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

 
2007 Annual International CHRIE Conference & Exposition 
537
Pre Test and Pilot Study 
The questionnaire was pre-pilot tested with a sample of 29 hospitality students. Minor revisions to the 
instrument were made following the pretest. A convenience sample of 128 undergraduate and graduate students 
enrolled in hospitality courses at a large university in the southwest were surveyed in a pilot test. Cronbach 
α
reliabilities for the different parts of the survey ranged from .78 to .89.
Factor analysis procedures were used to identify the underlying structures of the facilitator, constraint and 
gender issue items. Outliers and missing data were identified and eight cases with no hospitality work experience 
were filtered from the pilot test data. The data were evaluated for normality and homogeneity of variance-covariance 
matrices was examined; the results were satisfactory for both analyses. Multiple regressions were conducted to 
obtain the Mahalanobis distance with p<.001 to check for outliers and screen for multicollinearity. The Kaiser-
Meyer-Olkin (KMO) tests were performed to test the factorability of the correlation matrices. The KMO values were 
greater than .60, indicating that factor analysis was appropriate for examining the logical groupings of the items.
For the facilitator items, 2 factors, labeled 
Personal/External 
and 
Attitude
, were extracted using principal 
axis factoring based on a scree test. Target factor loadings ranged from .33 to .69 and from .50 to .76, respectively. 
A total of 34.62% of the variance was accounted for by the two factors. For the 15 items that constrain women’s 
career advancement, one variable (“credibility”) was deleted because it loaded on two factors with similar loadings. 
Three factors, labeled 
Equity
, 
Familial
, and 
Lack of support
, accounted for 56.33% of the variance. The loadings 
ranged from .50 to .83 for 
Equity,
.51 to .85 for 
Familial, 
and .50 to .92 for 
Lack of support. 
Factor analysis on the 
gender issue items supported a single factor solution, explaining 49% of the variance. Factor loadings for the 7 items 
ranged from .52 to .84; the factor was labeled 
Gender issues
. A subscale score for each factor was calculated by 
computing the mean of the variables of each subscale. Overall mean scores for six factors ranged from 3.32 to 4.47. 
The reliabilities (Cronbach 
α
) for the factors ranged from .55 to .86. (See Zhong & Couch, in press.)