Of the final group of 667 medical students Only 1% of the participants’ parents was not Turkish. Two-third of respondents have neighbors from different cultures, and 87.7% stated that they interacted with people from different cultures (Table 1). Of the participants, 79.3% reported that they spoke at least one foreign language other than Turkish. Within the last year, 64.8% of students examined patients from different cultures during their clinical practices. Only one-third of the students stated that they had heard the term ‘intercultural sensitivity’; 22.9% had learned from media, 16.8% from friends/family, and only 8,2% from university lectures (Table 1).
Table 1. Some Characteristics of Participants (Hacettepe Univ.-Turkey, 2018)
Characteristics
|
n
|
%1
|
Neighbor from different cultures (n=6582)
|
|
|
Yes
|
433
|
65.8
|
No
|
159
|
24.2
|
Don’t know
|
66
|
10.0
|
Interaction with people from different cultures (n=667)
|
Yes
|
585
|
87.7
|
No
|
82
|
12.3
|
Heard the term ‘intercultural sensitivity’ (n=6492)
|
Yes
|
200
|
30.8
|
No
|
449
|
69.2
|
Source of knowledge3
|
|
|
Media
|
153
|
22.9
|
Family/friends
|
112
|
16.8
|
University lectures
|
55
|
8.2
|
Others4
|
16
|
2.1
|
1 Column percentages
2 Various number of non-responses for each variable.
3 Multiple-choice answers; percentages were calculated separately from total (n=667).
4 Secondary education, foreign language course, TURKMSIC, IFMSA, social media, AFS Volunteers Association
Adaptation process of the scale
Language Validity
At the first stage of the study, items of the original scale were translated into Turkish by the researchers. At the second stage, a Turkish Language and Literature expert evaluated the draft translation with regard to ambiguity. As the third step, the items optimized in Turkish language were translated into English by a native English speaker professor at the American Culture and Literature Department at Hacettepe University. At the last step, the items translated back to English were compared with the original items of the scale before finalizing the Turkish version.
Exploratory Factor Analysis
For the factor analysis, exploratory and confirmatory analyses were conducted by randomly dividing the dataset into two approximately equal parts. In order to determine whether the items are compatible with the factor structure, exploratory factor analysis (EFA) based on polychoric correlations was applied for the first half of the scale, and oblique rotation method was used. The factor analysis was repeated by excluding 21 items with factor load smaller than 0.30 which were loaded to multiple dimensions. In the end, final version of the scale consisting of 23 items and 5 dimensions was obtained. In order to determine whether the data and sample are suitable for factor analysis, Kaiser-Meier-Olkin (KMO) test, which shows sampling adequacy, and Bartlett test, which shows sphericity were conducted. KMO value (0.874) shows that the sample size was sufficient for factor analysis. [24]. The result of Bartlett test has been found to be statistically significant, which supports the hypothesis that the correlations between items are different from zero.
The factor loads obtained from EFA, the eigenvalues and the variances explained by eigenvalues are shown in Table 2. As a result of the analysis, a five-dimensional structure with eigenvalues greater than 1 has been obtained; these eigenvalues explain 70.4% of the total variance. These five dimensions were as follows: Interaction Confidence, Interaction Awareness, Respect of Cultural Difference, Difficulty in Interaction and Interaction Enjoyment.
Table 2. Factor loadings, eigenvalues and the variance explained by eigenvalues
Factor Loadings
|
Items
|
Interaction Confidence
|
Interaction Awareness
|
Respect of Cultural Difference
|
Difficulty in Interaction
|
Interaction Enjoyment
|
Item 1
|
0.658
|
|
Item 2
|
0.499
|
|
Item 3
|
0.742
|
|
Item 4
|
0.838
|
|
Item 24
|
0.631
|
|
|
|
|
|
Item 11
|
0.872
|
|
Item 13
|
0.796
|
|
Item 15
|
|
0.599
|
|
|
|
|
Item 17
|
0.571
|
|
Item 18
|
0.888
|
|
Item 19
|
0.924
|
|
Item 20
|
|
|
0.749
|
|
|
|
Item 22
|
0.547
|
|
Item 23
|
0.641
|
|
Item 25
|
0.582
|
|
Item 30
|
0.610
|
|
Item 36
|
0.599
|
|
Item 38
|
|
|
|
0.602
|
|
|
Item 26
|
0.662
|
|
Item 33
|
0.537
|
|
Item 40
|
0.728
|
|
Item 41
|
0.771
|
|
Item 42
|
|
|
|
|
0.861
|
|
|
Eigenvalues
|
Total variance explained
|
Cumulative variance explained
|
|
Interaction Awareness
|
8.448
|
36.7%
|
36.7%
|
|
Difficulty in Interaction
|
3.045
|
13.2%
|
49.9%
|
|
Respect of Cultural Difference
|
1.759
|
7.7%
|
57.6%
|
|
Interaction Enjoyment
|
1.736
|
7.6%
|
65.2%
|
|
Interaction Confidence
|
1.197
|
5.2%
|
70.4%
|
|
KMO=0.874; Bartlett statistics= 3515.9 (p<0.00001)
|
Confirmatory Factor Analysis
In order to confirm the factor structure of the scale (23 items and 5 dimensions), confirmatory factor analysis (CFA) based on polychoric correlations was applied to the second half of the data, which was randomly divided into two parts. Factor loadings were found between 0.53 and 0.83. In Figure 1, the path diagram shows the path coefficients obtained from CFA analysis. In the confirmatory factor analysis, the fit indices were obtained to assess the fitness of the model constructed in the analysis. The (chi-square/degree of freedom) value was found as 2.96, which demonstrated a good model fit (i.e. between 1-3). The RMSEA (Root Mean Square of Error Approximation) fit index value was 0.077, reflecting a good fitness (<0.08). The Goodness of Fit Index (GFI) was found as 0.98, which shows that the model has a very good degree of fitness (>0.95 =very good). Other fit index values were as follows: NFI (Normed Fit Index) =1.00 (>0.95 very good), CFI (Comparative Fit Index) =1.00 (>0.95 very good) and RMR (Root Mean Square Residual) = 0.048 (<0.05 very good). According to these results, the model demonstrated a very good fitness and the construct validity of the 5-dimensions with 23-items scale was proven.
Item Statistics and Reliability Values
The level of difficulty of items ranged between 3.10 and 3.69 for the dimension ‘Interaction Confidence’; 3.57 and 3.69 for ‘Interaction Awareness’; 4.24 and 4.32 for ‘Respect of Cultural Difference’; 3.42 and 3.78 for ‘Difficulty in Interaction’ and 3.85 and 4.02 for ‘Interaction Enjoyment’ (Table 3). According to the discrimination values, the total value of items in each dimension were positive and above 0.40, which suggests that the items have a good degree of discrimination [25]. The items exemplify similar behaviors and have high levels of internal consistency. (Scale’s items were presented in Supplemental Material.)
Table 3. Item Statistics and Reliability Values
|
Difficulty
|
Discrimination Values
(Point Biserial Correlation)
|
Cronbach’s Alpha
(when item deleted)
|
Cronbach’s Alpha
|
Spearman-Brown Coefficient
|
Interaction Confidence
|
|
|
|
Item 1
|
3.69
|
0.671
|
0.819
|
0.851
|
0.869
|
Item 2
|
3.59
|
0.609
|
0.837
|
Item 3
|
3.10
|
0.654
|
0.824
|
Item 4
|
3.27
|
0.714
|
0.807
|
Item 24
|
3.45
|
0.676
|
0.817
|
Interaction Awareness
|
|
|
|
Item 11
|
3.57
|
0.645
|
0.650
|
0.772
|
0.706
|
Item 13
|
3.69
|
0.657
|
0.637
|
Item 15
|
3.59
|
0.524
|
0.780
|
Respect of Cultural Difference
|
|
|
|
Item 17
|
4.24
|
0.613
|
0.840
|
0.848
|
0.802
|
Item 18
|
4.32
|
0.752
|
0.779
|
Item 19
|
4.25
|
0.723
|
0.791
|
Item 20
|
4.31
|
0.663
|
0.816
|
Difficulty in Interaction
|
|
|
|
Item 22
|
3.50
|
0.617
|
0.793
|
0.826
|
0.819
|
Item 23
|
3.78
|
0.636
|
0.789
|
Item 25
|
3.42
|
0.520
|
0.813
|
Item 30
|
3.52
|
0.656
|
0.785
|
Item 36
|
3.45
|
0.559
|
0.805
|
Item 38
|
3.54
|
0.578
|
0.802
|
Interaction Enjoyment
|
|
|
|
Item 26
|
3.86
|
0.561
|
0.797
|
0.817
|
0.801
|
Item 33
|
3.85
|
0.530
|
0.803
|
Item 40
|
4.02
|
0.647
|
0.772
|
Item 41
|
3.86
|
0.621
|
0.778
|
Item 42
|
3.96
|
0.693
|
0.755
|
For all items of the scale
|
Cronbach’s Alpha
|
0.906
|
Spearman-Brown Coefficient
|
0.779
|
Cronbach’s alpha values were 0.851 for ‘Interaction Confidence’, 0.772 for ‘Interaction Awareness’, 0.848 for ‘Respect of Cultural Difference’, 0.826 for ‘Difficulty in Interaction’ and 0.817 for ‘Interaction Enjoyment’, and 0.906 for the total score. These values suggested that the scale has a high level of reliability [26]. The Spearman-Brown coefficients for the whole scale and the dimensions show that the scale’s split-half reliability values were also sufficient.
Intercultural Sensitivity Status of the Medical Students
The mean total score of students was 85.84±11.44, which shows a good level of intercultural sensitivity. (Table 4). However, there are some outliers (1% got 23 points).
Table 4. Dispersion Statistics of Total and Sub-Dimension Scores of the Students (Hacettepe Univ.-Ankara, 2017)
Scale and Sub-Dimensions
|
Median
|
1st–3rd Quarter
|
Min–Max
|
Interaction Confidence
|
17.10±3.67
|
17.00
|
15.00-20.00
|
5.00-25.00
|
Interaction Awareness
|
10.86±1.94
|
11.00
|
9.00-12.00
|
3.00-15.00
|
Respect of Cultural Differences
|
17.12±2.70
|
18.00
|
16.00-20.00
|
4.00-20.00
|
Difficulty in Interaction
|
21.21±4.22
|
21.00
|
18.00-24.00
|
6.00-30.00
|
Interaction Enjoyment
|
19.55±3.00
|
20.00
|
18.00-21.00
|
5.00-25.00
|
TOTAL
|
85.84±11.44
|
86.00
|
79.00-93.00
|
23.00-115.00
|
The scores of female students were significantly higher than males on ‘Respect of Cultural Differences’ and ‘Interaction Enjoyment’ dimensions (Mann-Whitney U test; p=0.002 and p=0.001, respectively)
Students who have neighbors from different cultures have significantly higher scores on the sub-dimensions of ‘Interaction Confidence’, ‘Difficulty in Interaction’ and ‘Interaction Enjoyment’ (Mann-Whitney U test; p=0.003, p=0.036, respectively, p<0.001). Students who have the chance to interact with people from different cultures had higher scores on the sub-dimensions of ‘Interaction Confidence’, ‘Interaction Awareness’ ‘Difficulty in Interaction’ and ‘Interaction Enjoyment’ (Mann-Whitney U test; p<0,001, p=0.001, p<0.001, p<0.001, respectively). The students who examine patients from different cultures had significantly higher scores on the sub-dimensions of ‘Interaction Confidence’ (p<0.001), ‘Interaction Awareness’ (p=0.004) and ‘Difficulty in Interaction’ (p=0.046) (Mann-Whitney U test) (Table 5).