Study on Interrelationship and Relative Importance of Factors Influencing Overall Comfort in Aircraft Cabins

Abstract

To explore the relationships among factors influencing overall comfort in aircraft cabins, a scale-based method was developed. First, a weighted multi-level measurement model was used to determine the hierarchical relationships between comfort factors. Second, a hybrid path analysis model was constructed to analyze the relationships between comfort factors and their impact on overall comfort. Finally, the validity and stability of the results were evaluated using multigroup analyses. The results show that service experience has the greatest impact on overall comfort, with food and beverage being the most influential. Additionally, both service experience and the physical environment affect passengers’ perception of cabin facilities. This study identifies key factors for optimizing overall comfort. Furthermore, the methods used are applicable for analyzing overall comfort in public transportation systems in general.

1. Introduction

Service quality has become a crucial factor in influencing the choice of public transportation [1,2,3]. Comfort is a key determinant of passengers’ acceptance of transportation systems, impacting their satisfaction and willingness to use the service again [4,5,6]. Research on comfort has evolved to consider both physiological and psychological aspects, influenced by the physical and social environment within the cabin [7]. The high-quality market positioning of air travel and the unique characteristics of the operating environment make factors affecting cabin comfort more complex, thus creating a multifaceted impact mechanism on overall comfort. Understanding these influencing factors and their relationships can provide insights into optimizing the comfort of civilian aircraft cabins, clarifying the paths and effects of both direct and indirect influences.

1.1. Concept of Comfort and Factors Affecting Cabin Comfort

Comfort and discomfort reflect passengers’ physiological and psychological perceptions of their surrounding physical and social environment [8,9]. The number of factors influencing comfort has grown, much like Maslow’s Hierarchy of Needs [10]. Once basic security needs are satisfied, individuals seek higher-level needs, such as quality service and social interaction [11,12,13]. Helander and Zhang et al. [14] suggest that discomfort is often related to the physical environment, such as seating posture and fatigue, while comfort stems from experiences that exceed expectations, such as additional services and the quality of meals.

Several studies have examined the factors affecting cabin comfort. Chen [15] identified onboard staff service and facilities as the most significant determinants of airline service quality, followed by service products. Vink [16] found that factors like legroom, hygiene, crew attentiveness, and personal space impact ride comfort. An [17] and Messner [18] noted that catering is a key factor that affects the satisfaction of aviation service quality. Additionally, Li et al. [19] emphasized that facilities, especially the in-flight entertainment system, significantly affect cabin comfort, while Wang et al. [20] identified various factors including physical environment, seats, interior layout, services, catering, and entertainment systems as contributors to comfort.

To differentiate specific aspects such as seat comfort, environmental comfort, and service comfort, the concept of “overall comfort” is introduced in this work. Overall comfort refers to the passengers’ holistic subjective evaluation of comfort, encompassing the cabin’s physical environment, seating, interior layout, food and beverages, crew service, and in-flight entertainment (IFE). This concept aligns more closely with passengers’ overall cabin experience. However, it lacks detailed suggestions for improving comfort, making it essential to identify key factors through studying the relationships between these comfort elements.

1.2. Methods of Investigating Overall Comfort

Unlike single-factor comfort studies, research on overall comfort acknowledges the interconnectedness and interaction of various influencing factors. Examining how these factors interact may reveal the degree of their influence and clarify the mechanisms behind overall comfort. However, the diversity of these factors, the ambiguity surrounding them, and the complexity of their interrelations make studying overall comfort challenging.

Previous studies have decomposed comfort factors into indicators to help respondents understand complex comfort dimensions and translate evaluation results into actionable design suggestions [20,21]. However, deconstructing comfort into numerous variables increases research complexity. Some studies have employed factor analysis to reduce dimensionality, grouping related indicators based on subjective judgments or statistical relationships, which results in latent variables [22,23]. These latent variables are crucial for building multi-level measurement models that help simplify overall comfort research.

Here are two primary approaches for modeling the impact of overall comfort: one involves linear regression or correlation analysis to examine individual factors [24,25]. However, comfort factors are rarely independent; they interact synergistically. Hence, research on overall comfort faces significant challenges in simultaneously modeling multiple determinants and analyzing multi-path mechanisms (direct and indirect effects). This issue can be addressed using advanced path analysis techniques, which analyze complex relationships and their collective impact on overall comfort [7,22,26].

1.3. Modeling the Relationship Between Factors Influencing Overall Comfort

Several studies have developed models to explore how various factors influence comfort in public transport. Lee et al. [16] used structural equation modeling (SEM) to examine the overall impact of physical environmental factors on the subjective ride experience in high-speed trains. Chen [15] used SEM to develop an aviation service quality model to explore the relationships between subjective factors such as service quality, perceived value, satisfaction, and behavioral intention. Cheng [27] applied SEM to analyze the impact of convenience, comfort, safety, service, and economy on bus transfer service quality. These studies, while varying in their assumptions, demonstrate the feasibility of using SEM to model causal relationships between comfort factors.

Before constructing a model, it is essential to clarify ambiguous concepts such as “overall comfort” and simulate the hierarchical relationships among influencing factors. Furthermore, after model construction, its validity must be assessed. Comfort surveys often cover diverse subgroups, but it remains unclear whether differences in data structure across these groups affect the model’s reliability [28]. Few studies have explored model invariance across specific aviation contexts. Differences in cabin configurations (wide-body versus narrow-body layouts) and flight distances (short versus long-haul) can significantly influence the structure and paths of passenger perception [29,30]. Consequently, a robust overall comfort model must undergo cross-group testing to ensure its validity and stability across different aircraft types and flight distances [31].

1.4. Current Study

This study aims to identify the interrelationships between factors influencing overall passenger comfort and to determine their relative importance using civil aircraft cabins as an empirical case. The study is organized around three primary research objectives:

(1)

To establish a hierarchical measurement model by decomposing complex comfort factors into quantifiable indicators;

(2)

To construct a hybrid path analysis model to estimate both the direct and indirect effects of internal environment, facilities, and service experience on overall comfort; and

(3)

To validate the model’s effectiveness and stability across different aircraft types and flight distances through multigroup analysis.

Section 2 details the proposed methodology. Section 3 presents the model fit results and cross-group validation. Finally, Section 4 discusses the theoretical and practical implications of the findings.

2. Materials and Methods

2.1. Description of Method

In this study, a method is proposed to identify the relationships between factors influencing overall comfort and to assess the relative importance of these factors through the analysis of subjective survey-scale data. The method is divided into three main parts (Figure 1). First, the factors influencing overall comfort were decomposed into distinct indicators. A weighted multi-level measurement model was used to define the hierarchical relationships between the factors that affect overall comfort. Second, based on the hybrid path analysis model, the potential paths of relationships between the model variables were proposed, and an initial hypothesized model of overall comfort was constructed. The Maximum Likelihood method was employed to fit the initial hypothesized model using survey data, and the model’s fit was tested. Third, multi-group analysis was used to assess the path invariance of the overall comfort model across different groups, validating the model’s robustness and stability.

Figure 1. Method for researching overall comfort in civil aircraft cabin.

This study uses civil aircraft cabins as an example to demonstrate the steps of the methodology and explore the interrelationships and relative importance of factors influencing overall comfort.

2.2. Hierarchical Relationship of Factors Influencing Overall Comfort

2.2.1. Decomposing Factors Affecting Overall Comfort in Civilian Aircraft

(1)

Decomposing factors into indicators

Previous studies have identified that factors affecting passengers’ physical and psychological comfort in the cabin include both physical factors (such as the environment, seating, interior layout, food, and beverages) and social factors (such as crew service and IFE). To provide a clearer understanding of overall comfort, these cabin factors were disassembled into measurable indicators. The indicators were refined based on expert opinions [32]. Experts included five frequent fliers and three transportation facility human-factor design specialists. Among the specialists, two were from the civil aviation industry’s cabin business departments, and one was from the cabin services department of an airline. The frequent fliers were from academic institutions focused on human-factor research in transportation.

(2)

Questionnaire design and indicator optimization

Two types of questionnaires were designed: a satisfaction questionnaire and an importance questionnaire. The satisfaction questionnaire assessed passengers’ satisfaction with each indicator and with the overall comfort on a five-point scale, ranging from “1 = very dissatisfied” to “5 = very satisfied” [33,34]. The importance questionnaire asked respondents to rate the importance of each factor on a scale of “1 = very unimportant” to “5 = very important.”

SPSS Statistics 24 was used to analyze the reliability of each questionnaire. Cronbach’s Alpha was used to assess the internal consistency of the items in each questionnaire. The rationality of the items was tested, and 93 wide-body aircraft indicators and 80 narrow-body aircraft indicators affecting overall comfort were identified (Table 1).

Table 1. Reliability statistics of scale.

Factor	Wide-Body Aircraft		Narrow-Body Aircraft
	Cronbach’s Alpha	Number of Indicators	Cronbach’s Alpha	Number of Indicators
Physical environment	0.869	6	0.887	6
Seat	0.952	29	0.973	27
Interior layout	0.969	22	0.974	22
Crew service	0.939	8	0.924	8
Food and beverage	0.757	8	0.940	8
In-flight entertainment (IFE)	0.912	20	0.954	9
Total scale	0.980	93	0.988	80

(3)

Questionnaire

The study involved data collection from three major full-service airlines in China (excluding low-cost carriers), covering six flights on wide-body aircraft and eight flights on narrow-body aircraft, totaling 14 flights. Narrow-body aircraft flights ranged from 1 to 2 h, during which snacks, drinks, and limited IFE were provided. Wide-body aircraft flights ranged from 2 to 5 h and provided optional meals, beverages, and personal IFE. There were no significant flight delays. To ensure informed consent, the chief flight attendant announced to passengers, within 10–20 min after meals and beverages were distributed, that a questionnaire would be handed out. Crew members assisted in distributing questionnaires and pens to each adult passenger, excluding those who did not wish to participate or were asleep. Ethical approval was obtained from the University Ethics Committee, and all procedures followed relevant regulations.

To minimize respondent fatigue and ensure data quality, strict screening rules were applied. Questionnaires with “straight-lining” (choosing the same answer for all items), significant missing data (more than 10% of items unanswered), or inconsistent responses were excluded from analysis [35]. For wide-body aircraft surveys, 749 satisfaction questionnaires were distributed, with 444 returned. For narrow-body aircraft, 591 satisfaction questionnaires were distributed, with 396 returned. For importance questionnaires, a total of 799 were distributed, with 530 returned.

2.2.2. Hierarchical Structure Model of Factors Influencing Overall Comfort in Civil Aircraft Cabins

(1)

Construction of measurement model

Constructing a model for overall comfort based on disassembled indicators would result in several variables and a complex structure. Some first-order measurement model factors (such as food and beverage, crew service, and IFE) are highly correlated. To simplify the model and ensure theoretical consistency, a second-order measurement model was proposed, incorporating indicator-factor-latent factors. Latent factors included internal environment, facilities, and service experience (Figure 2). In this model, indicators are the directly measured variables, while factors and latent variables are inferred and cannot be directly observed through the questionnaire. A list of cabin indicators, factors, and latent factors are shown in Appendix A.

Figure 2. Second-order measurement model of the overall comfort of civilian aircraft cabins.

(2)

Transformation of latent variables into measurement variables

To simplify the structure, second-order measurement indicators were weighted and aggregated to form composite scores for each factor. Passengers perceive overall comfort as a holistic psychological state, with varying contributions from different cabin attributes. According to Expectancy-Value Theory, the impact of each attribute on overall satisfaction is a function of both its performance (satisfaction score) and its subjective importance to the individual [7,23]. This method ensures that critical factors (e.g., legroom) influence the overall comfort score more than less important factors. Thus, we employed a weighting procedure based on importance scores. This transformation ensures that the resulting composite variables accurately reflect the passenger’s value system, thereby enhancing the ecological validity of the subsequent SEM analysis. The aggregation procedure is defined as follows:

Let $L_m$ denote the m-th latent factor which contains measurement indicators, denoted as I. The set is defined as $L_m$ = {$I_1, I_2$, …, $I_y$}. Let W denote the weight derived from the importance survey, and P denote the satisfaction score from the survey.

First, the weight of each measurement indicator ($W_{I_y}$) is calculated based on its mean importance score ($Z_{I_y}$) relative to the sum of importance scores for all indicators in that factor:

$$\begin{array}{c}{W}_{I_y}={\displaystyle \frac{Z_{I_y}}{{\sum }_y{Z}_{I_y}}}\end{array}$$

(1)

Next, the composite comfort score ($P_{L_m}$) for the latent factor $L_m$ for each respondent is calculated by summing the weighted satisfaction scores of the indicators:

$$\begin{array}{c}{P}_{L_m}={\sum }_y\left(P_{I_y}·W_{I_y}\right)\end{array}$$

(2)

Following this procedure, the internal environment and other specific sub-factors were transformed into observed composite variables. Consequently, the second-order measurement model was simplified into a first-order structure comprising factors and latent constructs.

After the factors in the second-order measurement model were transformed from latent variables into measurement variables, the second-order measurement model was transformed into the first-order measurement model factor-latent factor, whereas the internal environment of the latent variables only included one measurement variable: physical environment. The latent variable facility includes two measurement variables: seat and interior layout. The latent variable service experience includes three measurement variables: crew service, foods and beverages, and IFE (Figure 3). The measurement model in Figure 3 includes basic characteristics of overall comfort. The hierarchical structure model of factors influencing overall comfort in civil aircraft is shown in Figure 4.

Figure 3. Variables of the overall comfort model.

Figure 4. Hierarchical structure model of factors influencing overall comfort in civil aircraft.

2.3. Construction of Overall Comfort Model for Civilian Aircraft Cabin

2.3.1. Initial Model Construction

SEM is a statistical method used to analyze the relationships between variables based on their covariance matrix [36,37]. It allows for the examination of complex relationships, incorporating both measurement and structural models. The measurement model uses factor analysis to explore latent variables among the observed indicators, while the structural model uses path analysis (including multiple regression) to examine causal relationships between variables [38,39].

In this study, the overall comfort model for civilian aircraft cabins was developed by combining the weighted multi-level measurement model and path analysis methods. The primary objective was to construct relationships between the internal environment, facilities, and service experience, and their impact on overall comfort. The initial hypothesized model was proposed based on these relationships (Figure 5). This model, which integrates both latent variables and observed variables, serves as the foundation for understanding how various factors contribute to overall comfort.

Figure 5. Initial hypothetical model of the overall comfort of civilian aircraft cabins.

In this model, the internal environment is treated as an observed composite variable derived from weighted indicators (as described in Section 2.2.2). Facilities and service experience are modeled as latent constructs, measured by their respective sub-factors (Such as facilities are measured by seat and interior layout). Overall comfort is treated as a single-item observed variable directly obtained from the overall satisfaction question in the survey. This hybrid approach (Figure 6) allows for a comprehensive yet parsimonious estimation of the structural relationships among comfort factors.

Figure 6. Hybrid path analysis model for the overall comfort of civilian aircraft cabins.

2.3.2. Model Test

Once the measurement model was validated, the structural relationships within the overall comfort model were tested. The model fit was evaluated using several fit indices, including Chi-square (χ²), the ratio of Chi-square to degrees of freedom (χ²/df), Goodness-of-Fit Index (GFI), Comparative Fit Index (CFI), and Root Mean Square Error of Approximation (RMSEA). The following criteria were used to assess the model fit: χ² should not be statistically significant (p > 0.05), χ²/df should be less than 2, GFI and CFI should be greater than 0.9, and RMSEA should be less than 0.08 [40,41].

Following the measurement model validation, the structural relationships within the overall comfort model were tested. Model fit was evaluated using multiple indices: Chi-square (χ²), the ratio of Chi-square to degrees of freedom (χ²/df), Goodness-of-Fit Index (GFI), Comparative Fit Index (CFI), and Root Mean Square Error of Approximation (RMSEA). The criteria for an acceptable fit are listed in Table 2. The structural model analysis focused on decomposing effects into direct, indirect, and total effects to elucidate the mechanisms influencing overall comfort.

Table 2. Fitting indices of the model.

Index	Absolute Fit Measurement					Incremental Fit Measurement				Parsimonious Fit Measurement
	χ2	χ2/df	GFI	AGFI	RMSEA	NFI	RFI	TLI	CFI	CN
Standard	p > 0.05	<2	>0.9	>0.9	<0.08	>0.9	>0.9	>0.9	>0.9	>200

2.4. Multigroup Analysis of Overall Comfort Model

The overall fit of the model reflects the survey data, but the imbalance in sample sizes and survey items presents challenges in assessing the generalizability of the model. Therefore, it was necessary to determine whether the model applies to various subgroups, such as different aircraft types (wide-body and narrow-body) and flight distances (short, medium, and long haul). In the present study, multigroup analysis was conducted to evaluate the validity and stability of the overall comfort model across these groups. Multigroup analysis allows for evaluating whether a model suitable for one group is also applicable to other groups. By comparing groups based on aircraft types and flight distances, this analysis helps to assess whether the model can be generalized across different conditions [42,43].

2.4.1. Model Cross-Group Validity Test

The path map of the overall comfort model was created using AMOS Graphics™ 22, with group variables added to form a multigroup model with unrestricted parameters. Several restricted models were tested, including models with equal measurement weights, structural weights, structural variables, and structural residuals. These parameter restrictions helped assess the validity of the model across different groups. The maximum likelihood method was used to fit the multigroup models, and model fit was assessed using the same indices as for the initial model. If the fit indices meet the recommended criteria (Table 3), the hypothesis of cross-group invariance is supported, indicating that the model structure is valid across different subgroups.

Table 3. Overall fit index and standards of the multi-group model.

Index	χ2	χ2/df	NFI	RFI	IFI	TLI	GFI	AGFI	RMSEA	AIC	ECVI
Standard	p > 0.05	<2	>0.9	>0.9	>0.9	>0.9	>0.9	>0.9	<0.08	smaller is better	smaller is better

Note: Considering that the significance of χ² is affected by the sample size, a loose standard may not consider the significance of χ².

2.4.2. Model Cross-Group Stability Test

After confirming the validity of the parameter-unrestricted model, the stability of the model was tested by comparing it with restricted models (e.g., models with equal measurement weights, structural weights, and residuals). The significance of Chi-square differences (Δχ²) between the baseline and restricted models, along with the absolute difference in incremental fit indices, was analyzed to test model stability. If no significant differences were found, the model path coefficients across groups were considered stable. The test indicators and standards are shown in Table 4.

Table 4. Stability evaluation index of the multi-group model.

Index	Δχ2	ΔNFI	ΔRFI	ΔIFI	ΔTLI
Standard	p > 0.05	<0.05	<0.05	<0.05	<0.05

Note: Considering that the significance of Δχ² is affected by the sample size, a loose standard may not consider the significance of Δχ².

3. Results

3.1. Overall Comfort Model of Civilian Aircraft Cabin

3.1.1. Reliability and Validity Analysis of Second-Order Measurement Model

The reliability and validity of the second-order measurement model were evaluated using the reliability index (Composite Reliability, CR) and the validity index (Average Variance Extracted, AVE). The CR and AVE values for the latent variables and their respective measurement variables met the standard requirements (Table 5), indicating that the second-order measurement model is reliable and valid.

Table 5. Reliability and validity of the second-order measurement model.

Latent Variables	Measurement Variables	λ	Θ	CR		AVE
				Measurement Variables	Latent Variables	Measurement Variables	Latent Variables
Internal environment	Physical environment	1	0	0.637	1	0.570	1
Service experience	Crew service	0.690	0.524	0.942	0.759	0.672	0.513
	Food and beverage	0.788	0.379	0.938		0.682
	IFE	0.665	0.557	0.972		0.636
Facilities	Seat	0.785	0.385	0.971	0.777	0.544	0.647
	Interior interface	0.824	0.321	0.973		0.609

Note: where λ represents the standardized factor loadings, Θ represents the error variances, CR is the Composite Reliability, and AVE is the Average Variance Extracted. The CR values exceed the threshold of 0.6, and the AVE values exceed the threshold of 0.5, confirming that the measurement model is both reliable and valid.

3.1.2. Fitting Quality Analysis of Overall Comfort Model

The model was fitted using the Maximum Likelihood (ML) estimation method. The fit indices showed that the model had a good fit with the data (Figure 7). Specifically, the Chi-square test (χ²) was not statistically significant (p = 0.491), and the ratio of Chi-square to degrees of freedom (χ²/df) was 1.23, which is below the acceptable threshold of 2. The Goodness-of-Fit Index (GFI) and Comparative Fit Index (CFI) were both greater than 0.95, while the Root Mean Square Error of Approximation (RMSEA) was less than 0.05, indicating that the fit between the model and the data was ideal.

Figure 7. Comprehensive cabin comfort model and its path coefficient.

In the model, internal environment, facilities, and service experience jointly explained 98.6% of the variation in overall comfort. Further, internal environment and service experience jointly explained 80.9% of the variation in facility.

The analysis of the model’s effects shows that Facilities have the greatest direct effect on overall comfort (0.493), while Service experience has the greatest total effect (0.837). The substantial indirect effect of service experience on overall comfort, mediated by facilities (indirect effect = 0.370), indicates the importance of service-related factors in enhancing passenger comfort Table 6). Among service experience factors, food and beverage had the most significant contribution, followed by crew service and IFE.

Table 6. Effect values of latent variables of overall comfort of the civilian aircraft cabin.

Exogenous Variable	Factor	Endogenous Variable	Intervening Variable	Indirect Effects	Direct Effects	Total Effects
						Factor	Exogenous Variable
Internal environment	Physical environment	Overall comfort	Facilities	0.108	0.092	0.2	0.2
Service experience	Service	Overall comfort	Facilities	0.370	0.467	0.6	0.837
	Food & beverage					0.65
	IFE					0.585
Facilities	Seat	Overall comfort	--	--	0.493	0.39	0.493
	Cabin interface					0.403

3.2. Test of Overall-Comfort-Model Validity and Stability

3.2.1. Cross-Aircraft Type Validity and Stability Analyses of Overall Comfort Model

Aircraft type (wide-body aircraft, narrow-body aircraft) was used as the group variable. The overall fitness index of the multigroup model is shown in Table 7.

Table 7. Summary of overall fit index of the cross-aircraft validity test of the comprehensive cabin comfort model.

Model	χ2	DF	p	χ2/df	NFI	RFI	IFI	TLI	GFI	AGFI	RMSEA	AIC	ECVI
Unrestricted	17.58	18	0.484	0.977	0.995	0.989	1	1	0.994	0.981	0	93.58	0.114
Measurement weights	26.449	23	0.28	1.15	0.993	0.988	0.999	0.998	0.991	0.978	0.014	92.449	0.113
Structural weights	31.253	26	0.219	1.202	0.992	0.987	0.999	0.998	0.989	0.977	0.016	91.253	0.111
Structural covariances	34.218	28	0.194	1.222	0.991	0.987	0.998	0.998	0.988	0.976	0.016	90.218	0.11
Structural residuals	37.231	29	0.14	1.284	0.99	0.986	0.998	0.997	0.987	0.975	0.019	91.231	0.111
Measurement residuals	103.021	36	0	2.862	0.973	0.969	0.983	0.98	0.968	0.95	0.048	143.021	0.174

Note: when NFI, RFI, IFI, TLI, GFI, AGFI > 0.95, RMSEA < 0.05 indicates that fitness between model and data is very ideal.

The fitting index statistics of the parameter-unrestricted model showed that χ² was not statistically significant (p = 0.484), GFI, AGFI, NFI, RFI, IFI, TLI and CFI were greater than 0.95, and RMSEA was less than 0.05, suggesting that the fitting quality of the aircraft-type-unrestricted model is ideal. Specifically, the parameter-unrestricted model has cross-aircraft validity. The fitting index statistics of four parameter-restricted models—the measurement weights model, structural weights model, structural covariances model, and structural residuals model—all meet the model adaptation standards, suggesting that four parameter-restricted models have a good fit with the sample data. However, the fitting index statistics of the measurement residuals model showed that χ² was statistically significant, and χ²/df was greater than 2, which does not meet the adaptation standard.

The unrestricted model was used as the baseline model. The stability of this model was tested by comparing it with the four parameter-restricted models that met the adaptation criteria. The significance of Δχ² in all four parameter-restricted models was not significant. The absolute value of the difference in the incremental fit indices (ΔNFI, ΔRFI, ΔIFI, ΔTLI) was less than 0.05 (Table 8). Results suggest that there was no significant difference between the four parameter-restricted models and the parameter-unrestricted model. Therefore, the overall comfort model is the same for wide-body and narrow-body aircraft.

Table 8. Summary of differences between parameter restricted and unrestricted models.

Model	Δχ2	ΔDF	p	ΔNFI	ΔRFI	ΔIFI	ΔTLI
Measurement weights	8.869	5	0.114	0.002	0.001	0.001	0.002
Structural weights	13.673	8	0.091	0.003	0.002	0.001	0.002
Structural covariances	16.638	10	0.083	0.004	0.002	0.002	0.002
Structural residuals	19.651	11	0.0503	0.005	0.003	0.002	0.003

3.2.2. Cross-Distance Validity and Stability Analyses of Overall Comfort Model

The flight distance (short distance, medium distance, and long distance) was used as a group variable to determine whether the overall comfort model differed by flight distance. The overall fitness index between the multigroup model and the sample data is shown in Table 9.

Table 9. Summary of overall fit index of cross-distance validity test of the comprehensive cabin comfort model.

Model	χ2	DF	p	χ2/df	NFI	RFI	IFI	TLI	GFI	AGFI	RMSEA	AIC	ECVI
Unrestricted	35.29	25	0.083	1.412	0.991	0.976	0.997	0.993	0.988	0.960	0.022	153.297	0.187
Measurement weights	60.989	35	0.004	1.743	0.984	0.971	0.993	0.987	0.980	0.952	0.030	158.989	0.194
Structural weights	80.116	41	0.000	1.954	0.979	0.967	0.989	0.984	0.974	0.946	0.034	166.116	0.203
Structural covariances	86.993	47	0.000	1.851	0.977	0.969	0.989	0.986	0.972	0.950	0.032	160.993	0.197
Structural residuals	92.551	49	0.000	1.889	0.975	0.968	0.988	0.985	0.971	0.950	0.033	162.551	0.198
Measurement residuals	133.905	62	0.000	2.130	0.964	0.964	0.981	0.980	0.962	0.949	0.038	177.905	0.217

The fitting index statistics of the parameter-unrestricted model showed that χ² was not statistically significant (p = 0.083), GFI, AGFI, NFI, RFI, IFI, TLI and CFI were greater than 0.95, and RMSEA was less than 0.05, indicating that the fitting quality of the unrestricted model to the sample data is very ideal. Specifically, the parameter-unrestricted model has cross-distance validity. However, among the five parameter-restricted models, χ² was significant (p < 0.05). Because the significance of χ² is affected by the sample size, further examination is conducted to determine if other fitting indices meet the criteria. Among the parameter-restricted models, the measurement residuals model (χ²/df > 2, AGFI < 0.95) and the structural weights model (AGFI < 0.95) fail to meet the fitting criteria. Consequently, these two models are excluded from consideration. The fitting indices of the other three models all satisfy the fitting criteria.

The parameter-unrestricted model was used as the baseline model. The stability of this model was tested by comparing it with the three parameter-restricted models that met the adaptation criteria. In all three parameter-restricted models, Δχ² was significant (p < 0.05). However, as the significance of Δχ² is affected by the sample size, like the significance of χ², the absolute value of the difference in the value of the incremental fitting index between parameter-restricted model and parameter-unrestricted model, which is not affected by the model complexity, is considered to be the main assessment criterion. The results show that ΔNFI, ΔRFI, ΔIFI, and ΔTLI of measurement weights, structural covariances, and structural residuals are all less than 0.05 (Table 10). That is, these three parameters limit the model to be no different from the baseline model. Therefore, the model is stable. The AIC and ECVI values of the measurement weights model are the smallest, so it is identified as the best model.

Table 10. Summary of differences between parameter restricted and unrestricted models.

Model	Δχ2	ΔDF	p	ΔNFI	ΔRFI	ΔIFI	ΔTLI
Measurement weights	25.692	10	0.004	0.007	0.007	0.006	0.006
Structural covariances	51.696	22	0.000	0.014	0.014	0.007	0.007
Structural residuals	57.255	24	0.000	0.015	0.015	0.008	0.008

In the cross-distance group analysis, because χ² and Δχ² are significant, a loose cross-group stability test standard is adopted. In order to verify whether the models of the three groups of short distance, medium distance, and long distance are stable, we analyzed the short-distance, medium-distance, and long-distance grouping models in the measurement weights model, as shown in Figure 8. In addition, based on the direct effects, indirect effects, and total effects, the effects of factors influencing the overall comfort for different flight distances are compared (Table 11).

Figure 8. Model of overall comfort across different flight distances.

Table 11. Effect values of latent variables of overall comfort of the civilian aircraft cabin.

Flight Distance	Exogenous Variable	Endogenous Variable	Intervening Variable	Indirect Effects	Direct Effects	Total Effects
Short distance	Physical environment	Overall comfort	Facility	0.311	0.04	0.315
	Service experience	Overall comfort	Facility	0.372	0.38	0.75
	Facility	Overall comfort	--	--	0.51	0.51
Medium distance	Physical environment	Overall comfort	Facility	0.115	0.08	0.123
	Service experience	Overall comfort	Facility	0.37	0.48	0.85
	Facility	Overall comfort	--	--	0.50	0.5
Long distance	Physical environment	Overall comfort	Facility	1.225	0.14	0.263
	Service experience	Overall comfort	Facility	0.323	0.45	0.773
	Facility	Overall comfort	--	--	0.49	0.49

The results indicate that in the short-, medium-, and long-haul models, although there are differences in the direct and total effect values of cabin factors on the overall comfort, the degree of influence of cabin factors on the overall comfort, i.e., the relative importance of different cabin factors, remains unchanged. Therefore, the overall comfort model of aircraft cabins exhibits validity and stability across flight distances.

4. Discussion

4.1. Contribution and Significance of Research

This study contributes significantly to the literature on public transport comfort by making three key contributions. First, it introduces a weighted hierarchical measurement structure that simplifies complex comfort indicators into manageable constructs, offering a clearer framework for understanding the factors that influence overall comfort. Second, by employing a hybrid path analysis model, the study quantifies not only the direct impacts but also the indirect pathways through which comfort factors influence overall comfort. The results reveal that service experience significantly enhances overall comfort by positively influencing passengers’ perceptions of cabin facilities. Third, the study provides a robust cross-group validation, confirming the model’s stability across different aircraft types (wide-body and narrow-body) and flight distances (short, medium, and long haul). This methodological framework offers a replicable tool for identifying key optimization points in both aircraft cabin design and service management.

Furthermore, the proposed methodology is equally applicable to studies on comfort in general public transportation, including at the pre-travel, travel, and post-travel stages. It helps identify the key factors for optimizing travel comfort across various transport modes. Notably, in addition to assessing the validity and stability of the model, the cross-group analysis helps identify variations in the model’s structure and path coefficients across different subgroups, thereby enabling the development of differentiated strategies to enhance comfort for distinct passenger groups.

The method employed in this study was used to develop a comfort model for civilian aircraft cabins and to identify the critical factors influencing overall comfort. The model’s applicability can be evaluated from at least four perspectives: foundational theory, mathematical modeling techniques, empirical model characteristics, and parameter values [29,44]. The results of this study elucidate the basic features of comfort in civilian aircraft cabins by identifying the hierarchical relationships between factors and indicators that influence comfort. The study also explores the mathematical framework used to model comfort, employing a combined weighted multi-level measurement model and a hybrid path analysis model. Moreover, based on survey data, the relationships between comfort factors and their pathways were explored and verified, quantifying the effect of these factors on comfort by combining both direct and indirect effects. Ultimately, this study highlights the operational mechanisms influencing overall comfort and identifies the key factors for optimizing the passenger comfort experience.

4.2. Relationships Among Factors

The findings of this study indicate that facilities have the greatest direct impact on passenger comfort in civilian aircraft cabins, followed by service experience, with internal environmental factors having the least direct influence. However, when both direct and indirect effects are considered, service experience emerges as the most significant factor influencing passenger comfort. Specifically, service experience—which encompasses elements such as food and beverage quality, crew service, and in-flight entertainment (IFE)—has a substantial impact, surpassing the physical environment and facilities in terms of overall comfort [45]. Cabin facilities including seats and interior layout are factors with which passengers interact most frequently. This is consistent with earlier studies that emphasized the direct influence of physical factors on comfort [21,46]. However, it is important to note that internal environmental factors (such as temperature, noise, and air quality), which reflect fundamental comfort needs, are typically already standardized in air travel [11]. The transition in modern transportation, from space-oriented travel to an experience-oriented one, has led to a greater emphasis on service experience in both airline operations and passenger expectations [3].

The analysis also reveals an important mechanism: service experience and internal environment together explain 80.9% of the variance in perceived facilities. The high explanatory power of this model can be partly attributed to the weighting procedure that integrates importance scores. By incorporating these importance scores, the analysis moves beyond raw satisfaction ratings to focus on importance-weighted perceptions, amplifying the effect of high-priority factors (like service experience) and reducing the influence of less critical attributes. This approach helps provide a more accurate representation of passengers’ psychological value system compared to unweighted models. This suggests a psychological “halo effect”, where high-quality service can enhance passengers’ subjective evaluation of the physical environment. For instance, excellent catering (food and beverages) [18] and engaging IFE [47,48] can distract passengers from physical discomfort, such as from seats or cabin layout. The substantial path coefficient from service experience to facilities (0.751) quantitatively supports this interaction. Therefore, when physical optimization of facilities (such as seat pitch) is limited by economic factors, airlines can still effectively enhance overall comfort by improving service aspects—such as food quality and crew interactions—that compensate for physical constraints.

4.3. Main Influencing Factors

This study affirms that service experience has become the most important factor influencing overall comfort in civilian aircraft cabins. This result confirms the growing recognition that service-related factors now play a central role in shaping passengers’ perceptions of comfort, surpassing the impact of physical environmental factors [49]. Historically, research on cabin comfort focused on physical factors like vibration, air quality, and seating arrangement [11]. However, the anthropometry and human–machine adaptability of the cabin are increasingly being studied in current investigations. The evolution of comfort research can be conceptualized using the discomfort pyramid theory, which posits that satisfying low-level comfort needs leads to the development of higher-level comfort needs [11]. Vink and Brauer [12] also demonstrated that, in modern business competition, service factors might become the most important element.

Among the service experience factors, food and beverage had the most substantial impact on comfort, followed by crew service and IFE. The distinction between comfort and discomfort emphasizes that comfort is an additional experience—one that can be enhanced by factors like the quality of food and beverages [14]. Food and beverage services can also serve as a means of compensating for discomfort caused by other factors, such as seat discomfort. Furthermore, dining activities help alleviate the discomfort associated with prolonged sitting. Cabin crew service fulfills higher-level needs in Maslow’s hierarchy, and it plays a critical role in improving satisfaction at low cost. As competition in the airline industry intensifies, excellent crew service remains a key factor for improving passenger satisfaction [3].

With the increasing use of mobile devices, traditional IFE systems, primarily focused on audiovisual resources, have lost some of their competitive edge. However, personal IFE systems still play a crucial role in diverting passengers’ attention and alleviating boredom, especially in flights that do not offer Wi-Fi. The continuous enrichment and updating of entertainment sources are vital to keeping this service effective and engaging for passengers [47,48].

4.4. Limitations

In the multigroup analysis of the overall comfort model, the model had cross-types and cross-distance validity and stability, so it can be considered that the relationship between the comfort factors of different subgroups is the same. In fact, the comfort models had different path coefficients; however, the differences were not significant.

This study explores the comfort in current traditional full-service flights, and analyzes the validity and stability of models across aircraft types and flight distances. However, the research sample does not include small aircraft used for low-cost flights or regional flights, nor does it cover flights with significant time delays. In addition, when there are upgrades or major changes to cabin facilities and services, it is necessary to use the aforementioned method to reassess the validity and stability of the model.

5. Conclusions

Comfort is a key factor in passengers’ acceptance of transportation systems. Cabin comfort has become a comprehensive experience that includes both the physiological and psychological aspects of the cabin’s physical and social environment. A thorough analysis of cabin comfort is crucial for understanding the influence of various comfort factors. This study presents a method for researching comfort based on survey-scale data analysis. The factors influencing comfort in aircraft cabins, along with their pathways and degrees of impact, were examined. The results indicate that service experience, particularly food and beverage, IFE, and crew service, has the greatest impact on cabin comfort, surpassing the influence of facilities and the physical environment. Additionally, service experience plays a substantial role in shaping overall comfort. Furthermore, the results demonstrate cross-group validity and stability for different aircraft types and flight distances. This study explains the overall comfort in civilian aircraft cabins and quantitatively models the relationships among the factors influencing comfort, providing key insights for cabin comfort design and optimization. The methodology presented also offers a valuable approach for studying comfort in public transport more broadly.