Factor loading analysis - Descriptive Statistics

CHAPTER 4 RESULTS

4.3 Descriptive Statistics

4.3.2 Factor loading analysis

Factor validity is construct validity technique used in assessing the quality of questionnaire and it is obtained by means of factors analysis. The main measures used to test the validity of an instrument in factor analysis include:

Extraction communalities are estimates of the variance in each variable accounted by the components. The communalities values indicate that the extracted components represent the variables well. Thus, small values indicate variables that do not fit well with the factor solution, and should possibly be dropped from the analysis.

aiser-Meyer-Olkin Measure of Sampling Adequacy is a statistic that indicates the proportion of variance in the variables that might be caused by underlying factors. High values (close to 1.0) generally indicate that a factor analysis may be useful with the data.

If the value is less than 0.50, the results of the factor analysis probably won't be very useful.

matrix, which would indicate that the variables are unrelated and therefore unsuitable for structure detection. Small values (less than 0.05) of the significance level indicate that a factor analysis may be useful with the data.

Base on the original dimension of service quality, game quality, customer satisfaction and behavior intensions, the current study made on effort to examine underlying of four dimension. In this section we discuss the results of the exploratory factor analysis to assess the suitability of the instruments.

According the results analyze of factor loadings, the table 7 indicated the service quality dimension rotated component matrix, it was recognized that were two factor component to be correlated with all variable. The KMO for the service quality dimension of this study was .800. Bartlett's Test of Sphericity test ( The factor loadings of service employees factor (3 items) was ranking from (.864 - .841) with Cronbach’s Alpha of .853. For the experiences factor (3 items) the factor loadings was higher than .7, ranking from (.701 – .859) and the Cronbach’s Alpha at .733. In this dimension, nine items were deleted by the value is less than 0.50, the results of the factor analysis probably won't be very useful.

Table 7

Factor Analysis of the Service Quality Dimension

Factors Items Cronbach’s

Alpha The attitude of the employees at this

arena shows you that the understand

.841 You can rely on the arena employees

taking actions to address your needs

.843 The employees at this arena respond

quickly to your needs

.864 Factor 2

Experiences

The arena’s scoreboards are entertaining to watch

.733 .758

The arena’s provides good sightlines to watch the game

.701

The smell of the crowd is exciting .859

Kaiser-Meyer-Olkin Measure of Sampling Adequacy .800

Bartlett's Test of Sphericity

Approx. Chi-Square 1253.806

df 15

Sig. .000

% of Total Cumulative 71.71

From the table 8 indicated the factor analysis of the game quality dimension. The results showed factor loadings of eleven items were higher than .5 ranking from (.552 – .892), the factor loading of three items were less than .5. The KMO for the game quality dimension was .882. Bartlett's Test of Sphericity test ( After rotated component matrix, that was rotated 3 factor including opponent characteristics, player performance and game atmosphere. In the opponent characteristics factor (3 items) the results of factor loadings ranking from (.552-.892) with Cronbach’s alpha was .786. For the player performance factor (4 items) with the factor loading ranking from (.713-.850) and Cronbach’s alpha was .889. In the third factor of game quality dimension, game atmosphere (4 items) with the factor loadings ranking from (.746-.840), the Cronbach’s alpha also larger than .7 that was .884 of game atmosphere.

Table 8

Factor Analysis of the Game Quality Dimension

Factors Items Cronbach’s

Alpha

Opposing teams are high quality teams .552

Opposing teams have star players .892

Opposing teams have good win/loss records

Factors Items Cronbach’s

Alpha

Factor loadings

Your team gives 100% every game .850

Your team plays hard all the time .871

Players on your team always try to do their best At this stadium, you can rely on there

being a good atmosphere.

.811 This stadium’s ambiance is what you

want at a game.

.840 The (team name) understand that

atmosphere is important to you.

.820 You enjoy the excitement surrounding the

performance of the players

.746 Kaiser-Meyer-Olkin Measure of Sampling Adequacy

.882

% of Total Cumulative 74.84

The results showed in table 9 & table 10 present the detail the factor loadings of customer satisfaction dimension and behavior intension dimension. In the customer satisfactions dimension the results factor loading of all items higher than .8, it ranges from .918-.903. “I am very satisfied” (.895), “I am satisfied with my decision” (.903), “I am satisfied with the outcome” (.849), “I truly enjoy to going this event” (.818) and the Cronbach’s alpha was .889, “I would go on in the future” (.920). The KMO for the customer satisfaction dimension was .813. Bartlett's Test of Sphericity test (

For the behavior intension dimension, Cronbach’s alpha was highest with .932 and the factor loadings ranking from .904-.920. “I intend to visit this sporting event in the future” (.904), “I would say positive things” (.916), “I would go on in the future” (.920),

“I would recommend to other people” (.908). The KMO for the behavior intensions dimension was .855. Bartlett's Test of Sphericity test (

Table 9

Factor Analysis of Customer Satisfaction Dimension

Factors Items Cronbach’s

Alpha Based on all my experience in this arena,

I am very satisfied

.895 I am satisfied with my decision to attend

this games

.903 I am satisfied with the outcome of this

game

.849 I truly enjoy going to this sporting event .818

Kaiser-Meyer-Olkin Measure of Sampling Adequacy .813

Bartlett's Test of

Factor Analysis of Behavior Intensions Dimension

Factors Items Cronbach’s

Alpha

Factor loadings Behavior

Intensions

I intend to visit this sporting event in the future

.932

.904 I would say positive things about going

to this sporting event to others .916

I would go on sporting events in future .920 I would recommend going to this

sporting event to other people .908

Kaiser-Meyer-Olkin Measure of Sampling Adequacy .855

Bartlett's Test of Sphericity

Approx. Chi-Square 1767.142

df 6

Sig. .000

% of Total Cumulative 83.16

38 4.3.3. Correlation among Variables

A bivariate correlation was run between all continuous variables. The correlation coefficients (r) ranged from .683 to .833. No correlations were higher than .90 or lower than .10. The majority had a correlation greater than .6, a medium strength of relationship.

The correlation among the variable in the contrast are presented in table 11, all the variables were significant correlated and the listed as following. Service quality was positively correlated with game quality (r = .722, p < .01), customer satisfaction (r = .692, p < .01), behavior intensions (r = .683, p < .01). Game quality variables were positively correlated with customer satisfactions (r = .711, p < .01), behavior intension (r = .692, p

< .01). Finally, customer satisfactions was also positively correlated with behavior intensions (r = .833, p < .01).

Table 11

Correlation among the variable in the contrast

Measure Correlation Matrix

1 2 3 4

1. Service quality 1.00

2. Game quality .722** 1.00

3. Customer satisfactions .692** .711** 1.00

4. Behavior intensions .683** .692** .833** 1.00

**. Correlation is significant at the 0.01 level

4.3.4 CFA Model Parameter

The CFA measurement model was identified on the 7 factors. Table 12 presented the standardized CFA model results. All items loaded significantly (p < .001) on the 7 theorized latent variables. The values of the factor loadings ranged from .550 to .910.

Most values of the factor loadings were greater than .70; while only 3 values were smaller than .70. The results provided evidence that the indicators are good measures of their

conceptual constructs. In addition, each item only loaded on one factor, which suggests convergent and discriminant validity. Figure 4 depicts the CFA diagrams for the 7 factors, with factor loadings and error terms.

Table 12

Standardized CFA measurement model results

Constructs Indicators β p-value

Service employees

Figure 5. CFA diagrams for the 7 factors with factor loadings

41 4.3.5. Confirmatory Factor Analysis

A confirmatory factor analysis was conducted using AMOS for measurement model of determinants of customer satisfactions and behavior intension. Confirmatory factor analysis (CFA) serves as a measurement model for a structural equation model.

Anderson and Gerbing (1988) argue that, in customer behavior research it is common to analyze the measurement model before the structural models. CFA allows the researcher to assess the contribution of each observed variable and determine how well the observed variable measures its underlying latent construct. The purpose of CFA is to specify the relationship between each scale item and its underlying latent constructs (factors).

After the data were screened, the CFA was run for the measurement model in Figure 5 including the seven latent constructs: service employees, access, opponent characteristics, player performance, game atmosphere, customer satisfaction and behavior intension. The results are presented and discussed for the model fit indices, factor loadings (the correlation between the latent variable and the observer variable). The results of the CFA measurement model indicated a good fit of the data to the hypothesized structure. The value of CFI was .952, the value of GFI was .902, and the value of RMSEA was .60. All model indices exceeded the suggested criteria indicating a good fit. Table 14 provided the results of fit indices for the CFA measurement model and the recommended value of the good-of-fit indices.

Table 13

Results of model fit indices for CFA measurement model

Absolute fit indices Obtained Recommendations on fit

indices

CFI .952 > .90

GFI .902 > .90

RMSEA .060 < .08 or < .1

2.906 < 5

Absolute fit indices Obtained Recommendations on fit

indices 677.025 P < 0.05(N > 250)

d.f .000

Figure 6. Confirmatory factor analysis a measurement model

43 Table 14

Correlation matrix between the CFA measurement model

Correlations Estimates

Service employees Experiences .629

Service employees Opponent characteristics .665

Service employees Player performance .335

Service employees Game atmosphere .544

Service employees Customer satisfaction .619

Service employees Behavior intension .640

Experiences Opponent characteristics .611

Experiences Player performance .622

Experiences Game atmosphere .868

Experiences Customer satisfaction .773

Experiences Behavior intension .738

Opponent characteristics Player performance .385

Opponent characteristics Game atmosphere .539

Opponent characteristics Customer satisfaction .640

Opponent characteristics Behavior intension .684

Player performance Game atmosphere .635

Player performance Customer satisfaction .436

Player performance Behavior intension .473

Game atmosphere Customer satisfaction .795

Game atmosphere Behavior intension .750

Customer satisfaction Behavior intension .880

4.3.6 Estimated Correlations of Latent Variables for CFA

The model had 67 parameters to be estimated and 233 degrees of freedom. The correlations among 7 latent variables were estimated. Table 15 was presents the estimated correlation matrix for the 7 latent variables. The results show that all correlations among 7 latent variables were statistically significant at the .001 level and ranged from .334 to .948.

According Green & Salkind (2008) correlation coefficients of .10, .30, and .50 are usually interpreted as small, medium, and large coefficients respectively. Only 5 of 21 correlations were smaller .50 – the correlations between stadium employees and player performance (.309); between stadium employees and opponent characteristics (.476); between opponent

characteristics and player performances (.334); between opponent characteristics and game atmosphere (.389); and between player performance and customer satisfaction (.434). The rest of the correlations had high coefficient values (greater than .05). The results suggest that all latent variables were highly correlated to each other in the CFA measurement model.

Table 15

Estimated Correlation Matrix of Latent Variables for CFA

Measure Correlation Matrix

1 2 3 4 5 6 7

1. Stadium employees 1.00

2. Experiences .521** 1.00

3. Opponent characteristics .529** .476** 1.00

4. Player performance .309** .563** .334** 1.00

5. Game atmosphere .417** .653** .389** .533** 1.00

6. Customer satisfactions .564** .690** .547** .434** .658** 1.00

7. Behavioral intentions .639** .723** .642** .517** .681** .948** 1.00

4.3.7 The Second-order-factor Model

The second-order model represents the hypothesis that these seemingly distinct, but related constructs can be accounted for by one or more (Chen, 2005). CFA was used to confirm the expected relationship of service quality latent variable and game quality latent variable and their corresponding dimensions. As a two latent factor of service quality and three latent factor of game quality were significantly correlated, second order-factor-model was tested. Figure 7 & Figure 8 was showed that all the items loaded significant on their respective factors (p < .01) and the factor loading ranking from .56 to .89. The overall, the fit of the model to the data were moderate as following. For the service quality model (CFI = .981; GF I= .981; RMSEA = .075) and for the game quality model (CFI

= .964; GFI = .946; RMSEA = .076).

45 Table 16

Results of model fit indices for CFA second-order-model Service Quality

Model

Game Quality

Model Recommended Value

CFI .981 .964 > .90

GFI .981 .946 > .90

RMSEA .075 .076 < .08 or < .1

4.026 4.077 < 5

32.206 167.172 P < 0.05

d.f 8 41

Figure 7. The second-order-factor model for service quality

Figure 8. The second-order-factor model for game quality

4.3.8. Structure Equation Model

In this study formulated a SEM to analyze the proposed model using AMOS to test the relationship among the proposed model. According to many scholars, the first issue to consider in examining a structural model is to examine the goodness of fit (GOF) of the model (Hair, et al.2006; Patrick 1997). Benchmarks for recommendable values for an overall fit have been suggested in previous studies (Table 17).

The results of the standardized path coefficients (β) indicated that service quality had a significant influence on customer satisfaction (β = .43, p > 0.001) and behavior intensions (β = .22, p > 0.001); game quality had a significant on customer satisfaction (β

= .67, p > 0.001) and behavior intensions (β = .20, p > 0.001); customer satisfaction had a

direct significant influence on behavior intensions (β = .61, p > 0.001). The structural equation model was illustrated in Figure 8.

Table 17

Results of Model Fit Indices for SEM

Absolute fit indices Obtained Recommendations on fit

indices proposed service quality has a direct positive relationship with customer satisfaction and behavior intension. The results of structural equations model estimates in Figure 8 show that service quality is a significant predictor of customer satisfaction (H1: β = .43, p < .001) and behavior intensions (H2: β = .22, p < .001). Game quality has a direct significant influence on customer satisfaction (H3: β = .67, p < .001) and behavior intensions (H4: β

= .20, p < .001). Finally, in the hypothesis 5, customer satisfaction has a direct significant influence on behavioral intentions.

Table 18

Results of research hypothesis testing

Hypothesis β

Satisfaction .434 *** Direct Supported

H2: Service Quality  Behavior intension .222 *** Direct Supported H3: Game Quality  Customer Satisfaction .666 *** Direct Supported H4: Game Quality  Behavior intension .198 *** Direct Supported

Hypothesis β

p-value

Path Pattern

Test Outcome H5: Customer satisfaction  Behavior

intension .610 *** Direct Supported

Figure 9. The structural equation model

CHAPTER 5 CONCLUSION

This chapter summarizes the purpose of the study, the major findings and limitation. This chapter is divided into several sections. This information is presented in the following sections: a) overview of the study, b) implication of findings, c) limitations and future research.

5.1 Overview of the Study

The primary purpose of this study was to investigate sport spectators’ perceptions of the service evaluation variables: service quality, game quality and satisfaction in relation to their behavioral intentions while attending a college sporting event. A conceptual model with four dimensions was proposed, which were spectators’ perceptions of service quality, game quality, customer satisfaction, and behavioral intentions. The final scale generated a total of 37 items. All constructs were measured using seven-point Likert scale ranging from 1 (strongly disagree) to 7 (strongly agree).

The final scope of the study was to test the proposed conceptual model by using service quality, game quality, customer satisfaction and behavior intension questionnaire.

536 usable surveys were collected from spectators who was participated the 2014 Vietnam University Games held in Ho Chi Minh City.

The first part, the study was to understanding the satisfaction of spectators’

perceptions service quality and game quality during the game. Based on the evaluation of spectator indicated that they are most satisfied with (1) “player try to do best”, (2) “home team plays hard”, (3) “home team give 100% effort”, (4) “score board entertaining”, (5)

“home player associated with player”. And the spectator was most unsatisfied with (1)

“the food in arena”, (2) “guest team have start player”; (3) “quality of sound in arena”; (4)

“guest team have win/lose records”; (5) “attitude of employees”.

The second part, the data was analyzed using by structural equation modeling, consisting of two parts: measurement model and structural model. Confirmatory factor analysis was first assessed to determine the appropriateness of the measurement model which generated 4 factors. Finally, the structural model was assessed to identify the causal link among the 4 latent variables. The results reveal that the proposed conceptual model explained 63.1 % in customer satisfaction, 73.9% variance in behavioral intentions to the sporting event. In addition, the results indicated the model fits the data well. For the hypothesis testing, all other hypothesized paths were significantly with p-value small than .01.

5.1.1 Service Quality

Service quality refers to the element of the service quality in the sporting events content. The current study reveals that service quality had a positive influence on both customer satisfaction and behavior intension. Thus this findings support hypothesis 1 (β

= .434, p < .001) and hypothesis 2 (β = .222, p < .001). That implies which is high quality service is a key determinant of spectator’s satisfaction and behavior intension. This also consisted with the previous studies (Theodorakis et. al., 2013) the findings in this study suggest service quality lead to the customer satisfaction; customer satisfaction was a mediated factor influence on behavior intension. In this current study found service quality also has a direct effect on behavior intension; therefor, the results were suggested the path from service quality to customer satisfaction was significantly than the path from service quality to behavior intension. The above results suggested that the relationship between service quality dimension and behavior intension was partially mediated by customer satisfaction.

51 5.1.2 Game quality

Game quality in the spectator industry refers to the entertainment of the competition based on the game outcome, associated of player with excitement of the sporting event. The findings indicate that the relationship between game quality customer satisfaction and behavior intension was significantly different support hypothesis 3 (β

= .666, p < .001) and support hypothesis 4 (β = .198, p < .001). The game quality has been widely discussed by pervious researches (Yoshida & Jame, 2010; Theodorakis et. al., 2013), the results findings game quality also has a direct effect on customer satisfaction on behavior intension, and however, game quality was strongest effect among customer satisfaction and behavior intension. Thus, customer satisfaction also was a partially mediated factor to behavior intension. The results of game quality in this current study consisted with the previous research.

5.1.3 Customer Satisfaction and Behavior Intensions

A positive relationship between satisfaction and behavioral intentions has been confirmed by many researchers (Caro & Garcia, 2007; Hightower et al., 2002; Kaplanidou

& Gibson, 2010; Shonk, 2006; Shonk & Chelladurai, 2008; Yoshida &James, 2010;

Theodorakis et. al., 2013). The path between customer satisfaction and sports spectator behavioral intentions toward the sporting event was statistically significant. The findings support hypothesis 5 (β = .610, p < .01), indicating that when satisfied with the service and game will be more likely to recommend the sporting event to others and revisit the sporting event.

52 5.2 Implication of Findings

The study aimed to investigated customer satisfaction and behavior intension in the context of college games. Both dimensions of service quality and game quality were included in the model and were tested in the relationship to customer satisfaction and behavior intensions. Studies have provided on the measurement of the previous research and its influence on spectators’ satisfaction and behavior intensions (Yoshida &James, 2010; Theodorakis et. al., 2013).

The first theoretical contribution of our study was proposed two clear dimension service quality (service employees and experiences environment). These two dimensions were incorporated within an integrated model of service quality, as the experiences environment (sensorycape), which was proposed by Lee, Lee, Seo, and Green (2012) in the general service marketing literature. The both of service employees and experiences environment dimensions were found to be reliable, valid and applicable in this context. As previously research discussed, in this present study the service quality covered the sub-dimensions of employees and environment experiences. There are typical sub-dimensions that had been used in the previously research in a spectator sporting event and describe the process part of service quality (Lee, Lee, Seo, and Green, 2012; Yoshida & James, 2010).

A second theoretical contribution of the study was to get clarification the relationship between the three dimensions of game quality (opponent characteristics, player performance and game atmosphere). The results indicated that the game atmosphere was strongest significant of the game quality in the content of college games.

In the previously research such as Yoshida & James (2010), found the game atmosphere was significantly influence on game satisfaction with one sample in Japan and the other in United States. Thus, in the game quality dimension, the game atmosphere is an important factor in the context of college games; therefore, it should not be overlooked in the future studies and organizations.

A third theoretical contribution of our study is to support all theoretical models proposed before; service quality and game quality has direct influence on customer satisfactions and behavior intensions. The results clarified that game quality has a stronger influence on customer satisfaction than service quality; furthermore, customer satisfaction partially mediates the relationship among service quality, game quality and behavior intension. As a previously research discussed, the results had been reported so far for above relationship. Regarding our hypothesis, we found that customer satisfaction was a

在文檔中 A Confirmative Study of Customer Satisfaction and Behavior Intensions on Sporting Events (頁 42-0)