importance of entertainment in the shopping center experience: Evidence from Singapore, The

Journal of Real Estate Portfolio Management, Sep-Dec 2002 by Ibrahim, Muhammad F, Wee, Ng C

Interestingly, seven variables with mean scores above 5 belong to transport/travel attributes. These variables are mainly concerned with time issues and the comfort level during the journey to the shopping center. It is not surprising that these variables have high mean scores due to the humid weather in Singapore and the perception of time poverty among Singaporeans. Relating to the rest of the transport attributes, most have mean scores above 4 while "absence from stress" and "enjoyment of travel" scores a low 3.99 and 3.74, respectively.

Relating to the retailer attributes, the variables "high quality sales service," "availability of sale, promotion, discount and bargain," "wide variety of stores" and "availability of food-court/restaurants" top the list of importance ratings. Most of the variables relate to the atmosphere of the shopping center, such as "spaciousness of internal layout of shopping center," "good air quality in shopping center," etc., appear in the middle of the list. Variables such as the "availability of entertainment facilities," "availability of celebrations, activities and functions," "availability of unique store design" and "ease of finding car parking lot" have relatively low mean scores. However, these variables might be directly related to specific characteristics of the respondents.

Factor Analysis

Logically, it is important that all relevant attributes are included in the modeling of entertaining shopping experience. However, the problem of multi-collinearity among the attributes may pose a serious threat to its interpretation. Factor analysis serves two specific purposes. First, it links correlated variables into factors. Second, it serves as a solution for multi-collinearity so as to provide a better understanding of the underlying dimensions that influence the entertaining shopping experience. Exhibits 6-8 summarize the principal component analysis of the retailer, customer and transport mode/travel attributes.

Bartlett's test of sphericity and KMO are employed to determine the appropriateness of the data set for factor analysis (Kline, 1994; and Malhotra, 1996). High values (between 0.5 and 1.0) indicate that factor analysis is appropriate, while values below 0.5 imply that it may not be appropriate. On the other hand, the alpha value measures the reliability of the attributes in contributing to each factor (Cronbach, 1951). The latent root criterion (eigenvalues greater than one) has been adopted as the main method in extracting the number of factors for the analysis. Hair, Anderson, Tatham and Black (1998) regard the latent root criterion as the most appropriate and reliable method when there are twenty to fifty variables in the factor analysis.

Relating to the issue of factor loadings, which would be regarded as an important consideration in the factor analysis, Hair, Anderson, Tatham and Black (1998) state that factor loadings of 0.30 are considered significant if the sample size is 350 or greater. Thus, the attributes listed in association with each factor are those loading at or above 0.30. In analyzing the factor solutions, there are attributes that load more than 0.30 on more than one factor. These are indicated the lower loading by an asterisk next to the factor loading of the attribute. Finally, in most instances, the labels given to the factors are chosen to reflect the properties shared by the set of attributes loading above 0.30 within each factor.