3.1. Methodology
Our total sample was 396 students who met our inclusion criteria - tertiary students - and agreed to participate in our study. Their average age was 23.78 years (min = 19; max = 53). There was a majority of females (n= 252) compared to males (n=144). The majority were single (n=373) and a minority were married, in a common-law relationship (n=19) or divorced (n=4). Furthermore, most did not have children (n=382) and were not cohabiting with an elderly person (n=340). We also note that the majority perceived themselves to be in good health and did not suffer from any chronic diseases (n=371). Most of the students were studying to obtain a bachelor's degree (n=357), and the rest of the sample was distributed between "postgraduate" (n=19), "master's" (n=3) and "other" (n=88). Finally, concerning professional activity alongside studies, the results were more mixed, with 177 students declaring that they had a job compared to 218 (see Table 3).
Table 3, Descriptive statistics of the sample
Demographic characteristics
|
n (%)
|
Sex
|
Female
|
252 (63,64%)
|
Male
|
144 (16,36%)
|
Marital Status
|
Single
|
373 (94,19%)
|
Married / Common law
|
19 (4,80%)
|
Divorced
|
4 (1,01%)
|
widower
|
0 (0%)
|
Education Level
|
Bachelor/License
|
357 (76,45%)
|
Post-grade
|
19 (4,07%)
|
Master
|
3 (0,64%)
|
Doctorat
|
0 (0%)
|
Other
|
88 (18,84%)
|
Age
|
Average
|
23,78
|
Minimum
|
19
|
Maximum
|
53
|
25ème percentile
|
21
|
50ième percentile
|
23
|
75ième percentile
|
24
|
Dependent child(ren)
|
Yes
|
14
|
No
|
382
|
Cohabitation with an elderly person
|
Yes
|
56
|
No
|
340
|
Exercise a student job
|
Yes
|
177
|
No
|
218
|
Chronic Illness
|
Yes
|
25
|
No
|
371
|
3.2 Results: Reliability study
3.2.1. Measuring Fidelity
The Cronbach's alpha for our translation is equal to .850 and attests to good fidelity, since this score is higher than the .70 threshold recommended by Nunnaly (1978). This fidelity score is however lower than the original version of the HHI, since Herth (1992) ) obtained an alpha of .97. However, this score is close to other translations of the HHI, such as the German (alpha = .82) (Geiser et al., 2015) or Norwegian (alpha = .81) (Benzein & Berg, 2003).
No item deletion brings any really significant gain. Only items 3 and 6 would result in an alpha of .876 and .869 respectively. It should be noted that these are the two inverted items in the questionnaire.
3.2.2. Factors Analysis
A principal component analysis (PCA) was conducted to test the construct validity of the HHI-F. The objective of the PCA is to check whether the factor structure of our translation is similar to that of the original scale (Herth, 1992).
First, we obtain a Kaiser-Meyer-Olkin (KMO) index of .916, with a highly significant Bartlett's sphericity test (p < .000). This allows us to ensure that the HHI-F items are highly correlated with each other.
The Cattell (1966) scree test (Eigenvalues) suggests a 2-component structure for the HHI-F (see. Figure 4). Indeed, the first two components have an eigenvalue greater than 1, with an explanation of variance of 44.69% for the first component, and 9.13% for the second. The third component scores .978 and explains 8.15% of the variance. Nevertheless, after several principal component analyses, on the one hand based on the eigenvalue, and on the other hand based on a fixed number of 3 factors to be extracted in accordance with the number of dimensions of the original HHI scale (Herth, 1992), we do not obtain radically different structures to the one obtained from a PCA based on an eigenvalue higher than 1. Indeed, only item 12 explains a third component, the two other components being explained by the same items as the PCA based on the eigenvalue. Therefore, we decided to keep the eigenvalue-based PCA that we describe in these results.
Table 4 shows the two-factor Varimax rotation for the 12 HHI-F items, consisting of items 1, 2, 3, 4, 5, 6, 7, 8, 9 and 10 for the first factor, and items 11 and 12 for the second factor.
Tableau 4. 2-component Varimax rotation for the 12 HHI-F items
Item
|
Component
|
|
1
|
2
|
1
|
,814
|
,102
|
2
|
,801
|
,013
|
3
|
,800
|
,048
|
4
|
,784
|
-,022
|
5
|
,742
|
,095
|
6
|
,724
|
,044
|
7
|
,696
|
,080
|
8
|
,646
|
,091
|
9
|
,638
|
,050
|
10
|
,606
|
,006
|
11
|
,023
|
,764
|
12
|
,077
|
,719
|
Also, we do not find the three components that made up the original version of the HHI, which distinguishes items 1, 2, 6 and 11, items 4, 7, 10 and 12, and items 3, 5, 8 and 9. It should be noted, however, that these three components have never been found identically in the various translations of the HHI that have been produced. In fact, during our review of the literature concerning the various translations and validations of the HHI. We note that 3 studies report only 1 factor, (Geiser et al., 2015; Ripamonti et al., 2012; Soleimani et al., 2019; Viana et al., 2010), 4 studies report 2 (Benzein & Berg, 2003; Van Gestel-Timmermans et al., 2010; Wahl et al., 2004; Yaghoobzadeh et al., 2019), 5 studies suggest 3 factors (Aslan et al., 2007; Balsanelli et al., 2010; Chan et al., 2012; Herth, 1992; Hirano et al., 2007; Mousa et al., 2017), and only one study suggests 4 factors (Arnau et al., 2010) (see Table 1).
3.3. Results: Sensitivity Analysis
Contrary to Herth (1992), we obtained a significant difference between the overall HHI-F score and gender (t(348) = 3.105, p < 0.05), age (Pearson correlation coefficient r = 0.146, p < 0.01), and the level of education followed (F(3, 346) = 3.312, p < 0.05).
We also analyzed the influence of other variables than those studied by Hertz (1992). Marital status (single, married or common-law, or divorced) was not significantly related to the HHI-F score ((F2, 347) = 2.019, p = .134). In contrast, there was a significant difference between those with and without children (t(348) = -2.960, p < 0.01), as was being a working student or not (t(348) = -3.362, p < 0.001).
3.4. Results: Concurrent Validation
In the literature, the measure of hope is often associated and correlated with other measures of psychological health, such as self-efficacy, optimism, anxiety, depression, problem solving (Delas et al., 2015). For this reason, the HHI will be compared to the following questionnaires: the Depression, Anxiety and Stress Scale (DASS-21) developed by Lovibond et Lovibond (1995) ), which assesses mental health status; the Impact of Event Scale-Revised (IES-R) developed by Weiss and Marmar (1997) which assesses the psychological impact of traumatic events; and the Brief Resilient Coping Scale (BRCS) developed by Sinclair and Wallston (2004), which measures coping strategies.
For each of these concurrent validations, we chose to retain the original three-dimensional structure of the HHI (temporality and future; preparation; positive expectations and interconnection). Indeed, even if we do not find the same component structure for the HHI-F as for the HHI, we consider that the validation of the HHI-F should be based on the same assumptions as the original Herth scale.
3.4.1 Correlation between thek HHI-F and the DASS-21scale
The DASS-21 (Lovibond & Lovibond, 1995), which assesses a patient's state of depression, anxiety and stress is a three-dimensional scale with 21 items in its current version. For each item, participants are asked to position themselves on a 4-point Likert scale: 0 = Did not apply to me at all; 1 = Applied to me to some degree, or some of the time; 2 = Applied to me to a considerable degree or a good part of time; 3 = Applied to me very much or most of the time. The scores for each dimension are obtained by adding the scores ticked by the respondents for the corresponding items. Items 3, 5, 10, 13, 16, 17 and 21 are dedicated to the measurement of depression, items 2, 5, 7, 9, 15, 19 and 20 to the measurement of anxiety, items 1, 6, 8, 11, 12, 14 and 18 to the measurement of stress. According to several studies, the DASS-21 has good psychometric characteristics, especially with regard to its factorial validity (Lovibond & Lovibond, 1995; Osman et al., 2012).
For the first concurrent validity study of the HHI-F, we measured the correlation between the 3 dimensions of the HHI-F and the global score, with the scores of the 3 dimensions of the DASS-21. The french version of the DASS-21 that we used is the one proposed by Ramasawmy, Hicks and Gilles (2010).
The results show that all correlations are highly significant (p < .001) for all points of comparison that can be established between the HHI-F and DASS-21 scales (see Table 5).
Table 5. Results of correlations between the 3 dimensions of the HHI-F and its overall score, and the scores on the 3 dimensions of the DASS-21
|
|
|
HHI-F
|
|
|
|
Global score
|
Temporality and the future
|
Preparation
|
Positive expectations and interconnection
|
DASS-21
|
Depression
|
Pearson correlation
|
-,562**
|
-,532**
|
-,551**
|
-,413**
|
Sig. (two-sided)
|
,000
|
,000
|
,000
|
,000
|
N
|
350
|
350
|
351
|
351
|
Anxiety
|
Pearson correlation
|
-,409**
|
-,385**
|
-,424**
|
-,285**
|
Sig. (two-sided)
|
,000
|
,000
|
,000
|
,000
|
N
|
350
|
350
|
351
|
351
|
Stress
|
Pearson correlation
|
-,405**
|
-,389**
|
-,406**
|
-,288**
|
Sig. (two-sided)
|
,000
|
,000
|
,000
|
,000
|
N
|
350
|
350
|
351
|
351
|
The results seem consistent, since they indicate systematically negative correlations between the HHI-F and the DASS-21. Indeed, the higher the DASS-21 scores, the more they indicate a deteriorated state of mental health measured through the perception of depression, anxiety and stress. Conversely, the higher the HHI scores, the higher the level of hope. It therefore seems consistent that when a patient reports depression, anxiety and stress, their levels of hope for the future, preparedness and expectations are low.
3.4.1 Correlation between the HHI-F and IES-R Scale
The Impact of Event Scale-Revised (IES-R) developed by Weiss and Marmar (1997) assesses the psychological impact of traumatic events. It consists of 22 items, on which respondents are asked to rate themselves on a 5-point Likert scale, ranging from 0 "Not at all", 1 "Somewhat", 2 "Moderately", 3 "Fairly" to 4 "Extremely". The IES-R distinguishes three dimensions: 1. reliving (items 1, 2, 3, 6, 9, 14, 15 and 20); 2. avoidance (items 5, 7, 8, 11, 12, 13, 17 and 22); 3. hyperactivation (items 4, 10, 15, 18, 19 and 21). The score for each dimension is calculated by summing the corresponding items. An overall IES-R score is also calculated by adding up all the items. The higher the score, the more severe the symptoms are considered to be.
For the second concurrent validity study of the HHI-F, we measured the correlation between the 3 dimensions of the HHI-F and the global score, with the scores of the 3 dimensions of the IES-R and the global score. The French version of the IES-R that we used is the one proposed by Chiasson et al. (2018).
The results show significant correlations for all points of comparison that can be established between the HHI-F and IES-R scales (see Table 6).
Table 6. Results of the correlations between the 3 dimensions of the HHI-F and its overall score, and the scores of the 3 dimensions of the IES-R and its overall score.
|
|
HHI-F
|
|
|
Global Score
|
Temporality and the future
|
Preparation
|
Positive expectations and interconnection
|
IES-R
|
Global Score
|
Pearson correlation
|
-,301**
|
-,307**
|
-,307**
|
-,189**
|
Sig. (two-sided)
|
,000
|
,000
|
,000
|
,000
|
N
|
350
|
350
|
351
|
351
|
Reliving
|
Pearson correlation
|
-,267**
|
-,279**
|
-,276**
|
-,157**
|
Sig. (two-sided)
|
,000
|
,000
|
,000
|
,003
|
N
|
350
|
350
|
351
|
351
|
Avoidence
|
Pearson correlation
|
-,218**
|
-,237**
|
-,231**
|
-,121*
|
Sig. (two-sided)
|
,000
|
,000
|
,000
|
,023
|
N
|
350
|
350
|
351
|
351
|
Hyperactivation
|
Pearson correlation
|
-,371**
|
-,352**
|
-,356**
|
-,277**
|
Sig. (two-sided)
|
,000
|
,000
|
,000
|
,000
|
N
|
350
|
350
|
351
|
351
|
As in the concurrent validation with the DASS-21, all correlations between the IES-R and the HHI-F are negative. This can be explained by the fact that the more the psychological impact of the event is perceived as high by the patient and therefore traumatic, the more his hopes for the future, preparation and positive expectations decrease. These results are consistent.
3.4.2 Correlation between the HHI-F and the BRCS Scales
The Brief Resilient Coping Scale (BRCS) (Sinclair & Wallston, 2004) measures coping strategies. It consists of 4 items on which the patient is asked to position himself on a 5-point Likert scale, ranging from 1: "Does not describe me at all" to 5: "Describes me completely". The overall BRCS score is calculated by adding the scores of the 4 items. The higher the score, the greater the patient's resilience.
For the third concurrent validity study of the HHI-F, we measured the correlation between the three dimensions of the HHI-F and the global score, with the BRCS score. The French version of the BRCS that we used is the one proposed by Ionescu (2011).
The results show significant correlations for all points of comparison that can be established between the HHI-F and BRCS scales (see. Table 7).
Table 7. Results of the correlations between the 3 dimensions of the HHI-F and its overall score, and the BRCS score
|
HHI-F
|
|
Global Score
|
Temporality and the future
|
Preparation
|
Positive expectations and interconnection
|
BRCS
|
Pearson correlation
|
,552**
|
,453**
|
,540**
|
,479**
|
Sig. (two-sided)
|
,000
|
,000
|
,000
|
,000
|
N
|
350
|
350
|
351
|
351
|
The positive correlations attest to the validity of the HHI-F scale. Indeed, these results mean that the more coping strategies are developed in patients, the greater their sense of hope, which is a consistent link between these two psychological health factors.