Relationships among COVID-19 causal factors perceived by children, basic psychological needs and social anxiety

Participants

The study design is cross-sectional, and the sample collection was performed through non-probabilistic and incidental sampling. The sample consisted of 58 schoolchildren (M age = 10.18; SD = 0.77; 36 boys and 22 girls) from the Valencian Community in Spain. The children’s hours of study were 7.98 h per week, 10 suffered a COVID-19 infection (17.24%), 31 had a relative infected by COVID-19 (53.45%), 17 participants and their relatives were not infected by COVID-19 (29.31%), 41 practiced sports and physical activity (70.69%) and 17 were sedentary (29.31%). In particular, 18 were associated athletes (30.51%), 14 competed at regional level (24.14%) and five competed at national level (8.62%).

The sample was selected according to the study’s goals as it was intended to examine whether COVID-19 causal factors perceived by children predicted basic psychological needs and social anxiety; and if basic psychological needs predicted social anxiety. The only inclusion criterion was to evaluate schoolchildren as we wanted to examine the effects of the epidemic on the aforementioned variables in schoolchildren.

Instruments

To measure the perception of the population regarding the pandemic causality, it was utilized the “Questionnaire about the causality factors of COVID-19 pandemic” (CPFC-COVID-19) (Prieto, 2022). The questionnaire consists of 20 items with a Likert-type scale ranging from 1. totally disagree, 2. disagree, 3. agree and 4. totally agree. The CPFC-COVID-19 has four dimensions: 1) “Social distancing and protection” (DSP) with six items (e.g., I think that if I go down the street without a mask, I put no one at risk of contagion); 2) “Perceived psychological impact” (IPP) with five items (e.g., I think that stress influences the risk of contagion by COVID-19); 3) “Skepticism” (ES) with four items (e.g., I am sure I will not get infected if I go with my group of friends); and 4) “Credibility of the information received” (CIP) with five items (e.g., I think that the population is not well informed about the political and social situation related to curbing COVID). As Cronbach alpha increases with a high number of items, it was taken as an internal consistency marker the mean inter-item correlation (Clark & Watson, 1995; DeVellis, 2003). Several scholars have selected this measure in which an inter-item correlation higher than 0.15 is an adequate cut-off point. Thus, in this study the mean inter-item correlation of the distinct factors was: 0.20 (DSP), 0.27 (IPP), 0.50 (ES) and 0.40 (CIP). Finally, it is important to highlight that the authors have permission to use this instrument from the copyright holders and the instrument has shown appropriate reliability and validity in a previous study (Prieto, 2022).

The Spanish version of the Basic Needs Satisfaction in General Scale (BNSG-S) (Costa et al., 2022; Gagné, 2003). The original scale is made up of 21 items that measure the satisfaction of competency needs (r = 0.16; six items, e.g., “Most days I feel that I am successful in what I do”), autonomy satisfaction (r = 0.20; seven items, e.g., “I feel that I am free to decide for myself how to live my life”) and relatedness (r = 0.34; eight items, e.g., “I really like the people that I interact with”). In each of the dimensions, there were three items written in a negative way. The participants had to answer all the items thinking about how they related to their life and indicating how true they were for them on a Likert-type scale from 1 (not true) to 7 (totally true). Finally, it is important to highlight that the authors have permission to use this instrument from the copyright holders and the instrument has shown appropriate reliability and validity in previous studies (Gagné, 2003; González-Cutre et al., 2015).

The Psychological Needs Thwarting Scale was utilized (EFNP) (Bartholomew et al., 2011; Sicilia, Ferriz & Sáenz-Álvarez, 2013). The original scale is preceded by the phrase “In exercise…” but in this study to be adapted by life in general it was changed to: “In my life…”. It consists of 12 items that are distributed in four items for each of the three subscales in which it is composed: a) Frustration of the need for autonomy (r = 0.51; e.g., “ I feel obligated to follow the decisions of others”); b) Frustration of the need for competence (r = 0.48; e.g., “there are situations in which I feel incapable”); and c) Frustration of the need for relatedness (r = 0.25; e.g., “I feel that other people don’t like me”). For each item, the participants must indicate their response on a 7-point Likert scale ranging from 1 = do not agree at all, 2 = very slightly agree, 3 = slightly agree, 4 = moderately agree, 5 = agree, 6 = strongly agree to 7 = very strongly agree. Higher scores indicate a higher level of frustration of psychological needs. Finally, it is important to highlight that the authors have permission to use this instrument from the copyright holders and the instrument has shown appropriate reliability and validity in previous studies (Bartholomew et al., 2011; Sicilia, Ferriz & Sáenz-Álvarez, 2013).

The social anxiety questionnaire (CASO-N24) (Caballo et al., 2012) was used in this study. This measure presents two versions, one aimed at boys and the other at girls, both with 24 items with a Likert-type response scale from 1 to 4, being 1: not at all, 2: a little, 3: quite a lot, and 4: a lot. The items of this questionnaire were obtained from a review of the literature on childhood fears and problematic social situations in children who came for consultation for social phobia. The questionnaire contemplates six factors: Public speaking/Interaction with teachers (r = 0.53); Interaction with the opposite sex (r = 0.63); Being embarrassed or ridiculed (r = 0.37); Assertive expression of annoyance or anger (r = 0.49); Interaction with strangers (r = 0.54); and Acting in public (r = 0.33). Finally, it is important to highlight that the authors have permission to use this instrument from the copyright holders and it has shown appropriate reliability and validity in previous studies (Caballo et al., 2012).

Procedure

The study was approved by the ethics committee of Universidad Internacional de La Rioja (UNIR) whose approval code is PI024/2022. The study complied with the ethical standards of the American Psychology Association and with the rules of the Declaration of Helsinki (World Medical Association, 2013). Before beginning the process of administering the questionnaires, permission was requested from the school centre and the student’s parents. A signed in-person informed consent was obtained from all participants’ parents.

In the informed consent, the confidentiality and anonymity of the data was ensured. The questionnaires were administered individually through a Google forms link at the computer room in the school hours from February 2022 to April 2022. The administration of the questionnaires was carried out by teachers and research staff to assist with any possible doubts that might arise during the procedure. The length of the survey was around 20 min. Once the questionnaires were finished, the data were directly recorded in Google’s database.

Data analyses

The statistical software R (4.1.1) was utilized to conduct the distinct statistical analyses. To analyze the relationship across the pandemic causality factors, basic psychological needs satisfaction and frustration, and social anxiety, a Partial Least Square Path Modelling (PLS-PM) approach was taken (Sánchez, 2013). PLS-PM is a variance-based structural equation modelling technique which is not constrained by distributional assumptions (Martinent et al., 2019; Nicolas, Drapeau & Martinent, 2017; Sánchez, 2013). Thus, it signifies that PLS-PM is a good technique to work with little sample sizes. The significance of the parameter estimates is assessed by constructing 95% bias-corrected percentile confidence interval based on a bootstrap procedure with 100 replications (Martinent et al., 2019). Concerning all the variables (causality factors of COVID-19, basic psychological need satisfaction and needs frustration), two parcels were made (Coffman & MacCallum, 2005) in each factor. To create the parcels it was used a subset-item-parcel approach in which there were a random aggregate of items. This approach of parceling was taken because it improves the measurement properties of constructs, simplifies model complexity, enhances normality, reduces multicollinearity, and provides stable estimates in PLS-PM, especially when dealing with small sample sizes (Matsunaga, 2008).

In addition, different steps were followed to run the distinct analyses. First, the psychometric properties of the variables measured were examined. The indexes measured to ensure sufficient quality were: Standardized factor loadings, composite reliability values (ρ), average variance extracted (AVE) values and eigenvalue analysis of the correlation matrix of each set of manifest variables (Martinent et al., 2019; Nicolas, Drapeau & Martinent, 2017). Standardized factor loadings higher than 0.40, (Martinent et al., 2019) ρ values greater than 0.70, (Raykov, 2001) AVE values equal or greater than 0.50, (Fornell & Larcker, 1981) the first eigenvalue larger than 1 and the second one smaller than 1 (Nicolas, Drapeau & Martinent, 2017) indicate acceptable reliability of latent and manifest scores. Second, the structural model was tested to examine the connection among the latent variables examined in the study.

Descriptive statistics and correlations among the variables

Table 1 shows the descriptive statistics and correlations among the variables. The results revealed low-medium scores in social distancing and protection (DSP), perceived psychological impact (IPP), Skepticism (ES), credibility (CIP), autonomy frustration (THAUT), competence frustration (THCOMP), relatedness frustration (THRELA), public speaking/interaction with teachers (HABPUB), interaction with the opposite sex (SEX), being embarrassed or ridiculed (EVI), assertive expression of annoyance or anger (ASERT), interaction with strangers (DESCON) and acting in public (ACT). On the other hand, high scores were found in autonomy satisfaction, competence satisfaction and relatedness satisfaction. Moreover, correlational analyses showed that there was no multicollinearity between the study variables as correlations ranged from −0.54 to 0.71 and none of the confidence intervals (i.e., r ± two standard errors) were close to 1.0.

Inner PLS-PM model

Table 2 shows the results of the inner PLS-PM model in which the distinct indexes provided enough markers of the reliability and validity of the model. In particular, the standardized factor loadings ranged between 0.60 and 0.96 (M = 0.86; SD = 0.11), the ρ values ranged between 0.72 and 0.95 (M = 0.86; SD = 0.07), the AVE values ranged from 0.60 to 0.91 (M = 0.74; SD = 0.15) whereas the first eigenvalues ranged from 1.21 to 1.78 (M = 1.55; SD = 0.18) and the second eigenvalues ranged from 0.16 to 0.83 (M = 0.48; SD = 0.20).

PLS-PM model (relationships between the latent variables)

Table 3 shows the results of the PLS-PM model (Fig. 1: i.e., relationships between the latent variables). The results revealed that social distancing and protection significantly negatively predicted competence satisfaction (β = −0.13; p < 0.05). Besides, perceived psychological impact significantly negatively predicted relatedness satisfaction (β = −0.35; p < 0.05), significantly positively autonomy frustration (β = 0.34; p < 0.05) and competence frustration (β = 0.26; p < 0.05). In addition, perceived psychological impact significantly negatively predicted interaction with the opposite sex (β = −0.36; p < 0.05). Moreover, autonomy satisfaction significantly positively predicted interaction with the opposite sex (β = 0.48; p < 0.05). On the other hand, autonomy frustration significantly positively predicted public speaking/interaction with teachers (β = 0.44; p < 0.05), significantly positively interaction with the opposite sex (β = 0.35; p < 0.05), significantly positively being embarrassed or ridiculed (β = 0.46; p < 0.05) and significantly positively interaction with strangers (β = 0.49; p < 0.05).

Social distancing, protection and competence

Results revealed that social distancing and protection significantly negatively predicted competence satisfaction. These results are in line with the study of Scheid et al. (2020) as it states that the mask use and distance measures may influence the perception of competence satisfaction in schoolchildren. These results make sense because compliance with mask use and distance measures may hinder how children work at school and the way in which they interact with the environment and their peers (Coelho et al., 2022). For instance, interacting with children when working on a group task is different when wearing a mask or not. Thus, the experience of working at school and respecting the rules may hinder competence satisfaction as there are several measures to be respected that may worsen competence.

Perceived psychological impact and basic psychological needs

Perceived psychological impact significantly negatively predicted relatedness satisfaction, significantly positively autonomy frustration, and competence frustration. This is in line with previous studies in which the distinct outcomes caused by the epidemic have triggered a different influence on basic psychological needs (Šakan, Žuljević & Rokvić, 2020).

In this sense, perceived psychological impact can negatively influence a person’s sense of autonomy, making them feel less able to control their own actions and decisions. For example, people who experience high levels of perceived psychological impact may feel compelled to avoid social situations due to fear of judgment or negative evaluation. This avoidance can be seen as a lack of control over their own lives, which can diminish their sense of autonomy. In another way, perceived psychological impact can undermine the sense of competence, especially in social contexts, where the perception of not meeting social expectations can be debilitating. For instance, studies have shown that a lack of perceived competence in social interactions is closely related to high levels of social anxiety (Weeks et al., 2008). These results are in line with a previous study that highlights the increased difficulties to interact with others in pandemic times due to the impact on fear and anxiety (Calbi et al., 2021).

Perceived psychological impact and social interaction

In addition, results revealed that perceived psychological impact significantly negatively predicted interaction with the opposite sex. Specifically, may lead to avoiding social interactions to prevent situations that might exacerbate negative feelings, due to the risk of contagion. In this sense, the increase in stress, anxiety and depressive symptoms related to contagion (Chen et al., 2020) could generate, due to preventive measures, social conflict, isolation, devaluation, rejection and exclusion (Slavich, 2020). This avoidant behavior is a common characteristic of social anxiety, which has been shown to negatively affect interpersonal relationships (Alden & Taylor, 2004).

Otherwise, the interaction with the other sex may help to reduce those perceived psychological impact experienced in the epidemics, disconnecting and helping to reduce the maladaptive outcomes of the causal factors provoked by the pandemic (Luo et al., 2022). In this sense, appropriate social supports with alternative means for interpersonal communication (e.g., social media), frequent contact with colleagues, can alleviate stress (Yan et al., 2021) and promote interaction between people. Thus, despite the increase of fear and anxiety which minimize the interaction with others, it is needed to emphasize in children the need to socialize with others respecting the rules as a way to enhance their wellbeing.

Social interaction, autonomy and social anxiety

According to self-determination theory, higher autonomy satisfaction would be expected to lead to better social outcomes, and conversely, autonomy frustration would be associated with negative outcomes such as social avoidance. However, our findings indicate that social interaction is not significantly affected by autonomy levels, which challenges the typical expectations of self-determination theory. One possible theoretical explanation is that both autonomy satisfaction and frustration may motivate social interactions, albeit for different reasons. Autonomy satisfaction might drive positive relationships due to feelings of competence and security, whereas autonomy frustration might generate interactions in an attempt to compensate for a lack of social control or support. Therefore, both can be present in all relational contexts, as suggested by the cited study (Sibley & Overall, 2008).

Moreover, autonomy satisfaction and frustration significantly positively predicted social anxiety in the interaction with the opposite sex. This finding suggests that social anxiety in the interaction with the opposite sex may depend on other factors different than autonomy satisfaction or frustration. It is possible that social interactions with the opposite sex may be maintained regardless of autonomy level, due to external motivations such as social norms, interpersonal attraction, or the need for affiliation, such as the frequency of feeling supported by using technologies, seeing friends, creating challenges or carrying out the same activities simultaneously (Muñoz-Fernández & Rodríguez-Meirinhos, 2021). Furthermore, results revealed that autonomy frustration was positively related to anxiety to being embarrassed or ridiculed. The pandemic context likely influenced results, as increase stress (Chen et al., 2020), social isolation, and uncertainty may have heightened social conflict, rejection and exclusion (Slavich, 2020), regardless of frustration and autonomy satisfaction, with women being more susceptible to COVID-19-related mental health problems (Yalçın et al., 2021), and gender differences being evident in exposure to potentially traumatic events (Street & Dardis, 2018). Besides, autonomy frustration positively predicted anxiety in public speaking/interaction with teachers and significantly positively interaction with strangers. In particular, when a person experiences frustration in their autonomy, they may feel they do not have control over their decisions. As a consequence, they may develop social anxiety, especially if they struggle to maintain social relationships, in line with the following study (Gao et al., 2022). On the other hand, those who are restricted in their social interaction, for example, due to pandemic isolation, may experience fear and social anxiety due to reduced exposure to frequent social situations (Rahm-Knigge, Prince & Conner, 2021). In this sense, encouraging autonomy and socialization, especially in educational contexts and during periods of isolation such as a pandemic, is crucial for promoting emotional well-being and prevent disorders such as social anxiety.

Some of the limitations are that the usage of self-report measures may lead to some bias such as: social desirability, acquiescence and dishonestly (Latkin et al., 2017). Nevertheless, the measures taken were the most appropriate to the population selected. On the other hand, the utilization of a specific school may hinder the generalization of the results to other different populations and contexts. As such, the results should be understood in the context in which the study was conducted. Finally, the small sample size analyzed in this study hinders the generalization of the findings obtained and minimizes the power of the conclusions shown. Thus, in future studies, it would be interesting to increase the sample size, adding more schools from different regions and students with diverse sociodemographic characteristics to facilitate the generalization of the findings obtained. Nevertheless, the results obtained are valuable as they show a big picture of the relationship among COVID-19 causal factors, basic psychological needs and social anxiety, which may help in future epidemics to make decisions to enhance children’s well-being. It is also worth noting that the literature also indicates that female adolescents show higher levels of social anxiety than male adolescents, so future research could make these comparisons (Prieto, 2020).