Sample
Two hundred seventy-eight participants out of a total of 300 responded the questionnaire, resulting in an overall response rate of 92.6%. Gender, only child and ethnicity characteristics are listed in Table 1.
Score distributions
The CTDA descriptive data is shown in Table 2. Across dimensions, “systematicity and analyticity” obtained the highest score (43.93±5.71), whereas “maturity and skepticism” scored the lowest (28.41±3.96). The skewness and kurtosis coefficients of the whole questionnaire were acceptable, with the former ranging from -0.98 to -0.32 and the latter ranging from -0.13 to 2.05. There were no floor effects in the three categories. However, item 19 showed a significant ceiling effect (23.74%).
Table 1. Sociodemographic characteristics of participants (n=278)
Characteristic
|
No.
|
%
|
Gender
|
|
|
Male
|
113
|
40.6
|
Female
|
165
|
59.4
|
Only child
|
|
|
Yes
|
148
|
53.2
|
No
|
130
|
46.8
|
Ethnicity
|
|
|
Han nationality
|
214
|
77.0
|
Ethnic minorities
|
64
|
23.0
|
Table 2. Item scores and internal consistency of each factor (n=278)
Item
|
Mean±SD
|
Skewness
|
Kurtosis
|
Floor (%)
|
Ceiling (%)
|
Factor 1: “Systematicity and analyticity”
|
43.93±5.71
|
-0.61
|
1.04
|
0.00
|
1.79
|
(1) Find a conflicting statement in an article
|
5.27±1.09
|
-0.50
|
0.26
|
0.00
|
11.87
|
(2) Analyze all information before making a judgment
|
5.47±.96
|
-0.93
|
2.05
|
0.35
|
10.07
|
(3) Evaluate the rationality of the conclusion
|
5.43±0.95
|
-0.52
|
0.50
|
0.00
|
10.79
|
(4) Make sure the information is reliable
|
5.56±1.05
|
-0.76
|
0.63
|
0.00
|
17.26
|
(5) Know details about controversial problems
|
5.65±.98
|
-0.79
|
0.87
|
0.00
|
17.98
|
(6) Draw conclusions by logical thinking
|
5.39±.99
|
-0.62
|
1.12
|
0.35
|
11.15
|
(7) Examine advantages and disadvantages of each opinion
|
5.68±.96
|
-0.82
|
1.09
|
0.00
|
17.62
|
(8) Contemplate the right and wrong of things
|
5.49±1.01
|
-0.85
|
1.55
|
0.35
|
13.30
|
Factor 2: “Inquisitiveness and conversance”
|
32.67±4.59
|
-0.32
|
-0.13
|
0.00
|
2.15
|
(9) Learn as much as possible
|
5.12±1.25
|
-0.61
|
0.21
|
0.71
|
11.51
|
(10) Make an effort to collect all relevant information before a decision
|
5.64±1.05
|
-0.97
|
1.40
|
0.35
|
19.06
|
(11) Solve a problem with updated information
|
5.64±.95
|
-0.54
|
0.54
|
0.00
|
17.62
|
(12) Delve into any novel viewpoint
|
5.65±1.00
|
-0.72
|
0.95
|
0.00
|
20.14
|
(13) Make a decision with proper rules
|
4.93±1.15
|
-0.59
|
0.48
|
0.71
|
5.39
|
(14) Strive to seek potential solutions before deciding
|
5.70±.92
|
-0.91
|
1.23
|
0.00
|
15.82
|
Factor 3: “Maturity and skepticism”
|
28.41±3.96
|
-0.73
|
1.48
|
0.00
|
8.27
|
(15) Raise questions and respond to others’ opinions
|
5.45±1.06
|
-0.80
|
1.17
|
0.35
|
13.66
|
(16) Comprehend and listen to different opinions before communicating
|
5.72±.94
|
-0.92
|
1.42
|
0.00
|
18.34
|
(17) Tolerate different viewpoints
|
5.80±.87
|
-0.82
|
1.77
|
0.00
|
19.78
|
(18) Correct the viewpoint with enough evidence
|
5.70±1.09
|
-0.98
|
0.89
|
0.00
|
23.02
|
(19) Question my prejudice, assumption, or belief
|
5.75±1.01
|
-0.94
|
1.85
|
0.35
|
23.74
|
Reliability
The overall Cronbach’s a coefficient of the CTDA was acceptable (0.923) and showed great internal consistency. Three domains were considered to show excellent internal consistency (0.814-0.861). The overall split-half reliability coefficient of the CTDA was excellent (0.889). The retest response rate was 83.6% (41/49), and test-retest reliability revealed a statistically significant ICCs for the three domains. In this study, the ICCs ranged from 0.710 (“Inquisitiveness and conversance”) to 0.795 (“Maturity and skepticism”) (p<0.01) (Table 3).
Table 3. CTDA internal consistency and test-retest reliability
|
Cronbach's a coefficient(n=278)
|
ICCs (95%CI) (n=43)
|
Dimensions
|
|
|
Factor 1: Systematicity and analyticity
|
0.861
|
0.722(0.357-0.788) **
|
Factor 2: Inquisitiveness and conversance
|
0.814
|
0.710(0.336-0.809) **
|
Factor 3: Maturity and skepticism
|
0.852
|
0.795(0.387-0.847) **
|
CTDA
|
0.923
|
0.792(0.336-0.809) **
|
**p<0.01. ICCs, intraclass correlation coefficients.
Validity
Construct validity
We performed a confirmatory factor analysis of the three-factor structure with 19 items to demonstrate that the structure showed an acceptable fit with the data (χ2=410.747, df=149, CMIN/DF=2.757, CFI=0.896, AGFI=0.828, p<0.001, RMSEA=0.080[90% CI: 0.071 to 0.089]). factor loadings were higher than 0.40 and ranged from (r = 0.50 - 0.85), as illustrated in Fig.1.
Correlation analysis between CTDA dimensions
The CTDA showed significant correlation between any two assessment categories (r = 0.605-0.722), with p values less than 0.01. The correlations between assessment categories based on Pearson’s correlation are shown in Table 4.
Table 4. Correlations between CTDA dimensions(n=278)
Dimension
|
1
|
2
|
3
|
1
|
—
|
0.722**
|
0.613**
|
2
|
0.722**
|
—
|
0.605**
|
3
|
0.613**
|
0.605**
|
—
|
**p < 0.01.
Concurrent validity
The correlation coefficients of the three dimensions with the SSACT scale and subscales are shown in Table 5. This analysis demonstrated a significantly correlation (p<0.01) between the whole CTDA scale and subscales and the whole SSACT scale. The correlation coefficients were higher than 0.40 (r=0.459-0.775), indicating that the concurrent validity of the CTDA was satisfactory.
Table 5 Concurrent validity of the CTDA (n=278)
Dimensions
|
Correlation coefficient
|
Active learning
|
Critical thinking
|
SSACT
|
Factor 1: Systematicity and analyticity
|
0.652**
|
0.769**
|
0.771**
|
Factor 2: Inquisitiveness and conversance
|
0.620**
|
0.661**
|
0.691**
|
Factor 3: Maturity and skepticism
|
0.459**
|
0.542**
|
0.543**
|
CTDA
|
0.669**
|
0.763**
|
0.775**
|
**p < 0.01.
Convergent validity and discriminant validity
Based on item-domain correlations, the scores of each item correlated with their own dimension to an acceptable degree (r=0.651-0.857, p<0.01), and the convergent validity of the CTDA was excellent. In addition, whole items showed a higher correlation with their own categories than with other categories, and the discriminant validity was satisfactory, as shown in Table 6.
Table 6. Convergent and discriminant validity of the CTDA (n=278)
Domain
|
Correlation coefficient range
|
Convergent validity
|
Discriminant validity
|
Convergent validity
|
Discriminant validity
|
Success/total
|
Percentage (%)
|
Success/total
|
Percentage (%)
|
1
|
0.663-0.786**
|
0.360-0.578**
|
8/8
|
100
|
8/8
|
100
|
2
|
0.651-0.806**
|
0.272-0.629**
|
6/6
|
100
|
6/6
|
100
|
3
|
0.720-0.857**
|
0.436-0.545**
|
5/5
|
100
|
5/5
|
100
|
**p < 0.01.