| This study attempts to construct a theoretical model that can be used to prove the validity of the constructs of service quality, attitudinal loyalty, and behavioral loyalty, as well as to examine the causal relationships among them. To achieve these purposes, the researcher used structural equation modeling (SEM) techniques. According to Bollen (1989), Hair et al. (1998), and Byrne's (1998) viewpoints, this is the most powerful method for evaluating a measurement for its reliability and validity, as well as the structural relationship for a theoretical model in current statistical techniques.;For research purposes, there were two structural hypotheses for the proposed model. They were: Hypothesis I: Service quality influences attitudinal loyalty and behavioral loyalty. Hypothesis II: Attitudinal loyalty influences behavioral loyalty.;Three instruments were used in this study to gather information on consumers' perceptions of service quality, attitudinal loyalty, and behavioral loyalty in Taiwanese recreational sport/fitness clubs. They are as follows: (a) the Service Quality Survey of Participant Sport/Fitness Clubs (SQSPSFC), a modified version of Ko's survey of service quality, measured the perception of service quality of participants, (b) attitudinal loyalty (or commitment) was measured by the Attitudinal Loyalty Survey of Participant Sport/Fitness Clubs (ALSPSFC), which is a modified version of Allen and Meyer's (1990) twenty-four-item organizational commitment scales, and (c) behavioral loyalty was measured by the Behavioral Loyalty Survey of Participant Sport/Fitness Clubs (BLSPSFC) which combines participants' frequency, intensity, duration, and continuation of participation within an organization.;Six hundred questionnaires were distributed to twenty clubs that were randomly selected from a pool of one hundred twenty-nine recreational sport/fitness clubs. Five hundred and eight questionnaires were collected. SEM requires that the data set have no missing values. Therefore, the questionnaires that had any missing information were dropped. Finally, four hundred and sixty-four questionnaires that did not have any missing values made up the sample of this study.;The results showed that the three measurement models for service quality, attitudinal loyalty, and behavioral loyalty were valid and reliable, indicating that they are valid constructs. In testing structural relationships, the results showed that service quality had a significant impact on both attitudinal loyalty and behavioral loyalty. This means that Hypothesis I was supported. Unfortunately, the path coefficient of attitudinal loyalty on behavioral loyalty was not statistically significant. Hypothesis II did not hold. Further examination showed that there existed a spurious relationship. This examination founded that service quality cancelled out the effects of attitudinal loyalty on behavioral loyalty. |
