CHAPTER 3 METHODOLOGY
4.2 STAGE 2: SCALE ITEM GENERATION
4.4.3 RESULTS OF THE FIVE-FACTOR SOLUTION
The five factors explained 9.01% (the beverage fit), 9.19% (the sight fit), 10.23% (the food fit), 9.02% (the sound fit), and 10.21% (the smell fit) of the variance in the solution after rotation. SPSS did not output the total amount of variance accounted because of the correlation and overlap among the factors with oblique rotation. All of the items reported reasonably high communality values with the minimum of 0.406 and the maximum of .781. Communality values show the percentage of common variance
explained (Hair et al., 2010). For example, values above .40 show that the item is sharing at least 40% of the variance with the set of factors. Costello and Osborne (2005)
recommend retaining items with communalities of .40 or above and the common magnitude of communality values in the social sciences ranges from low to moderate of .40 to .70. When an item has a communality of less than .40, it may not be related to the other items or suggest the possibility of an additional factor to be explored. All items in the analysis were satisfactory to Costello and Osborne’s (2005) criterion.
The items signified strong markers of the factors with relatively high loading values. Loadings for each item represent the level of correlation between the item and its relevant factor, which also means the contribution of factors to each item (Hair et al.,
2010). According to the results, each factor was defined by at least five items, where factor structures matched the prior assumption by the researcher. The standardized loading and communality values are reported in Table 4.7. The highest loading value of each item on the five factors is highlighted in bold.
Table 4.7. Factor loadings and communalities for 26 items of the SIF Scale. Beverage Fit Sight Fit Food Fit Sound Fit Smell Fit Communality BEVER2: The
beverages at the event were part of the local culture.
.882 .014 -.033 -.009 -.020 .729
BEVER3: The
beverages at the event were unique to the local culture.
.811 .106 -.117 .009 .021 .699
BEVER4: The
beverages at the event left a strong impression of the local culture.
.729 -.101 .238 .066 -.032 .749
BEVER1: The
beverages at the event were representative of the local culture.
.715 .030 .144 .019 .005 .742
BEVER5: The
beverages at the event were consistent with the image of the local culture.
.538 .050 .378 -.118 .026 .726
SIGHT2: When I looked around the event, I was reminded of the local culture.
.071 .769 -.065 -.004 .117 .699
SIGHT3: The interior design and decorations at the event were representative of the local culture.
.011 .768 .086 .051 -.130 .628
SIGHT1: The
architecture at the venue was representative of the
SIGHT4: The sights at the event were unique to the local culture.
-.024 .638 .034 .090 .043 .582
SIGHT5: The sights at this event left a strong impression of the local culture.
-.066 .470 .248 .061 .027 .517
SIGHT6: The event provided a good view of the surrounding of the local area.
.048 .461 .136 .047 -.082 .406
FOOD1: The food at the event was representative of the local culture.
.078 .021 .855 -.025 -.054 .738
FOOD4: The food at the event left a strong impression of the local culture.
.015 -.029 .854 -.023 .090 .764
FOOD5: The food at the event was consistent with the image of the local culture.
.087 .091 .740 .036 -.110 .681
FOOD3: The food at the event was unique to the local culture.
.143 -.028 .590 .040 .121 .733
FOOD2: The food at the event was part of the local culture.
-.036 .121 .478 .009 .117 .517
SOUND2: When I listen to the music at the event, I was reminded of the local culture.
.157 -.072 -.215 .860 .174 .762
SOUND3: The music played at the event was unique to the local culture.
.130 -.084 -.050 .781 .110 .732
SOUND5: The music at the event was consistent with the image of the local culture.
-.058 .016 .260 .764 -.089 .764
SOUND1: The music at the event was
representative of the local culture.
-.156 .294 -.076 .719 -.090 .625
impression of the local culture.
SMELL2: When I smelled the
surroundings at the event, I was reminded of the local culture
-.028 .060 -.049 -.050 .938 .766
SMELL5: The smells at the event were
consistent with the image of the local culture.
.011 -.035 .035 .024 .871 .778
SMELL3: The smells at the event were uniquely part of the local culture.
-.015 -.109 .174 .144 .709 .767
SMELL1: The smells at the event were
representative of the local culture.
.198 .146 -.114 .012 .664 .704
SMELL4: The smells at the event left a strong impression of the local culture.
-.140 .145 .388 -.030 .565 .781
According to Hair et al. (2010) revision of a factor solution may be required if a factor has too few items, an item has no significant loadings, an item’s communality is too low, or an item has a cross-loading. Neither the number of items in each factor nor the loading values of items were an issue in the solution although a few of items showed somewhat low loadings (SIGHT5, SIGHT6, FOOD2). Those low loadings were still higher than the recommended benchmark of .40 to show that those items were
sufficiently related to latent variables. The only problematic issue of the solution was that BEVER5 loaded on both the beverage fit (.538) and the food fit (.378). As said by
Costello and Osborne (2005) a cross-loading is when the loading value is .32 or higher on two or more latent variables. A simple structure with items loaded cleanly on one factor
acceptable as long as the relationship among the cross-loaded item and the factors is explainable (Costello & Osborne, 2005). In this case, the cross-loading of BEVER5 make sense considering food and beverage are two highly interrelated elements. Furthermore, there was just one cross-loaded item, and the solution was easy to interpret. After
evaluating multiple criteria, the researcher decided not to re-specify the factor model and moved onto the next step of reliability check in which addition assessment was executed to check whether any revision or removal of items was necessary.