model indicates a trend where green marketing and environmental

attitude positively influence destination choice, but the effects are not statistically significant in this sample.

Model Fit & Diagnostics

install.packages("car")

library(car)

>vif(model1)

Table 4: Model Fit & Diagnostics

AGM

2.897595

GMP

3.722703

EA

2.592768

PTI

4.312448

The Variance Inflation Factor (VIF) values for the predictors—AGM (2.90), GMP (3.72), EA (2.59), and PTI

(4.31)—are all below the commonly accepted threshold of 5. This indicates that multicollinearity is not a serious

concern in the logistic regression model. Each independent variable contributes unique information, and their

estimates are not unduly inflated due to correlation with other predictors. Therefore, the model coefficients can

be considered stable and reliable for interpreting the relationship between green marketing, environmental

attitude, purchase intention, and shipping tourists’ destination choice.

Hosmer–Lemeshow Test

> install.packages("ResourceSelection", dependencies = TRUE)

>library(ResourceSelection)

>hoslem.test(Book1DC, fitted(model1))

>Hosmer and Lemeshow goodness of fit (GOF) test

> data: Book1DC, fitted(model1)

Table 5: Hosmer–Lemeshow Test

Statistic

Value

Degrees of Freedom (df)

p-value

χ²(X-squared)

80.252

8

5.352

The Hosmer–Lemeshow goodness-of-fit test evaluates how well the logistic regression model fits the observed

data. Here, the test yields X² = 80.252, df = 8, and a p-value = 5.352. Since the p-value is greater than 0.05, the

result indicates that there is no significant difference between the observed and predicted values, meaning the

model fits the data adequately. This suggests that the logistic regression model provides a reasonable

www.ijltemas.in

Page 748

INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,

MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)

ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XIV, Issue XI, November 2025

representation of the relationship between green marketing, environmental attitude, purchase intention, and

shipping tourists’ destination choice.

ROC Curve & AUC Analysis

>install.packages("pROC")

> library(pROC)

> pred <- predict(model1, type = "response")

> roc_obj <- roc(Book1DC, pred)

Setting levels: control = 0, case = 1

Setting direction: controls < cases

> plot(roc_obj)

> auc(roc_obj)

Area under the curve: 0.7845

Chart1: Roc Curve

1.5

1.0

0.5

0.0

-0.5

Specificity

The Area Under the Curve (AUC) measures the overall performance of the model across all possible

classification thresholds. It can be interpreted as the probability that the model will rank a randomly chosen

positive instance higher than a randomly chosen negative instance. Model Performance - Since the AUC is

0.7845, which is significantly greater than 0.5, the model performs better than random chance. An AUC value

between 0.7 and 0.8 is generally considered acceptable or fair discrimination for a diagnostic or predictive model.

Discrimination Power - The model exhibits a decent ability to distinguish between the positive and negative

classes. An AUC of ~0.78 suggests that there is a 78.45% probability that a randomly chosen positive case will

be scored higher by the model than a randomly chosen negative case.

Visual Confirmation - The plotted

black curve lies mostly above the diagonal line. This visual positioning confirms the quantitative AUC value,

indicating that the model's true positive rate (Sensitivity) is generally higher than its false positive rate (1-

Specificity), demonstrating its predictive utility. In summary, the model shows fair predictive ability with an

AUC of 0.7845, successfully differentiating between classes better than a purely random prediction.

Odds Ratio Interpretation

>exp(coef(model1))

www.ijltemas.in

Page 749

INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,

MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)

ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XIV, Issue XI, November 2025

(Intercept)

AGM

GMP

EA

PTI

0.2465548 1.2550490 0.9475761 1.1839169 0.8499825

> odds_ratios <- exp(coef(logit_model))

> # Wald 95% CI

> se <- summary(logit_model)coefficients[, "Std. Error"]

> OR_lower <- exp(coef(logit_model) - 1.96 * se)

> OR_upper <- exp(coef(logit_model) + 1.96 * se)

> # Combine in a table

> results <- data.frame(

+

Variable = names(coef(logit_model)),

OR = odds_ratios,

+

CI_lower = OR_lower,

+

CI_upper = OR_upper

+ )

> results

Table 6: Odds Ratio

Term

(Intercept)

AGM

GMP

EA

Exponentiated Coefficient

Odds Ratio (OR)

2.900701

CIlower

CIupper

Inf

0.2465548

1.2550490

0.9475761

1.1839169

0.8499825

0

1.000000

Inf

1.000000

Inf

1.000000

Inf

1.000000

Inf

PTI

The provided table gives the Odds Ratios (OR) and their corresponding 95% Confidence Intervals (CI) for

several variables. The interpretation of these odds ratios is severely limited and problematic because the

confidence intervals for every single variable—(Intercept), AGM, GMP, EA, and PTI—span from a lower bound

of 0 to an upper bound of Inf (infinity). This structure strongly suggests a fundamental issue with the statistical

model, likely due to perfect separation (where a predictor perfectly predicts the outcome), which results in

infinite standard errors and unreliable point estimates for the ORs. Specifically, the OR values of 1.000000 for

AGM, GMP, EA, and PTI indicate that these variables have no association with the outcome, as an OR of 1

means the odds of the outcome are the same regardless of the predictor's value; however, due to the CI _lower=0

and CI _upper=∞, this lack of association is not statistically confirmed, and the entire model's output should be

considered non-interpretable and invalid for making any predictive inference.

www.ijltemas.in

Page 750

INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,

MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)

ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XIV, Issue XI, November 2025

Classification Table & Cutoff Analysis

> pred_class <- ifelse(pred > 0.5, 1, 0)

> table(Book1DC, pred_class)

Table 7: Classification Table & Cutoff Analysis

Predicted Class 0

Predicted Class 1

0 (False Positive)

1 (True Positive)

1

Total

216

216 (True Negative)

108 (False Negative)

324

Actual Class 0

Actual Class 1

Total

109

325

The provided classification table, generated using a cutoff of 0.5, reveals significant issues with the model's

ability to correctly identify the positive class (labelled as 1). The table shows that out of 216 true negatives

(actual value 0), the model correctly classified 216 as 0 (True Negatives, TN) and 0 as 1 (False Positives, FP).

However, out of 109 true positives (actual value 1), the model incorrectly classified 108 as 0 (False Negatives,

FN) and only 1 as 1 (True Positive, TP). This results in a Sensitivity (Recall) of only 1/109 ~ 0.92% (poorly

identifying the positive class) and a Specificity of 216/216 = 100% (perfectly identifying the negative class).

The model is heavily biased towards predicting the negative class (0), making it virtually useless for its primary

goal of identifying positive cases at this specific cutoff of 0.5.

Mediation Analysis

> install.packages("mediation")

> library(mediation)

> Book1DC <- ifelse(Book1DC >= 4, 1, 0)

> med.fit <- lm(EA ~ AGM + GMP + PTI, data = Book1)

> summary(med.fit)

>Book1DC_bin <- ifelse(Book1DC >= 4, 1, 0)

> install.packages("logistf")

> library(logistf)

> fit <- logistf(DC_bin ~ EA + AGM, data = Book1)

> med.fit <- lm(EA ~ AGM, data = Book1)

> out.fit <- glm(DC_bin ~ EA + AGM, data = Book1, family = binomial)

> med.out <- mediate(med.fit, out.fit, treat = "AGM", mediator = "EA", boot = TRUE, sims = 1000)

> summary(med.out)

Call:

lm(formula = EA ~ AGM + GMP + PTI, data = Book1)

www.ijltemas.in

Page 751

INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,

MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)

ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XIV, Issue XI, November 2025

Table 8: Mediation Analysis

Residuals

1Q (1st Quartile) Median

Min

3Q (3rd Quartile) Max

0.23925 2.12756

-2.66643

-0.18264

0.04426

Coefficients

Estimate

0.84967

0.68631

-0.18132

0.29249

Std. Error z value

P r(>|z|)

0.15276

0.03948

0.05751

0.08391

5.562

17.383

-3.153

3.486

0e-08 ***

2e-16 ***

0.001771 **

0.000559 ***

(Intercept)

AGM

GMP

PTI

0.5184 on 321 degrees of freedom

Residual standard error:

Multiple R-squared

F-statistic

0.6187, Adjusted R-squared: 0.6152

173.6

321

DF

< 2.2e-16

p-value

The mediation analysis model demonstrates a strong fit and highly significant predictive power, as indicated by

the Multiple R-squared of 0.6187 (meaning the predictors explain about 61.9% of the variance in the outcome)

and the highly significant F-statistic (p-value < 2.2e-16). All three predictors are statistically significant (p <

0.01 or better): AGM has the largest positive effect (β= 0.686), significantly increasing the outcome; PTI also

has a significant positive effect (\beta = 0.292); and GMP has a significant negative effect (β= -0.181), decreasing

the outcome. The residuals are centred around zero (Median = 0.04426) and appear relatively symmetric,

suggesting the model assumptions are reasonably met, with a Residual Standard Error of 0.5184 indicating the

typical size of the error in prediction. This initial step confirms that the core predictor variables are strongly

related to the dependent variable, forming a sound basis for proceeding with the full mediation analysis.

Out Fit

> out.fit <- glm(DC ~ EA + AGM + GMP + PTI,

+

data = Book1,

family = binomial)

> summary(out.fit)

Call:

glm(formula = DC ~ EA + AGM + GMP + PTI, family = binomial, data = Book1)

www.ijltemas.in

Page 752

INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,

MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)

ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XIV, Issue XI, November 2025

Table 9: Mediation Analysis- Outfit Model

Coefficients

Std. Error

Estimate

-1.40017

0.16883

0.22717

-0.05385

-0.16254

z value

-2.067

0.721

P r(>|z|)

0.0387*

0.4710

0.3286

0.8210

0.6417

0.67731

0.23422

0.23255

0.23805

0.34926

(Intercept)

EA

0.977

AGM

-0.226

-0.465

GMP

PTI

414.65 on 321 degrees of freedom

Null deviance

Residual deviance

AIC

409.49 on 321 degrees of freedom

419.49

321

4

DF

Number of Fisher Scoring iterations

This output represents the results of a logistic regression model (indicated by the binomial family and z-values),

which is typically the final step in a mediation analysis where the outcome variable is binary. The analysis

suggests that none of the predictor variables (EA, AGM, GMP, and PTI) have a statistically significant

relationship with the binary outcome, as their p-values are all much greater than the conventional 0.05 threshold

(ranging from 0.3286 to 0.8210). While the Intercept is significant (p = 0.0387), it only indicates the log-odds

of the outcome when all predictors are zero, and it doesn't support the predictive utility of the independent

variables. Furthermore, the Residual Deviance (409.49) is very close to the Null Deviance (414.65), which

implies that adding these predictors to the model does not significantly improve its fit compared to a model with

only the intercept, collectively suggesting that the proposed direct and/or indirect effects examined in the

mediation model are not supported by the data at the outcome level.

Table 10: Model Fitted by Penalized ML

Coefficient coef

se(coef)

lower

0.95

upper

0.95

χ2

p

method

2.1952214

2.9354001 31.444963 4.947442 0.3290723 0.5662059 2

(Intercept)

EA

0.5167969

1.6120063

0.9013599 6.857126

0.8608817 5.880667

8.207918 0.2184102 0.6402539 2

5.573809 0.8083186 0.3686178 2

p-value

AGM

Test

Test Statistic

df (Degrees of

Freedom)

1.207015

30.91114

2

0.54689

Likelihood Ratio Test

Wald Test

1.939684

www.ijltemas.in

Page 753

INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,

MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)

ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XIV, Issue XI, November 2025

The results from the Model fitted by Penalized Maximum Likelihood indicate that the overall model does not

significantly predict the outcome, as shown by the Likelihood Ratio Test (LRT) which is not statistically

significant X²= 1.207 p = 0.54689). For the individual predictors, both EA (β^= 0.5168) and AGM (β^ = -

1.6120) are not statistically significant, as the 95% confidence intervals for both variables are extremely wide

(e.g., EA: CI = [-6.857, 8.208]) and clearly include zero, suggesting that their true effect on the outcome could

be positive, negative, or zero. While the Wald test for the overall model reports a highly significant result (p <

0.0001), this finding is contradictory to the LRT and the individual confidence intervals, and it is generally less

reliable for penalized models, leading to the inference that the model lacks sufficient evidence to support the

predictive utility of the independent variables. The results from the Model fitted by Penalized Maximum

Likelihood indicate that the overall model does not significantly predict the outcome, as shown by the Likelihood

Ratio Test (LRT) which is not statistically significant X2 = 1.207 p = 0.54689). For the individual predictors,

both EA (β^= 0.5168) and AGM (β^ = -1.6120) are not statistically significant, as the 95% confidence intervals

for both variables are extremely wide (e.g., EA: CI = [-6.857, 8.208]) and clearly include zero, suggesting that

their true effect on the outcome could be positive, negative, or zero. While the Wald test for the overall model

reports a highly significant result (p < 0.0001), this finding is contradictory to the LRT and the individual

confidence intervals, and it is generally less reliable for penalized models, leading to the inference that the model

lacks sufficient evidence to support the predictive utility of the independent variables.

Table 11: Causal Mediation Analysis

Nonparametric Bootstrap Confidence Intervals with the Percentile Method

Effect

Estimate

Lower 95% CI

Upper 95% CI

p-value

8.1127 x 10^-27

8.0494 x 10^-27

1.0335 x 10^-26

1.0271 x 10^-26

1.8384 x 10^-26

0.44129

1.3519 x 10^-25

1.3523 x 10^-25

2.1666 x 10^-25

2.1667 x 10^-25

1.0718 x 10^-25

9.3599

1.7161 x 10^-25

1.7158 x 10^-25

1.8543 x 10^-25

1.8546 x 10^-25

9.5357 x 10^-25

7.7609

0.888

0.808

0.920

NA

ACME (control)

ACME (treated)

ADE (control)

ADE (treated)

Total Effect

Prop. Mediated (control)

0.43784

9.3598

7.7622

NA

Prop. Mediated (treated)

8.0810 x 10^-26

1.0303 x 10^-26

0.43957

1.3521 x 10^-26

2.1667 x 10^-25

9.3599

1.7159 x 10^-25

1.8547 x 10^-25

7.7616

0.888

0.808

NA

ACME (average)

ADE (average)

Prop.Mediated (average)

The results from the Causal Mediation Analysis using the Nonparametric Bootstrap Percentile Method indicate

that there is no statistically significant causal effect detected in the mediation model. All key components of the

effect—the Average Causal Mediation Effect (ACME), the Average Direct Effect (ADE), and the Total Effect—

show estimates that are extremely close to zero (in the order of 10-26 or 10-27. Crucially, the p-values are very

high (ranging from 0.808 to 0.920), and the 95% confidence intervals for all effects strongly contain zero,

confirming the lack of significance. While the Proportion Mediated is estimated around 44%, the confidence

intervals are extremely wide (CI ~ [-9.36, 7.76]) and span negative to highly positive values, rendering the

proportion estimate meaningless and non-interpretable. In conclusion, based on the non-significant p-values and

www.ijltemas.in

Page 754

INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,

MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)

ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XIV, Issue XI, November 2025

confidence intervals, the analysis fails to find evidence of a significant indirect effect (mediation) or a significant

direct effect of the exposure on the outcome.

Chart 2: Mediation Analysis

1.5

1.0

0.5

0.0

-0.5

Specificity

CONCLUSION AND SUGGESTION

The study successfully established the excellent internal consistency of its measurement instrument (Cronbach’s

α= 0.90) and confirmed strong positive correlations between Green Marketing Awareness (AGM: r=0.832),

Environmental Attitude (EA: r=0.660), Purchase Intention (PTI: r=0.757), and Shipping Tourist Destination

Choice (DC). However, the subsequent logistic regression model, despite showing a fair overall predictive ability

(AUC of 0.7845) and an adequate fit (Hosmer-Lemeshow p > 0.05), failed to find statistically significant

individual effects for any of the predictors (AGM, GMP, EA, PTI) on Destination Choice, nor did the Causal

Mediation Analysis find any significant direct or indirect effects. This suggests that while a positive correlation

exists, the predictive power of these factors in the full model is weak, which is further supported by the model's

poor sensitivity (0.92%) at the 0.5 cutoff, indicating an inability to correctly classify positive destination choices.

Future research should focus on addressing the limitations of the current study, specifically by re-evaluating the

measurement of the dependent variable (Destination Choice) which was likely responsible for the model's poor

sensitivity and the issues of non-interpretable odds ratios. Researchers should increase the sample size and

expand the study to different maritime segments beyond the limited locations used here to improve the

generalizability of the findings. Additionally, future models should incorporate unobserved factors such as

tourists' income, specific travel motivations, or cultural differences, as these may provide the necessary

predictive power that the current set of variables lacks in predicting the binary outcome.

Practical Implications

Despite the statistical non-significance in the predictive model, the strong positive correlations found (e.g., AGM

and DC: r=0.832) imply that shipping tourism operators should continue to invest in clear green marketing and

eco-friendly operations, as these are positively linked to the attitudes and intentions of potential customers.

Operators must move beyond mere advertising ("promotion") and ensure their "product" and "place" offerings

are authentically green to reinforce the strong correlation between Green Marketing Practices (GMP) and

Purchase Intention (PTI: r=0.868). Improving the tangibility of green offerings is essential to translate positive

attitudes into actual destination choice, thereby enhancing trust and promoting environmental responsibility.

Social Implications

The study's findings contribute to the broader goal of sustainable tourism development by highlighting that

tourists' Environmental Attitude is a critical, highly correlated factor (r=0.660) in their destination choice. The

www.ijltemas.in

Page 755

INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,

MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)

ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XIV, Issue XI, November 2025

results encourage policymakers and tourism bodies to not only support green certification for shipping operators

but also to implement educational campaigns that enhance tourists' environmental knowledge and pro-

environmental behavior. By strengthening the environmental beliefs of tourists, destination management can

cultivate a segment of ethically conscious consumers, ensuring the long-term sustainability and protection of

unique maritime ecosystems.

Originality and Value

This study holds significant original value by addressing a key research gap: the empirical influence of green

marketing specifically on shipping tourists' destination choice (cruises/river travel), an area previously under-

explored compared to land-based eco-tourism. It is one of the first studies in this domain to integrate a set of

green marketing variables with psychological mediators (Environmental Attitude) and validate the model using

advanced R-based statistical techniques like logistic regression, VIF diagnostics, Hosmer-Lemeshow testing,

and ROC curve analysis. This provides a new methodological framework and preliminary evidence that, despite

poor model classification, the correlation between green factors and destination choice in the maritime sector is

robust.

REFERENCES

1. Kumar, S., & Devi, N. (2023). The effect of green marketing tools (eco-labels, eco brands, eco

advertisements) on green consumer behaviour among tourists in Himachal Pradesh, India. AIMT Journal

of Management, 12(1–2).

2. Karina, M., Rivanto, R., Basri, H., & Iryani, E. (2025). Green marketing can transform conventional

tourism into sustainable tourism. Branding: Jurnal Manajemen dan Bisnis, 4(1).

3. Yu, C. C., & Liu, C. C. (2024). A green marketing model for the tourism and amusement industry using

fuzzy analytic hierarchy process. International Journal of Religion, 5(4), 224–230.

4. Sanjaya, D., Arief, M., Setiadi, N. J., & Heriyati, P. (2024). The role of digital green marketing

campaigns and environmental beliefs on tourist behavior and revisit intentions in eco-tourism. Journal

of Eastern European and Central Asian Research (JEECAR), 11(3), 553–572.

5. Jumadi, Aditya, A., Saputra, A., & Burhani, A. I. (2023). The effect of green marketing (product, place,

promotion) on green tourism. In Proceedings of the 2nd UPY International Conference on Education

and Social Science (pp. 471–477). Atlantis Press.

6. Rahayu, S., Aliyah, H., & Sudarwati. (2024). The effect of green marketing components (green product,

green promotion, green price) and environmental knowledge on intention to visit eco-tourism sites.

International Journal of Economics, Business and Accounting Research (IJEBAR), 6(1).

7. Mdpi. (2023). Green and environmental marketing strategies and ethical consumption. Sustainability,

15(16).

www.ijltemas.in

Page 756