國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
Appendix A (The inventory of Character Attachment, CA)
A.1. Original questions
The CA, developed by the researcher and Yu-chu Yeh, was revised from the Character Attachment Scale (Lewis, Weber & Bowman, 2008). The CA was employed to measure the participants’ character attachment to the avatar in this study. It was a 6-point Likert scale with response options ranging from “totally disagree” to “totally agree”. With a total of 15 items, the CA included four factors: “identification and friendship” (5 items), “suspense of disbelief” (4 items), “control (3 items), and “Responsibility” (3 items). The test items included statements such as “I sometimes forget my own feelings and take those as my character.”, and “I enjoy controlling my character”
A.2. The processes of scale development A.2.1. Construct validity
This scale was revised from the Character Attachment Scale (Lewis, Weber & Bowman, 2008). One PhD student from linguistics department, one PhD student from computer science department, and one professor educational psychology department were asked for expert validity. Some questions were revised for the wordings.
A.2.2. Scoring
It was a 6-point Likert scale with response options ranging from “totally disagree” to
“totally agree” which represented 1 point to 6 points. The higher the sum of the scores, the higher character attachment to the avatar.
A.3. Questions deletions and selections
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
A.3.1. Item analysis
Total fifteen questions translated into Chinese according to the literature review for character attachment scale (CA). High score group and low score group was divided by using 27% of the total 113 participants as the cutting point. Independent t test was employed to test the difference between high score and low score group. The result showed that CR vales of item 6, 7, 8, and 9 were below 3.0. The researcher deleted one item each time. Item 6 was deleted first time, followed by item 9, 8 and 7. In the end, item 6, 9, 7 and 8 were deleted.
After deleting these items, all the CR values were greater than 3.0 (ps < .05). It suggested that the scores of high score group can be discriminated from those of low score group. Besides, the correlations between each item and total scores were higher than .468, ps < .001. Therefore, eleven items were kept for the next step analysis.
A.3.2. Exploratory factor analysis
This study employs exploratory factor analysis (EFA) to test the construct validity of the scale of perceived mastery experience. The results showed that Kaiser-Meyer-Olkin measure of sampling adequacy (KMO) was .751 which suggested the sample could be representative of the entire population and therefore it was suitable for factor analysis.
Besides, the value of Bartlett’s test of sphericity was 434.489, ps <.001 which suggested that the item correlations of PME could be used for factor extraction. Following this, principle axis factor (PAF) and direct oblmin rotation were used for factor extraction. This study decided to extract three factors according to the literature review. The first result showed that the factor loading of item 12 was greater than .35 across two different factors. Item 12 was deleted for the second EFA analysis. After item 12 was deleted, the second results showed that the factor loading of item 3 was greater than .35 across two different factors. Therefore, item 3 was deleted for the third time EFA analysis. The third results showed that KMO was .753.
‧
Bartlett’s test of sphericity was 331.116, ps <.001. PAF and direct oblmin rotation were used again for factor extraction. Three factors were extracted. All items’ factor loading were in the range of .626 and .876(See Table 1). The cumulative variance explained of two factors were 68%. Three factors were named “identification, “fantasy and control”, and “responsibility”, respectively. Besides, the correlations between each factor were greater than .294, p<.005 and the correlations between each factors and total scores were between .656 to .847, ps <.001(See Table 2).
Table A1
Pattern Matrix of Character Attachment
Questions Factor loading Communalities
1 2 3
1. I enjoy controlling my character.
(Original item 10) .876 .781
2. I control my character.
(Original item 11) .776 .744
3. I enjoy pretending I am my character.
(Original item 4) .744 .628
4. I daydream about my character.
(Original item 5) .626 .530
5. I know what my character needs
(Original item 14) .839 .701
6. .I make decisions with my character's best interests in mind
(Original item 15) .784 .673
7. I know what my character wants.
(Original item 13) .752 .683
8. .I sometimes forget my own feelings and take those as my character.
(Original item 1) .872 .731
9. I enjoy pretending that my character is a real person.(Original
item 2) .742 .689
Eigenvalues
%of variance explained
%of cumulative variance explained
3.485 1.487 1.188 38.719 16.522 13.202 38.719 55.240 68.443
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
Table A2
Correlations of Different Components of Character Attachment
Identification Fantasy& Control Responsibility Total
Identification 1.000 .347** .294** .656**
Fantasy&
Control .347** 1.000 .349** .847**
Responsibility .294** .349** 1.000 .702**
CA total .656** .847** .702** 1.000
**p< .01.
A.3.3. Reliability analysis
After exploratory factor analysis, total nine questions were left for reliability analysis. The Cronbach’s α coefficients was .783 for CA scale. It was .609, .761, and .778 for the factors of “identification, “fantasy and control”, and “responsibility”, respectively.
Furthermore, the corrected item-total correlations of all the items were.302 to.627. The corrected item-total correlations of identification were between .303 to .491; The corrected item-total correlations of fantasy and control were between to .302 to .627; The corrected item-total correlations of responsibility were between .396 to .550.
‧
Corrected Item-Total Correlations and Cronbach's α of Character Attachment
Factors and Questions Corrected
Item-Total Correlation
Cronbach's Alpha if Item
Deleted Factor: Identification (Cronbach’s ‘s α=.609)
1. I sometimes forget my own feelings and take those as my character.
2. I enjoy pretending that my character is a real person
.303 .491
.786 .759 Factor: Fantasy and control (Cronbach’s ‘s α=.761)
3. I enjoy pretending I am my character 4. I daydream about my character.
5. I enjoy controlling my character 6. I control my character
.562 Factor: Responsibility (Cronbach’s ‘s α=.778)
7. I know what my character wants 8. I know what my character needs
9. I make decisions with my character's best interests in mind
‧
A.4. Chinese version of character attachment scale