If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Corresponding author at: University of Valencia, Department of Personality, Assessment and Psychological Treatments, Av. Blasco Ibáñez, 21, València 46010, Spain.
An original scale with adequate psychometric properties (W-COV) was constructed.
•
It assesses worries about health, economic and psychosocial consequences of COVID.
•
We used a cross-cultural design with a large sample from 5 Spanish-speaking countries.
•
W-COV was validated in adolescents and young adults aged between 12 and 25 years old.
•
Pandemic-related worries negatively affect life satisfaction of young Ibero-Americans.
Abstract
The pandemic context presents remarkable psychological challenges for adolescents and young adults. The aim of the present work was to construct and study the psychometric properties of a scale in Spanish language (W-COV) to measure their worries related to the pandemic. Participants were 5559 people aged between 14 and 25 years old (M = 19.05; SD = 3.28). Self-report data were collected using a cross-sectional and cross-cultural design. Participants were from 5 Spanish-speaking countries. Instruments were W-COV to assess worries about COVID-19 and its consequences; DASS-21 for anxiety, depression and stress; and SWLS for life satisfaction. Exploratory, confirmatory and multi-group factor analyses were conducted to determine the factorial structure of the W-COV and its measurement invariance (configural, metric, scalar and error variance). Correlational and regression analyses were also performed to study convergent and predictive validity. The results suggest that W-COV presents a bifactorial structure: (1) a general factor of worries about COVID-19; and (2) three different factors: worries about health, economic and psychosocial consequences from COVID-19. The internal reliability indices Cronbach's α and Omega were adequate. With respect to the invariance results, the instrument can be used interchangeably in the five countries considered, in both genders and in two different age groups (12–17 and 18–25). Regarding validity, W-COV factors were positively associated with anxiety, depression and stress, and negatively predicted life satisfaction. In conclusion, W-COV is a reliable and valid instrument for researchers and health care professionals to assess the psychological impact of the pandemic on mental health of young Ibero-Americans.
The coronavirus disease (COVID-19) outbreak has had a profound impact worldwide since the World Health Organization (WHO) announced its pandemic status on March 11, 2020 (
). Particularly in Spain, there have been 219 thousand confirmed cases and 25.600 deaths due to COVID-19 at the end of April 2020, being one of the most affected countries in the world by the first wave of the virus (
). Latin America has also been heavily hit by the virus, with severe consequences such as the collapse of sanitary systems and notable economic recession (
). In fact, the European Union remarked not only the need to reduce the number of COVID-19 infected cases, but also the importance of dampen as much as possible the strong impact of the pandemic on mental health (
). Multiple sources of stress have emerged from the pandemic, such as the immediate health threat, job or academic uncertainty, economic concerns, and changes in social routines (
). Consequently, recent studies have highlighted the negative impact of the pandemic on people's mental health, informing of increased levels of anxiety, depression, insomnia, irritability, and alcohol consumption, among others (
However, it seems that COVID-19 does not impact the mental health of all people equally, since not everyone is equally concerned about the pandemic. Many people manifest worries about the immediate and long-term consequences of the virus (
). These worries range from getting sick or dying from coronavirus (both oneself and family members) to the potential consequences in sanitary, economic, and psychosocial areas. Although it is natural to be concerned about this critical situation, if such worries become disproportionate in intensity, frequency, or duration, they could interfere with personal functioning, affecting one's emotional state, problem-solving skills, or goal-driven behaviors (
). For instance, excessive worry about one's health can lead to inappropriate health care seeking behavior that may add further pressure to the already burdened health care system (
How health anxiety influences responses to viral outbreaks like COVID-19: What all decision-makers, health authorities, and health care professionals need to know.
Although research on COVID worries has developed strongly in a short time, there are significant gaps in this area. Most of the research has been conducted in general adult population above 18 years old (e. g.
Diabetes and COVID-19: Psychosocial consequences of the COVID-19 pandemic in people with diabetes in Denmark—what characterizes people with high levels of COVID-19-related worries?.
Can measures of sleep quality or white matter structural integrity predict level of worry or rumination in adolescents facing stressful situations? Lessons from the COVID-19 pandemic.
). According to the literature, older people are more capable of reorienting their emotions and cope with the stress caused by the disease, while young people seem to be especially susceptible to the development of anxiety and depression because of concern about COVID (
). Worries related with the possible consequences of a new disease, as well as the limitations of social interactions, particularly influenced young people lives (
, as a result of the pandemic a considerable number of young people have experienced symptoms of depression, anxiety and distress, post-traumatic stress symptoms, obsessive-compulsive symptoms, eating problems, changes in sleep patterns and substance abuse. Considering that youth is a time when the consolidation of mental disorders can take place, the attention to these sub-clinical symptoms becomes especially important for their prevention (
A relevant question arises from the review of previous research: how are worries about COVID-19 measured in these studies? Indeed, the majority of previous investigations used ad hoc questions (e.g.,
However, some studies have developed their own questionnaires to assess COVID-19 worries in populations with specific characteristics, such as Type 1 Diabetes (
Diabetes and COVID-19: Psychosocial consequences of the COVID-19 pandemic in people with diabetes in Denmark—what characterizes people with high levels of COVID-19-related worries?.
), which assesses the level of concerns regarding being infected or infecting others by COVID-19. The validity of this questionnaire has been replicated in Bangladeshi population and shows adequate psychometric properties (
). However, this instrument measures worries about the disease itself, it does not assess worries about economic or psychosocial consequences of the pandemic.
There are no studies that provide information about people's worries about COVID-19 of the Ibero-American population. Research on COVID-19 related worries has been conducted in North American population (United States and Canada), European population (Italy, Denmark, Sweden), Asian population (Singapore, Bangladesh, Israel), and Australian population. Only one cross-cultural study has been identified (
), carried out on health workers from different countries over the world, although it does not include neither Spanish nor Latino populations.
Therefore, the aim of the present study was to construct and validate an instrument to assess worries about COVID-19 and its sanitary, economic, and psychosocial consequences in adolescent and young population of Ibero-America using a cross-cultural approach. For this purpose, the instrument was tested first in Chilean population, given the large sample size, and its psychometric properties were validated for Colombian, Ecuadorian, Mexican and Spanish population, testing also wheatear the factor structure was equivalent for both men and women across different age groups.
Method
Participants
Regarding total sample, participants were 5559 adolescents and young adults (83.2 % female; 15.6 % male, 1.2 %; non-binary gender) aged between 12 and 25 years (M = 19.05; SD = 3.28). The distribution of participants across countries was as following: 66.5 % of the participants were from Chile, 11.2 % from Colombia, 8.4 % from Ecuador, 5 % from Mexico and 5.6 % from Spain. Table 1 shows sample characteristics for each country. The study included a convenience sample, using a snowball-sampling technique.
Table 1Sample information for each country.
Country
N
Female (%)
Male (%)
Non-binary (%)
Age M (SD)
Chile
3699
87.4 %
10.9 %
1.6 %
19.42 (3.26)
Colombia
621
82.8 %
17.1 %
0.2 %
19.93 (2.69)
Ecuador
558
69.4 %
30.1 %
0.5 %
17.68 (2.74)
México
334
69.8 %
30.2 %
0 %
17.31 (3.69)
Spain
347
74.4 %
25.4 %
0.3 %
17.38(3.11)
The table shows frequencies of participants from each country, their gender and age (mean and standard deviation).
Worries about COVID-19 and its consequences were assessed using the Worries about COVID-19 and its consequences Scale (W-COV), developed in an original Spanish form (Online Resource 1). The scale consists of 16 items with 5-points Likert scale (1 = Almost never; 5 = Almost always). This instrument is composed by three factors: (1) Health worries: worries about getting sick or dying from COVID (both oneself and/or a family member), quality health care, using hygienic measures properly and lack of information about the virus (5 items; e. g. “I worry about getting sick or dying from COVID-19”); (2) Economic worries: worries about the economic situation due to COVID-19, not being able to meet basic needs, having product shortages, losing one's job or having to drop out of school (5 items; e. g. “I am worried that COVID-19 will harm my or my family's economic situation”); (3) Psychosocial worries: worries about the situation of confinement, uncertainty about the consequences of COVID-19, worries about the impact of the pandemic on social relationships, psychological state, and academic/work performance (6 items; e. g. “I am worried about my psychological state –anxiety, insomnia, irritability, sadness–”).
Psychological distress
Psychological distress was assessed by the short-form version of the Depression Anxiety Stress Scales (DASS-21;
). The scale consists of 21 items with a 4-points Likert scale assessing the symptoms during the last week. This instrument is composed of three factors: depression; anxiety; and stress. The internal consistency of the Spanish version of the instrument is good (α = 0.84 for depression; α = 0.70 for anxiety; and α = 0.82 for stress).
) was used to assess the individual's overall evaluation of their own life. The 5-item scale, with seven response alternatives, consists of one global dimension calculated by the sum of all items. The internal consistency of the Spanish version of this scale is good (α = 0.88).
Procedure
The study was carried out following the guidelines of the Helsinki Declaration and the Council of Europe Convention on Human Rights (
The first step of this research consisted of developing the instrument to assess a wide range of public worries about COVID-19 and its consequences on people's life. This instrument was developed according to an expert group formed by four psychologists, who proposed possible areas of concern for adolescents and young adults. From the initial proposal, 16 items were selected to be part of the scale. All items were positively formulated in order to facilitate the easy understanding of the scale.
Data were collected through an online survey on the free software Limesurvey. Following the Checklist for Reporting Results of Internet E-Surveys (CHERRIES;
), we can report some of the most relevant aspects of the online survey used. First, the survey was approved by the university's ethics committee. Data were collected from May to August 2020, during the first wave of COVID. Before answering the survey, participants were informed about the objective of the study and consented their participation. The survey did not ask about personal data. This was an open survey, available to all those who wished to answer voluntarily, so the sample used was one of convenience. The survey was disseminated through the website and social networks of the research teams, as well as among educational centers which had previous contact with researchers. There was no incentive for participation. Participants could access the survey from their own mobile or computer devices from home, and the time expected for complete the survey was approximately 15 min.
Data analysis
First, we conducted an Exploratory Factor Analysis (EFA) to determine the factorial structure of the W-COV. FACTOR software (version 10.5.01) was used to perform the EFA (
). The main advantage of FACTOR in comparison to other statistical software is the possibility of performing the EFA on the basis of the tetrachoric/polychoric correlation matrix; this option is preferable when modeling dichotomous and/or ordinal data (such as in the case of the W-COV;
). To determine the number of factors to be retained during the EFA, we used Parallel Analysis (PA). To verify the applicability of the EFA to the W-COV, we estimated the Kaiser–Meyer–Olkin (KMO) index and the Bartlett sphericity test. According to these indices, an EFA is applicable when the KMO is >0.80 and the significance of the Barlett's test of sphericity is p < .05. Once confirmed its applicability, factors were extracted through Robust Unweighted Least Squares (RULS) and applying an oblique rotation (direct oblimin). All these analyses were conducted on the basis of the polychoric correlation matrix. Due to the large sample size, the EFA was conducted using the Chilean sample. To do so, the Chilean sample (n = 3699) was divided into two random samples equally distributed in number of participants: the first Chilean subsample (validation sample, n = 1849) was used to perform the EFA, whereas the second subsample (confirmation sample, n = 1840) was used to confirm the initial factorial structure through CFA.
To examine the adequacy of fit of the factor model derived from the EFA, confirmatory factor analyses (CFAs) were then conducted for each country, as well as for the overall sample. These analyses were conducted through EQS 6.4 (
). As factorial models derived from an EFA admit different factorial structures (e.g., a three-factor model derived from an EFA may be structured as three correlated factors, three factors under a second order factor, or both [bifactor]), we compared the fit of different factorial solutions. Non-normal distribution of categorical data was addressed by applying robust estimation methods (robust Maximum Likelihood, MLR;
) and to ensure the comparability between countries, we did not apply any modification to the models based on modification indices, even when minor changes (e.g., correlations between error terms) significantly increased the models' fit. Goodness of fit for the CFA models was assessed through the following indices: the root mean square error of approximation (RMSEA), the comparative and incremental fit indices (CFI and IFI, respectively), and the standardized root mean square residual (SRMR). An excellent model fit was identified when the CFI and the IFI were ≥0.95, the RMSEA ≤0.05, and the SRMR ≤0.05 (
). For the sake of transparency, Satorra-Bentler chi-square (X2), general model significance (p), and relative chi-square (X2/df) were reported; however, given that X2 is highly sensitive to sample size (
The golden rule is that there are no golden rules: A commentary on Paul Barrett’s recommendations for reporting model fit in structural equation modelling.
), these indices were not employed to assess the adequacy of the CFA models.
To assess whether the factor structure of the W-COV were valid for their use across different ages (adolescents vs. young adults), countries (Spain, Chile, Mexico, Ecuador, and Colombia), and in both genders (male vs. females), multi-group CFAs were conducted. Specifically, we tested four levels of measurement invariance: (1) configural (test whether items load on the same factor across groups), (2) metric (test whether item factorial loadings are equal across groups), (3) scalar (test whether item intercepts are equal across groups) and (4) error variance invariance (test whether items measurement error are equal across groups). The adequacy of the increasingly constrained models was assessed through the difference between pairs of nested models (∆) in the RMSEA, CFI and SRMR. A change ≥0.01 in the CFI, ≥0.015 in the RMSEA, and ≥0.03 in the SRMR indicates a significant decrease in the model fit when testing for measurement invariance (
Finally, the general mean score of the items, subscales, and total score of the W-COV were explored, as well as the magnitude of the differences according to the age, gender, and country. Different reliability indices were calculated; in particular, we estimated the Cronbach's α and Omega (Ω). Convergent validity was explored by correlating (Pearson's r) the W-COV subscales with other related measures (DASS-21). Finally, multiple regressions analysis was performed to determine the predictive power of the W-COV over life satisfaction (SWLS).
Results
Exploratory Factor Analysis (EFA)
The EFA was conducted using the Chilean validation sample (n = 1849). To verify the applicability of the EFA to the W-COV, the Kaiser–Meyer–Olkin index (KMO = 0.855) and the Bartlett's sphericity test (χ2120 = 7520.5, p ≤ .001) were calculated. Both indices indicated that the W-COV admitted a factorial solution. PA suggested to retain three factors (factor eigenvalues >1.34). The factorial solution derived from the RULS factor analysis revealed that this three-factor structure explained 53 % of the total variance of the W-COV. Table 2 includes the standardized factor loadings obtained in the EFA.
Table 2W-COV factorial loadings.
EFA on the Chilean subsample (Validation subsample, n = 1849)
Factorial loadings from the AFC included in this table correspond to those obtained in the bifactor model (in particular, to factorial loadings of each item on a specific factor). As the bifactor model also includes a relationship between each item and a general dimension (i.e., items' variance is shared by a specific COVID-worry factor and a general COVID-worry dimension), these figures are not comparable to those obtained in the EFA.
These figures correspond to the CFA conducted in the Chilean confirmation subsample.
(n = 1850)
Colombia (n = 621)
Ecuador (n = 558)
Mexico (n = 334)
Spain (n = 347)
Factor 1: health worries
Item 1
0.80
0.72
0.76
0.43
0.58
0.80
Item 2
0.76
0.46
0.54
0.40
0.55
0.54
Item 3
0.38
0.17
0.28
0.26
0.28
0.25
Item 4
0.32
0.25
0.25
0.25
0.18
0.27
Item 13
0.28
0.08
0.09
0.02
0.11
0.09
Factor 2: economic worries
Item 5
0.89
0.64
0.70
0.58
0.50
0.95
Item 6
0.82
0.61
0.47
0.36
0.56
0.29
Item 7
0.31
0.02
0.13
0.06
0.06
0.02
Item 8
0.43
0.13
0.12
0.21
0.18
0.26
Item 9
0.52
0.24
0.38
0.24
0.18
0.13
Factor 3: psychosocial worries
Item 10
0.63
0.49
0.50
0.22
0.27
0.39
Item 11
0.69
0.63
0.62
0.72
0.58
0.66
Item 12
0.69
0.53
0.45
0.39
0.44
0.45
Item 14
0.53
0.32
0.30
0.20
0.25
0.05
Item 15
0.64
0.40
0.55
0.34
0.48
0.28
Item 16
0.42
0.26
0.28
0.19
0.31
0.18
The table shows the factor loadings of W-COV items in both exploratory (performed only with Chilean sample) and confirmatory analysis (with five different countries). Items were presented according the three main factors of the scale (worries about health, economic and psychosocial consequences of COVID-19).
a Factorial loadings from the AFC included in this table correspond to those obtained in the bifactor model (in particular, to factorial loadings of each item on a specific factor). As the bifactor model also includes a relationship between each item and a general dimension (i.e., items' variance is shared by a specific COVID-worry factor and a general COVID-worry dimension), these figures are not comparable to those obtained in the EFA.
b These figures correspond to the CFA conducted in the Chilean confirmation subsample.
Factor I (“Health worries”), which was composed of 5 items (items 1, 2, 3, 4, and 13) and had an eigenvalue of 5.37, explained a 33.56 % of the W-COV variance. Items within this factor reflect fears related to the health consequences derived from the COVID pandemic. Factor II, called “Economic Worries”, comprised 5 items (items 5 to 9) assessing worries related to the economic impact of the COVID pandemic (eigenvalue = 1.78; explained variance of 11.13 %). Factor III (“Psychosocial Worries”) integrated the other 6 items (items 10 to 16) and had an eigenvalue of 1.34 and an explained variance of 8.36 %. This factor comprises questions about the social and psychological impact of the pandemic.
Confirmatory Factorial Analysis (CFA) and measurement invariance
To determine the most suitable factorial structure behind the three-factor solution, we performed a CFA comparing three different models: (a) a model in which the three first-order factors derived from the EFA were correlated; (b) a model in which the three first-order factors were grouped under a second order factor (i.e., an higher order factor depicting a general COVID-worries dimension); and (3) a bifactor model comprising a non-hierarchical COVID-worries dimension in addition to the three specific factors. The adequacy of the proposed models was tested using the Chilean confirmation sample (n = 1850). Table 3 shows goodness-of-fit indices for the different models.
Table 3Results from the CFA.
n
χ2
df
p
χ2/df
RMSEA (CI)
CFI
IFI
SRMR
Step 1: comparison of different factorial solutions (Chilean confirmation sample)
Three correlated 1st order factors
1850
1050.47
101
<.001
10.40
0.071 (0.067;0.075)
0.861
0.862
0.062
Three 1st order factors under a 2nd order factor
1850
1028.62
100
<.001
10.28
0.071 (0.067;0.075)
0.864
0.865
0.062
Bifactor model
1850
646.21
88
<.001
5.27
0.059 (0.054;0.063)
0.918
0.919
0.041
Step 2: goodness of fit of the bifactor model in the different study samples
Colombia
621
300.99
88
<.001
3.42
0.062 (0.055;0.070)
0.921
0.921
0.047
Ecuador
558
191.32
88
<.001
2.17
0.046 (0.037;0.055)
0.952
0.952
0.037
Mexico
334
233.65
88
<.001
2.65
0.071 (0.059;0.081)
0.910
0.912
0.053
Spain
347
233.18
88
<.001
2.64
0.069 (0.058;0.080)
0.892
0.894
0.056
All countries
5559
1767.62
88
<.001
20.08
0.059 (0.056;0.061)
0.925
0.925
0.037
Note. CFA = confirmatory factor analysis; χ2 = Satorra-Bentler chi-square; df = degrees of freedom; p = general model significance; χ2/df = normed chi-square; RMSEA = root mean square error of approximation; CFI = comparative fit index; IFI = incremental fit index; SRMR = standardized root mean square residual.
The table shows, in the first place (Step 1), the fit indices obtained for the three factorial models tested in the Chilean sample. Being the bifactorial model the one with the best fit, secondly (Step 2), the fit indices of the bifactorial model are compared for the other countries.
The factorial solution with the most satisfactory fit indices was the bifactor model. In this model, the RMSEA was <0.08 (figure required to consider a model acceptable), the CFI and the IFI reached a value of 0.918 and 0.919 respectively (close to the cut-off point established to consider that a model reaches an excellent fit), and the SRMR was 0.041 (below the 0.05 value required by the strictest criteria to consider a model parsimonious). The resulting bifactorial model is depicted in the Fig. 1.
Fig. 1Bifactor model for the W-COV. R2 is expressed as a percentage outside the main endogenous variables' boxes. Coefficients are reported in standardized format. All parameters were significant at p < .001. Error terms are not included in order to facilitate interpretation.
Then, we verified the applicability of the bifactorial solution to samples from the other countries. As displayed in Table 3, the RMSEA and the SRMR were below the thresholds of 0.08 and 0.10 in all the country-based datasets as well as in the whole sample. Furthermore, in all the datasets, the CFI and the IFI were above –or close to– the threshold of 0.90 to consider a model parsimonious. The best adjustment according to these indices was obtained for the sample from Ecuador (RMSEA = 0.046; SRMR = 0.037; CFI = 0.952; IFI = 0.952).
Finally, to test measurement invariance of the W-COV according to gender, age, and country, we conducted a series of multi-group CFAs. As displayed in Table 4, gender, age, and country configural invariance of the W-COV was fully supported (RMSEA = 0.058; CFI = 0.925; SRMR = 0.038 [according to gender]; RMSEA = 0.061; CFI = 0.920; SRMR = 0.039 [according to age]; RMSEA = 0.060; ∆CFI = 0.920; ∆SRMR = 0.048 [according to country]), so we subsequently estimated models with increasing levels of constraints to test higher levels of invariance. Regarding metric invariance, changes in the RMSEA, CFI, and SRMR did not show a significant worsening in the model fit neither for gender (∆RMSEA = 0.003; ∆CFI = 0.004; ∆SRMR = 0.010) nor for age (∆RMSEA = 0.005; ∆CFI = 0.006; ∆SRMR = 0.006), or country invariance (∆RMSEA = 0.003; ∆CFI = 0.011; ∆SRMR = 0.015). Similarly, the models' fit did not significantly decrease when subsequent levels of age invariance were tested (∆ in CFI, RMSEA, and SRMR were always below 0.010, 0.015 and 0.03, respectively), thus supporting a complete equivalence of the W-COV in adolescents and young adults. Scalar invariance according to gender was also fully supported (∆RMSEA = 0.005; ∆CFI = 0.006; ∆SRMR = 0.024). However, the significant ∆ in SRMR when scalar invariance according to country was tested (0.045) or when error invariance according to gender and country was tested (0.032 and 0.056 respectively) suggested the presence of differences at these levels of measurement according to gender and country.
Table 4Multigroup CFAs according to gender, age, and country.
χ2
df
p
χ2/df
RMSEA (CI)
CFI
IFI
SRMR
Comparisons
∆RMSEA
∆CFI
∆SRMR
Measurement invariance according to gender
Configural invariance
1802.89
176
<.001
10.24
0.058 (0.056;0.060)
0.925
0.925
0.038
NA
Metric invariance
1920.27
204
<.001
9.41
0.055 (0.053;0.058)
0.921
0.921
0.048
Conf. Vs Metric
0.003
0.004
0.010
Scalar invariance
2392.59
220
<.001
10.87
0.060 (0.058;0.062)
0.926
0.927
0.072
Metric Vs Scalar
0.005
0.006
0.024
Error variance invariance
1797.35
192
<.001
9.35
0.055 (0.053;0.058)
0.923
0.923
0.040
Scalar Vs Error
0.005
0.003
0.032
Measurement invariance according to age
Configural invariance
1997.06
176
<.001
11.34
0.061 (0.059;0.063)
0.920
0.921
0.039
NA
Metric invariance
2154.94
204
<.001
10.56
0.059 (0.056;0.061)
0.915
0.915
0.045
Conf. Vs Metric
0.005
0.006
0.006
Scalar invariance
2585.25
220
<.001
11.75
0.059 (0.057; 0.062)
0.916
0.916
0.046
Metric Vs Scalar
0.000
0.001
0.001
Error variance invariance
2220.93
192
<.001
11.56
0.062 (0.059; 0.064)
0.912
0.912
0.043
Scalar Vs Error
0.003
0.004
0.003
Measurement invariance according to country
Configural invariance
1614.49
440
<.001
3.66
0.060 (0.057; 0.063)
0.920
0.921
0.048
NA
Metric invariance
1878.41
552
<.001
3.40
0.057 (0.054; 0.060)
0.910
0.910
0.063
Conf. Vs Metric
0.003
0.011
0.015
Scalar invariance
3631.68
616
<.001
5.89
0.062 (0.059; 0.064)
0.918
0.918
0.108
Metric Vs Scalar
0.005
0.005
0.045
Error variance invariance
1791.64
504
<.001
3.55
0.059 (0.056; 0.062)
0.916
0.917
0.052
Scalar Vs Error
0.003
0.002
0.056
Note. CFA = confirmatory factor analysis; χ2 = Satorra-Bentler chi-square; df = degrees of freedom; p = general model significance; χ2/df = normed chi-square; RMSEA = root mean square error of approximation; CFI = comparative fit index; IFI = incremental fit index; SRMR = standardized root mean square residual; ∆ RMSEA = change in RMSEA compared with the previous model (expressed in absolute values); ∆ CFI = change in CFI compared with the previous model (expressed in absolute values); ∆ SRMR = change in SRMR compared with the previous model (expressed in absolute values).
The table shows the four levels tested of measuring invariance (configural, metric, scalar and error variance) to observe if model fit indices remain similar in participants of different gender, country or age group (adolescents aged 12–17 or young adults aged 18–25). For ease of comparison, the increase in the indexes is also showed.
Reliability analysis showed an adequate internal consistency for all three dimensions in the five countries studied (Table 5). The instrument's Cronbach's alpha did not increase with the removal of any of the items, which indicates that all of them contribute positively to the internal consistency of the scale.
Table 5Descriptive statistics, internal consistency indexes and test-retest correlations.
M (SD)
Asymmetry
Kurtosis
α
ω
Entire sample
Women
Men
Chile
(N = 3699)
(n = 3234)
(n = 404)
0.83
0.83
Health-W
3.32 (0.79)
3.36 (0.78)
3.06 (0.86)
−0.31
−0.21
0.689
0.686
Economic-W
3.24 (0.89)
3.27 (0.88)
2.99 (0.91)
−0.05
−0.63
0.743
0.748
Psychosocial-W
3.52 (0.86)
3.55 (0.84)
3.24 (0.96)
−0.37
−0.39
0.739
0.744
Colombia
(N = 621)
(n = 514)
(n = 106)
0.85
0.85
Health-W
2.95 (0.82)
3.00 (0.81)
2.69 (0.82)
−0.01
−0.23
0.713
0.709
Economic-W
3.10 (0.89)
3.12 (0.89)
2.97 (0.88)
0.10
−0.61
0.776
0.795
Psychosocial-W
3.07 (0.91)
3.15 (0.90)
2.67 (0.85)
0.03
−0.60
0.769
0.774
Ecuador
(N = 558)
(n = 387)
(n = 168)
0.86
0.86
Health-W
3.11 (0.86)
3.23 (0.87)
2.83 (0.79)
0.02
−0.42
0.704
0.705
Economic-W
3.33 (0.92)
3.49 (0.86)
2.94 (0.93)
−0.09
−0.73
0.771
0.782
Psychosocial-W
3.01 (0.88)
3.14 (0.86)
2.72 (0.87)
−0.05
−0.65
0.736
0.740
Mexico
(N = 334)
(n = 233)
(n = 101)
0.88
0.88
Health-W
3.13 (0.91)
3.28 (0.86)
2.81 (0.93)
−0.08
−0.48
0.758
0.751
Economic-W
3.07 (0.95)
3.16 (0.88)
2.86 (1.08)
0.16
−0.74
0.797
0.789
Psychosocial-W
3.00 (0.94)
3.11 (0.92)
2.71 (0.97)
0.07
−0.77
0.782
0.769
Spain
(N = 347)
(n = 258)
(n = 88)
0.84
0.84
Health-W
2.83 (0.82)
2.96 (0.81)
2.46 (0.84)
0.14
−0.50
0.714
0.712
Economic-W
2.52 (0.78)
2.63 (0.76)
2.21 (0.74)
0.46
0.02
0.676
0.684
Psychosocial-W
2.80(0.88)
2.97 (0.85)
2.32 (0.80)
0.20
−0.58
0.748
0.756
Note. 61 people identified themselves with the gender “Other” (non-binary) in Chile, 1 in Colombia, 3 in Ecuador and 1 in Spain. ** p < .001.
The table shows descriptive data of the participants from the five countries. In addition, two internal consistency indices (α and ω) of the W-COV factors are observed for each country.
Levels of worries about COVID-19 and their consequences are similar among adolescents (aged 12–17 years) and the young adults (aged 18–25 years) (Fig. 1). In general, worries are moderate-high, scoring above 3 on a scale ranged from 1 to 5.
Validity analyses
Convergent validity was studied, comparing the square root of the AVE from W-COV dimensions (health worries, economic worries, and psychosocial worries) with values from correlations between pairs of factors, indicating adequate indexes. In addition, the trifactorial structure of the scale shows that factor loads were high and significant among cross-cultural samples (Table 2); that is, scale factors strongly correlate with the latent variable to be evaluated, worries about COVID-19 and its consequences. Furthermore, Pearson's correlations were conducted to compare the W-COV with another instrument (DASS-21) measuring emotional distress (symptoms of depression, anxiety, and stress).
The results showed that high levels of health, economic and psychosocial worries are significantly and positively correlated with symptoms of depression (r = 0.22, r = 30, r = 0.51, p < .001), anxiety (r = 0.30, r = 30, r = 0.48, p < .001) and stress (r = 0.29, r = 31, r = 0.56, p < .001), showing the most strongly correlations with psychosocial worries.
Multiple regression was conducted to analyse predictive validity. The results indicated that all dimensions of W-COV were significant predictors of life satisfaction explaining 11 % of its variance (R2 = 0.11). Health worries were positively associated with life satisfaction (ß = 0.10, p < .001), while economic (ß = −0.16, p < .001) and psychosocial worries (ß = −0.28, p < .001) were negatively related to life satisfaction.
Discussion
Since the coronavirus disease (COVID-19) has been declared a global pandemic by the WHO in March 2020, it has become one of the most important health problems all around the world. It is hard to predict the course that this current health crisis is going to take and to what extend it will affect young people's future well-being and health. Therefore, reliable and valid assessment tools are needed to investigate the effects of COVID-19 on youth's mental health across different Ibero-American countries, allowing a valuable insight into the perception of the coronavirus and its consequences in such a vulnerable population.
The aim of the present study was to examine the psychometric properties of the Worries about COVID-19 and its consequences Scale (W-COV) by analysing its factor structure and measurement invariance in adolescents and young people across five countries, including Chile, Colombia, Ecuador, Mexico and Spain, testing also wheatear the factor structure was equivalent for both men and women across different age groups. The proposed three-factor structure of the W-COV with factors assessing health, economic, and psychosocial consequences, was confirmed across all five Ibero-American countries. These results indicate that the W-COV measures worries about COVID-19 and its consequences in the same way across the five countries. The scale also has shown good internal consistency, with reliability indexes similar across the cross-cultural samples. In addition, overall measurement invariance according to gender and across age groups was supported for both, thus implying that, male and female adolescents and young adults interpreted the W-COV items in a conceptually similar manner.
Furthermore, the results indicate that the W-COV meets the criteria for convergent and predictive validity: 1) W-COV was positively related to symptoms of depression, anxiety and stress; 2) predicted subjective well-being. The convergent validity was analysed to provide evidence of construct validity. First, the 16 items of the W-COV correlate significantly and highly with the latent variable they intend to evaluate. The intercorrelations between the three factors of the W-COV have adequate values. Second, the results confirm the relationship between the emotional distress (DASS-21) and COVID-19 related worries (W-COV), demonstrating that the dimensions of both scales measure similar but different concepts. These results provide evidence that W-COV is a valid measure of worries about COVID-19 and its consequences.
The predictive validity of the instrument was determined by performing multiple regression analysis between COVID-19 related worries (W-COV) and life satisfaction, an indicator of subjective well-being. All dimensions of the W-COV are significant predictors of life satisfaction, with psychosocial worries being the dimension with the strongest association. Although health worries were positively related to life satisfaction, the relationship with economic and psychosocial worries was negative. These findings are consistent with a recent study that suggests a negative impact of fear of COVID-19 on life satisfaction (
Our study has some limitations worth noting. First, the findings of this study were based on self-report data through online measure, which makes the data collection process more accessible and widespread, but also increase the risk of invalid data and source bias. Future research might use a multimethod approach (combining qualitative and quantitative methods) and/or multirater (parents and peers) designs to improve the validity of our findings. Second, although more than one reliability analyses were carried out, the test-retest reliability of the scale cannot be examined given the cross-sectional design.
Therefore, future studies might consider a longitudinal design in order to gather information regarding the temporal reliability of the scale. Third, another problematic issue of the study concerns the convenience sampling of participants, which resulted in a disproportional share of female participants, thus limiting generalizability of our results. Samples from some of the countries were relatively small and should be enhanced in future studies, both in terms of their size and heterogeneity, for more robust conclusions. Finally, although our results indicate that the W-COV scores can be meaningfully compared across Ibero-American countries, only one European country was included in the present study. Future studies should examine measurement invariance of the W-COV in other European samples as well.
Conclusions
Overall, our findings indicate that the Worries about COVID-19 and its consequences Scale (W-COV) is a valid and reliable self-report measurement of how adolescents and young adults from different Ibero-American countries perceive the coronavirus and its consequences. Thus, this novel questionnaire can serve as a suitable tool for researchers and health care professionals to assess the psychological impacts of COVID-19 on especially vulnerable individuals during adolescence and youth. Additionally, this cross-cultural approach allows the comparison between results from studies conducted in different cultures and countries, expanding scientific knowledge internationally.
Funding
This research was supported by grants from the Spanish Ministry of Science, Innovation and Universities ( PSI2017-84005-R ) and the European Regional Development Fund (FEDER) from the European Union.
Data availability statement
The datasets generated and analysed during the current study are not publicly available due to the fact that they constitute an excerpt of research in progress but are available from the corresponding author on reasonable request.
CRediT authorship contribution statement
All the authors equally contributed to the study design, participants' recruitment, data collection, analysis/interpretation of data, and writing up the study and the paper. All the authors read and approved the final version of the manuscript.
Informed consent
The Ethics Committee of Research in Humans of the Ethics Commission in Experimental Research of the University of Valencia approved study consent, procedures and research protocol (Reference number: 1595575567385, July 23th 2020). Before accessing the survey, participants gave their informed consent to participate in the study.
How health anxiety influences responses to viral outbreaks like COVID-19: What all decision-makers, health authorities, and health care professionals need to know.
Can measures of sleep quality or white matter structural integrity predict level of worry or rumination in adolescents facing stressful situations? Lessons from the COVID-19 pandemic.
Diabetes and COVID-19: Psychosocial consequences of the COVID-19 pandemic in people with diabetes in Denmark—what characterizes people with high levels of COVID-19-related worries?.
The golden rule is that there are no golden rules: A commentary on Paul Barrett’s recommendations for reporting model fit in structural equation modelling.
00088-7&id=ga1.jpg){kind=link}
00088-7&id=gr1.jpg){kind=link}