Microsoft Word 2007 ichrie conference Proceedings Final-Final 06-06-07. doc

METHODOLOGY 
Sampling Frame and Data Collection 
Data were collected from hotel guests in Seoul, Korea. We chose five-star hotels as a sample, because 
successful service recovery from failure is a high priority subject for luxury hotel managers to maintain a long-term 
guest relationship. The researchers contacted the director of marketing for all 11 luxury hotels and four hotels 
allowed a research team to conduct surveys to their guests. A screening question was asked to see if respondents 
encountered any service failure in the past six months during their stay at a five-star hotel. A group of six hotel 
managers and four research faculty participated in pilot testing to refine the instrument. The survey provides 
respondents no monetary incentive for completing and returning the questionnaire, and a total of 542 responses were 
returned to survey administrators. Out of 542 collected, 35 questionnaires were either incomplete or the answers 
were found to be unreliable, leaving a remaining 507 questionnaires that were retained for further data analysis.
Measurements
Multiple item scales were used to measure each construct. If possible, validated scales from previous 
literature were employed after a slight modification. In this study, the items in all scales except for loyalty and an 
item for distributive justice were measured on a 7-point Likert-type scale ranging from 1 (strongly disagree) to 
7(strongly agree). Customer loyalty was measured on a 7-point scale anchored with “least likely” (1) and “extremely 
likely” (7). One item to measure distributive justice, “compared to what you expected, the discount offered was”, 
was anchored with “less than expected” (1) and “more than expected” (7). A total of 28 items were employed to 
capture the 7 latent constructs: 3 exogenous (perceived justice) and 4 endogenous constructs (relationship variables). 
The perceived distributive justice was measured by a four-item scale adopted from Blodgett et al. (1997) 
and Smith et al. (1999). The five-item scale, the perceived interactional justice, was adapted from Tax et al. (1998), 
Smith et al. (1999), and Severt (2002). To measure the perceived procedural justice construct, we used a four-item 
scale adapted from Blodgett et al. (1997) and Smith et al. (1999). The satisfaction with service recovery scale was 
measured by four items and was adapted from Oliver and Swan (1989a) and Davidow (2000). The four-item scale 
for trust was adapted from Wong and Sohal (2002), Crosby et al. (1990), and Morgan and Hunt (1994). The 
commitment, adapted from Morgan and Hunt (1994) and Wong and Sohal (2002), was measured by four items. 
Finally, the 3-item scale, loyalty, was adapted from Hennig-Thurau, Gwinner, and Gremier (2002).