A total of 475 Iranian athletic adolescent girls participated in the survey. The participants' mean (SD) age was 15.68 (1.89). Regarding sports status, 246 (52.9%) of the participants were Individual sports, and 229 (47.1%) played team sports (see Table 1).
TABLE 1 Socio-demographic characteristic (n=475)
|
N
|
%
|
M
|
SD
|
Sport Type
|
|
|
|
|
Individual Sport
|
246
|
52.9%
|
68.00
|
16.17
|
Team Sport
|
229
|
47.1%
|
65.17
|
14.61
|
Age Group
|
|
|
|
|
14
|
113
|
23.5%
|
65.94
|
15.42
|
15
|
166
|
35.8%
|
68.20
|
15.56
|
16
|
148
|
30.8%
|
65.95
|
12.93
|
17
|
48
|
9.9%
|
65.02
|
21.57
|
3.1 Factor Structure
The Kaiser- Meyer-Olkin (KMO) index was 0.908, above the recommended value of 0.6, and Bartlett’s test for sphericity reached statistical significance (x2 =2855.39,p≤0.001) that the data were suitable for factor analysis. We conducted CFA on a sample of 475 athletic adolescent girls. The initial CFA began with five factors and 25 items. The final CFA showed five factors, 25-item solutions: the factor loadings and fit statistics(Table 2).
TABLE 2. CFA of Connor-Davidson Resilience Scale (CD-RISC)
Items
|
Factor Loading
|
Eigenvalue
|
|
Total
|
% of variance
|
Cumulative %
|
Overall Alpha
|
ITEM 1
|
0.43
|
9.377
|
37.509
|
37.509
|
|
ITEM 2
|
0.41
|
1.635
|
6.538
|
44.047
|
|
ITEM 3
|
0.83
|
1.279
|
5.118
|
49.165
|
|
ITEM 4
|
0.63
|
1.167
|
4.148
|
53.313
|
|
ITEM 5
|
0.62
|
1.016
|
3.988
|
57.301
|
|
ITEM 6
|
0.58
|
.882
|
3.526
|
60.828
|
|
ITEM 7
|
0.57
|
.865
|
3.461
|
64.288
|
0.912
|
ITEM 8
|
0.46
|
.788
|
3.153
|
67.441
|
|
ITEM 9
|
0.65
|
.768
|
3.070
|
70.511
|
|
ITEM 10
|
0.76
|
.696
|
2.783
|
73.295
|
|
ITEM 11
|
0.70
|
.651
|
2.602
|
75.897
|
|
ITEM 12
|
0.70
|
.609
|
2.437
|
78.334
|
|
ITEM 13
|
0.51
|
.584
|
2.334
|
80.669
|
|
ITEM 14
|
0.75
|
.520
|
2.080
|
82.748
|
|
ITEM 15
|
0.76
|
.514
|
2.057
|
84.805
|
|
ITEM 16
|
0.69
|
.504
|
2.014
|
86.820
|
|
ITEM 17
|
0.65
|
.480
|
1.919
|
88.738
|
|
ITEM 18
|
0.47
|
.430
|
1.721
|
90.460
|
|
ITEM 19
|
0.78
|
.412
|
1.648
|
92.108
|
|
ITEM 20
|
0.72
|
.378
|
1.511
|
93.619
|
|
ITEM 21
|
0.76
|
.358
|
1.433
|
95.052
|
|
ITEM 22
|
0.67
|
.345
|
1.380
|
96.432
|
|
ITEM 23
|
0.39
|
.320
|
1.279
|
97.711
|
|
ITEM 24
|
0.71
|
.295
|
1.182
|
98.892
|
|
ITEM 25
|
0.59
|
.277
|
1.108
|
100.000
|
|
The standardized factor loading for the five CD-RISC items can be seen in Figure 1, where all the factor loadings were significant and in the expected direction.
TABLE 3. Confirmatory Factor Analysis (CFA) and Fit indexes
model
|
RMSEA (CI 90%)
|
sbX2
|
RMR
|
SRMR
|
CFI
|
NFI
|
IFI
|
RFI
|
25 items
|
0.078 (0.067 -0.077)
|
915.40
|
0.057
|
0.051
|
0.97
|
0.96
|
0.97
|
0.95
|
Legend: RMSEA: Root Mean Square Error of Approximation; RMR: Root Mean Square Residual; SRMR: Standardized RMR; CFI: Comparative Fit Index; NFI: Normed Fit Index; IFI: Incremental Fit Index; RFI: Relative Fit Index
The CFA results for a five-factor structure are shown in Table 3. These results are acceptable because the factor loadings of all items were significant, and all items except item 23 were above 0.40. Model fit was estimated using the following fit indices: Root Mean Square Error of Approximation (RMSEA; criterion 0.08) and its confidence level of 90%, Standardized Root Mean Square Residual (SRMR; criterion 0.09), Comparative Fit Index (CFI; criterion 0.90), Normed Fit Index (NFI; criterion 0.90), Incremental Fit Index (IFI; criterion 0.90), Relative Fit Index (RFI; criterion 0.90)[51]. The CFA results also showed that the five-factor structure fit the data well. In the present study, the fit indices of the model were RMSEA= 0.072; SRMR =0.051, RMR =0.057, CFI =0.97, NFI=0.96, IFI=0.97, RFI=0.95. All items of the loadings showed a significant factor(Table 3).
3.2 Content Validity
According to their content validity, items must cover the objectives of the topic in which they are included. The content validity ratio and content validity index is the most common quantitative methods to evaluate the content validity of the content on Likert scales[52]. Seventeen experts were asked to rate the following items: 1) Not necessary, 2) Useful, but not necessary, and 3) Necessary to evaluate the content validity ratio. Based on Lawshe’s [53] table and the number of experts, it is essential if an index number is more significant than 49% to have the corresponding item at p<0.05. The minimum acceptable CVR for 15 and 20 people is recommended to be 0.49 and 0.42, respectively. The following formula determines the validity ratio: ne is the number of experts who graded the items at three, and N is the number of professionals who paid attention to the item.
Content validity ratio = =
On the basis of Lawshe's [53] table, when eight panels of experts are involved, the CVR should be equal to or higher than 0.78. This is the reason why no items were eliminated in this selection process. To qualify as valid, an item must have a score greater than 0.79 on the CVI, which measures these three factors ("simplicity and fluency", "relevance", and "clarity")[54].
In addition, the content validity index (CVI) was calculated to assess the relevance of the items. CVR is calculated for each item separately, while CVI is calculated for the entire scale. There are two types of CVI: item-level CVI (I-CVI) and scale-level CVI (S-CVI).
Lynn [55] argues that if a panel of eight experts gave an I-CVI, it should be 0.78. The S-CVI is calculated in two ways: S-CVI/UV (based on Universal Agreement) and S-CVI/Ave (based on average). Using the average method, there are two ways for me to calculate S-CVI: 1. By summing the scores from the I-CVI and the number of items, 2. A ratio between the proportional relevance rating and the number of experts is calculated [56, 57]. According to this survey, the I-CVI for all items was higher than 0.79. Furthermore, the S-CVI/Ave value was 0.91, indicating satisfactory content validity. The CVI did not require any adjustments or omissions.
There is significant evidence of higher CVR than what has also been observed, suggesting acceptable content validity (see Table 4). It's important to emphasize that each of the 25 items has acceptable content validity. Furthermore, the S-CVI/Ave (0.947) value was higher than 0.9.
Table 4 CVI and CVR values for the items of CD-RISC scale
|
CVI
|
CVR
|
Items
|
Difficulty
|
Ambiguity
|
Relevancy
|
Urgency
|
ITEM 1
|
1
|
0.9
|
0.8
|
0.8
|
ITEM 2
|
0.9
|
0.9
|
1
|
1
|
ITEM 3
|
1
|
0.9
|
1
|
1
|
ITEM 4
|
1
|
0.9
|
1
|
1
|
ITEM 5
|
0.9
|
0.9
|
1
|
1
|
ITEM 6
|
1
|
0.9
|
1
|
1
|
ITEM 7
|
1
|
0.9
|
1
|
1
|
ITEM 8
|
1
|
0.9
|
0.9
|
0.8
|
ITEM 9
|
1
|
0.9
|
1
|
1
|
ITEM 10
|
1
|
0.9
|
1
|
1
|
ITEM 11
|
1
|
0.9
|
1
|
0.8
|
ITEM 12
|
1
|
0.9
|
1
|
1
|
ITEM 13
|
1
|
0.9
|
1
|
1
|
ITEM 14
|
0.7
|
0.9
|
1
|
0.6
|
ITEM 15
|
1
|
0.9
|
0.9
|
0.8
|
ITEM 16
|
1
|
0.9
|
1
|
1
|
ITEM 17
|
1
|
0.9
|
1
|
1
|
ITEM 18
|
1
|
0.9
|
1
|
1
|
ITEM 19
|
1
|
0.9
|
1
|
1
|
ITEM 20
|
1
|
0.9
|
1
|
1
|
ITEM 21
|
1
|
0.9
|
1
|
1
|
ITEM 22
|
1
|
0.9
|
0.9
|
0.8
|
ITEM 23
|
1
|
0.9
|
1
|
1
|
ITEM 24
|
1
|
0.9
|
1
|
1
|
ITEM 25
|
1
|
0.9
|
1
|
0.8
|
3.3 Face Validity
Using face validity, this study evaluated its appropriateness, representativeness, readability, and clarity. Evaluation of surveys can be accomplished in a variety of ways. The most common method is cognitive interviews. Before collecting data from a large sample, cognitive interviews are useful for researchers to clarify the items, ensure adequate coverage of the content, and modify the questionnaire if any questions are unclear[58]. Twelve athletic adolescent girls participated in cognitive interviews to determine item complexity, vagueness, and comprehensibility of interview items. Twenty-five questions were ultimately compiled as a final scale. As a result, no changes had to be made to the Persian scale when designing the CD-RISC's final Persian version, as there were no unclear Persian terms.
3.4 Internal Consistency Reliability
The internal consistency reliability of the Iranian Connor-Davidson Resilience Scale (CD-RISC) was assessed using Cronbach’s alpha for all participants and was 0.908. The internal consistency of the Iranian Connor-Davidson Resilience Scale (CD-RISC) was similar.
3.5 Follow-up Study and Test-Retest Reliability
Temporal stability using a test-retest strategy in a small subsample of 105 participants in the main study was randomly selected and asked to complete the CD-RISC again after two weeks. The results showed that after this period, the calculated test-retest coefficient was 0.81(CI=0.79-0.83).
3.6 Convergent Validity
The convergent validity of the CD-RISC was assessed by correlation with the quality of mindfulness and General Self-efficacy (GSE) scores. Significant positive correlations of the CD-RISC subscales with Quality of Mindfulness and General Self-efficacy (GSE) ranging from 0.19 to 0.48 indicated acceptable convergent validity (Table 5).
TABLE 5. Pearson correlation between CD-RISC with Quality of Mindfulness, General Self-efficacy (GSE), and Alexithymia among participants
|
F1
|
F2
|
F3
|
F4
|
F5
|
CD-RISC Total
|
Quality of Mindfulness
|
0.28**
|
0.23**
|
0.21**
|
0.19**
|
0.20**
|
0.31**
|
GSE
|
0.41**
|
0.37**
|
0.28**
|
0.23**
|
0.25**
|
0.44**
|
Alexithymia
|
-0.31**
|
-0.28**
|
-0.22**
|
-0.20**
|
-0.19**
|
-0.34**
|
*p<0.05 **p<0.01
|
|
|
|
|
|
|
3.7 Discriminant Validity
As shown in Table 5, the negative correlations between the CD-RISC subscales and the alexithymia score ranged from -0.19 to -0.34. All correlations were statistically significant at p<0.01, indicating acceptable discriminant validity.
TABLE 6. Multiple regression analysis for prediction of aggression by subscales of CD-RISC. (N=475)
|
B
|
S. E
|
Beta
|
T
|
P
|
R
|
R2
|
F
|
P
|
Constant
|
42.352
|
3.123
|
|
6.347
|
.001
|
|
|
|
|
F1
|
.171
|
.058
|
.146
|
3.914
|
.001
|
.457
|
.208
|
19.64
|
.001
|
F2
|
.159
|
.062
|
.127
|
3.531
|
.001
|
|
|
|
|
F3
|
.154
|
.054
|
.114
|
2.531
|
.023
|
|
|
|
|
F4
|
.161
|
.059
|
.101
|
2.344
|
.039
|
|
|
|
|
F5
|
.112
|
.064
|
.093
|
1.762
|
.087
|
|
|
|
|
Multiple linear regression was calculated to predict aggression based on the components of CD-RISC; the results showed that the components of CD-RISC significantly predicted aggression in Iranian athletic adolescent girls (F (5, 470) = 19.64, P≤ 0.001), with R2 of 0.21 (see Table 6).