The pandemic context resulting from COVID-19 has depicted a global emergency, from the health point of view, with more than 6,987,222 deaths to date [World Health Organization (WHO), 2023], and in terms of the quality of life and mental health of people, due to the set of stressors that arose in this context (i.e., physical, social, economic, and psychological) (Bavel et al., 2020; Ornell et al., 2020; Wang et al., 2020; Wetherall et al., 2022). Within these stressors are those that emerged from the indirect effects of health policies and containment efforts, specifically, the policies of confinement and social distancing (Caqueo-Urízar et al., 2020), designed to reduce personal interactions and movements (Maier and Brockmann, 2020; Mayr et al., 2020; Badenes-Plá, 2022; Yu et al., 2023), which generated changes in the social and daily activities of the population (e.g., studies, work, intimate relationships, financial management, home habitability) (Ammar et al., 2020; Gloster et al., 2020; Marroquín et al., 2020; Akbari et al., 2021; Amerio et al., 2021; Gruber et al., 2021; Lal et al., 2021; Manchia et al., 2022; Quintana, 2022; von Keyserlingk et al., 2022).

As is well known, the social environment systematically influences health, considering social, psychological, economic, demographic, local and cultural aspects (Wild, 2005; McMichael, 2011; Pettini and Mazzocco, 2022; Whipple and Evans, 2022; Eyre et al., 2023; Gudi-Mindermann et al., 2023; Ibanez and Eyre, 2023). In this sense, any variation in these aspects will impact the health status of people, being detrimental or beneficial to the population. For example, there is ample evidence on how contexts of physical and social isolation (i.e., plague, influenza, cholera, leprosy, or others) and subjective social isolation are associated with negative impacts on mental health (Cacioppo and Hawkley, 2009; Cacioppo and Cacioppo, 2014; Cacioppo et al., 2015; Bzdok and Dunbar, 2022; Gilbar et al., 2022). The growth of such literature has been exponential during and following the COVID-19 pandemic (Jakovljevic et al., 2020; Mukhtar, 2020; Wilder-Smith and Freedman, 2020; Clair et al., 2021; Liu et al., 2021; Neville et al., 2021; Takian et al., 2021), highlighting negative psychological impacts (i.e., distress, anxiety, depression, and high levels of stress) that are primarily attributed to side effects of confinement (Cuadra-Martínez et al., 2020; Gloster et al., 2020; Kujawa et al., 2020; Kokkinos et al., 2022; Nanath et al., 2022). These effects, which have been reported in all types of populations (e.g., children, adolescents, adults, pregnant women, senior citizens, among others) (Manchia et al., 2022), would be explained by the increase in environmental stressors and the variability of coping resources (Verdolini et al., 2021; McLaughlin et al., 2022; Tracy et al., 2022; Yang et al., 2022; Delhey et al., 2023). Now, stress refers to an emergent relationship between the person and the environment (Lazarus and Cohen, 1977) involving environmental stimuli, their evaluation and the organism’s response (Cohen et al., 1983, 2007, 2019; Segerstrom and O’Connor, 2012). Likewise, environmental stress is defined as the physiological, cognitive, and emotional response that people may experience to various environmental situations, whether at the macro-level (e.g., population density in a city) or in the immediate environment (e.g., housing conditions; Gatersleben and Griffin, 2017).

A wide range of environmental stressors derived from confinement are observed in the literature (Ellen et al., 2021; Hussong et al., 2021; Kumar and Shah, 2021; Kunzler et al., 2021; Salazar et al., 2021; Sheek-Hussein et al., 2021; Szkody et al., 2021; Valdés-Florido et al., 2022; Morgado et al., 2023). Researched are: (1) economic stressors (ES), which refer to the perceived economic impact that a variation in the estimated household budget generates, and thus impacts on job insecurity and economic livelihood in households (Bazzoli et al., 2021; Friedline et al., 2021; Low and Mounts, 2022) (2) everyday activities (EA), understood as the impact on the performance of routine activities, being considered an individual facet of social practice (Rieger and Wang, 2020; Ellen et al., 2021); (3) social activities (SA), which refers to the perceived impact on social recreation activities such as interaction with others and leisure (Kunzler et al., 2021; Nielsen et al., 2021); (4) home habitability (H), referring to the operational housing conditions and comfort (i.e., conditions necessary to satisfy the physical, biological, psychological, and social well-being) of those who inhabit a dwelling (Corral-Verdugo et al., 2011; Zulaica and Oriolani, 2019); and (5) virtual media exposure (VME), understood as the level of person’s exposure to, or interaction with, virtual or technological media (e.g., television viewing, computer use, use of social networks, websites and mobile applications; Dubey, 2020; Rivest-Beauregard et al., 2022).

Consequently, the frequency of environmental stressors, whether higher or lower, defines how confinement or other situations that cause stressful environmental changes in people’s habitability (e.g., population displacement due to natural disasters and/or in search of shelter) will be experienced, thereby affecting their health (Choi et al., 2020; Kim et al., 2021; Bzdok and Dunbar, 2022). Thus, the evaluation of these environmental factors will provide a more complete view of how the health of the population is affected, and how to counteract this situation. In this way, it will be possible to generate preventive actions, either at the individual or collective level, that are aimed at the well-being of people.

In this line, and considering that efforts to study this phenomenon lack a psychometric instrument with evidence of validity, the present study aims to design a scale that assesses the perception of change in environmental conditions of confinement incorporating a cross-sectional approach, and thus reduce the existing gap in research on the measurement of the impact of confinement (Manchia et al., 2022). To this end, updated evidence of both reliability and validity is presented, following the guidelines of ethical and methodological standards recognized in the field of psychometric evaluation (Prieto and Delgado, 2010; American Educational Research Association, 2014; Muñiz and Fonseca-Pedrero, 2019).

First, based on a qualitative analysis, it was decided to discard one item of the dimension because it loaded positively on the factor despite being classified as an inverse item. Then, a 29 items ESEM model (Model 1; M1) was estimated, as shown in Table 1. This model showed an excellent statistical fit according to the parameters proposed by Hair et al. (2019), for the CFI estimator (CFI = 0.972; TLI = 0.958). However, three items of the everyday activities factor presented relevant cross-loadings (i.e., λ > 0.5) on the social activities factor (i.e., perform physical activity; obtain medical or health care easily; take walks/visits).

Consequently, after a qualitative analysis (i.e., item relevance and construct definition), the items above became part of the social activities factor. Likewise, the remaining items of the everyday activities factor (e.g., carrying out procedures typically) were discarded due to their redundancy, leaving 26 items.

Second, a 4-factor model with 26 items was estimated (Model 2; M2), which presented better internal structure adjustments in the CFI and TLI indicators (CFI = 0.972; TLI = 0.960), compared to M1. As previously mentioned, items with lower factor loadings referring to the factor (λ > 0.5) and that presented cross-loadings (λ ≤ 0.3) were iteratively eliminated to reduce the number of items in the scale. In total, 20 items were retained, estimated in Model 3 (M3).

The ESEM analysis of M3 presented better internal structure adjustments in the CFI and TLI indicators (CFI = 0.977; TLI = 0.962) than the previous models. This structure was confirmed by performing a CFA, which evidenced a satisfactory fit for the CFI, TLI, and RMSEA indicators (CFI = 0.978; TLI = 0.975; RMSEA = 0.069) (Hair et al., 2019). Finally, the ECSS-20 comprises four dimensions (i.e., ES, SA, H, and VME) and five items per dimension (i.e., 20 items in total).

Table 2 presents the factor loadings with their corresponding factorial covariances and reliability coefficients of the M3. Factorial loadings for this model proved adequate for each factor, and no relevant cross-loadings were observed. Also, structural relationships between dimensions were moderate (r > 0.30), mild (r > 0.10; Cohen, 1988), and null, and reliability estimates were adequate (ω > 0.89; α > 0.89; Cho and Kim, 2015).

Third, a multigroup CFA model was estimated between M3 men and women; the results are presented in Table 3. This model was first tested for configural invariance, i.e., a baseline model was fitted for each group separately. Compared with the configural model, the metric model showed no relevant changes in the RMSEA differential or CFI, thus confirming that the factor loadings of the items are the same in both groups. Finally, the scalar model compared with the configural model also showed no relevant changes in the RMSEA differential or CFI, which means that the intercepts of the items are the same in both groups.

Thus, strong measurement invariance is demonstrated by the existence of metric and scalar invariance, i.e., the equivalence between factor loadings and thresholds for those who identified themselves as female or male is sustained (Cheung and Rensvold, 2002; Chen, 2007; Dimitrov, 2010; Leitgöb et al., 2023).

Finally, the SET-ESEM model that estimated the association between the latent dimensions of the ECSS-20 and the PSS-14 one-dimensional showed comparative and absolute fit indices far from the recommendations [χ²(469) = 1780.710; CFI = 0.911; TLI = 0.893; RMSEA = 0.094; 90% CI = (0.090, 0.099); SRMR = 0.094]. The observed mismatches could be attributed to the factor loadings shown by the PSS-14 (see Figure 1). Finally, significant direct and inverse loadings are observed for the ES factor (λ = 0.323) and the H factor (λ = −0.301) concerning PSS-14. The details of the standardized relationships between the latent dimensions and the PSS-14 indicators are shown in the Figure 1.

The present study focused on developing and validating the Environmental Stressors in Confinement Scale (ECSS-20), which assesses the perception of change in environmental stressors produced by confinement circumstances. Theoretical and practical contributions, limitations, and future lines of research emerging from this study are discussed below.

In first place, the final structure of the ECSS-20, composed of four dimensions: economic stressors (ES), social activities (SA), habitability (H), and exposure to virtual media (VME), proved to be robust and consistent with previous literature evidencing the existence of environmental stressors stemming from confinement (Ellen et al., 2021; Kunzler et al., 2021; Sheek-Hussein et al., 2021). The reason for this is that the observed factor loadings indicate that each dimension uniquely impacts the perception of environmental stress in confinement situations. McDonald’s omega (ω > 0.89) and Cronbach’s alpha (α > 0.89; Cho and Kim, 2015), internal consistency values were also shown to be satisfactory, providing evidence of the reliability of the instrument. That is, the ECSS-20 can be applied to the Chilean adult population experiencing confinement measures.

Second, the application of gender invariance tests supported the equivalence of factor loadings between women and men, suggesting that the ECSS-20 can be used in both groups without distinction. Thus, this strengthens the instrument’s usefulness, as it demonstrates that it assesses the impact of environmental stressors accurately and comparatively, as has been evidenced in other studies analyzing gender invariance in confinement contexts (Prime et al., 2021).

Third, validity tests based on the association with other variables were established by demonstrating significant relationships between the ECSS-20 and PSS-14 dimensions. Specifically, direct significant relationships were observed between the economic stressors (ES) dimension and the perception of stress. Therefore, the greater the perception of variation in the estimated household budget, the greater the feeling of stress. Likewise, inverse relationships were observed between the dimension of habitability (H) and the perception of stress. In other words, the higher the perception of well-being with those who share a dwelling, the lower the feeling of stress. As a result, this supports previous research that suggests these factors impact the change in people’s mental health (Kunzler et al., 2021; Sheek-Hussein et al., 2021). Thus, this demonstrates that habitability and economic conditions could function as protective and risk factors in the perception of stress during confinement situations, such as during the COVID-19 pandemic. Therefore, it is proposed to consider these subscales to assess these factors in confinement circumstances.

In contrast, one of the limitations of this study is the mismatch in the RMSEA values of M1, M2, and M3 in the ESEM analyses, suggesting that the model structure and accuracy of fit could be improved. This discrepancy could be explained by factors such as the complexity of the interrelationships between dimensions or the presence of variables not considered in the model (Shi et al., 2020). In addition, the sample size and the absence of validity tests on a sample other than those collected in this study should be considered as a limitation of this study. Finally, a limitation of this research is that the survey was conducted during the medium health impact phase mandated by the Chilean government, characterized by the reduction of social interactions through measures such as social distancing or confinement, including a distance of one meter between two people and the use of a permit demonstrating current vaccination status for transit in the city (Ministerio de Salud, 2022).

This context of dynamic and sometimes irregular confinements, as evidenced in studies such as Patrono et al. (2024), demonstrates the heterogeneity of the confinement experience in the country, with differentiated impacts on the mental and physical health of the population (Duarte and Jiménez-Molina, 2022; Gutiérrez-Pérez et al., 2024). The research by Dagnino et al. (2020) underscores the significant psychological impact and the high demand for psychological support in Santiago, reflecting a critical need that could influence the structure of the ECSS-20 under different confinement intensities. Governmental policies, criticized for their improvised and unequal approach, particularly in terms of gender equity (Undurraga and López-Hornickel, 2023), and the special vulnerability of minority communities (Anandarajah et al., 2024), require careful consideration in interpreting the ECSS-20 data. Recently, Rodman et al. (2024), offered a longitudinal perspective on the deterioration of youth psychopathology due to reduced socialization, a factor that must be considered when assessing the validity of the ECSS-20 in future research. Additionally, studies on the impact of the social environment, such as that by Choi et al. (2024), highlight how specific neighborhood characteristics, such as socioeconomic deprivation and disorder, can increase the risk of dementia, mediated in part by subjective loneliness. This link underscores the importance of considering how urban environments and socialization dynamics influence mental health at all life stages (Ibanez et al., 2024; Migeot et al., 2024). The ECSS-20 ability to capture variations in the perception of environmental stress could be crucial not only for a better understanding of mental health disorders in youth but also for exploring longitudinal connections with cognitive risks in later life stages, influenced by social isolation and neighborhood conditions.

Finally, future lines of research could apply the ECSS-20 in contexts where stress is generated by confinement and/or population displacement, such as those generated by natural disasters (i.e., tsunamis, fires, landslides, extreme heat, hurricanes and tornadoes) (Sandoval-Díaz and Martínez-Labrín, 2021; Birkmann et al., 2022) or by sociopolitical situations of the countries of residence (i.e., political asylum, immigration) (Kwok and Ku, 2008; Kim et al., 2021). These events, marked by critical sociopolitical dynamics and needs for rapid adaptation, present a fertile ground to assess variations in subjective well-being, coping, and fields of spatial justice and habitability (Astudillo Pizarro and Sandoval Díaz, 2019; Sandoval-Díaz et al., 2021, 2022, 2024). Moreover, employability circumstances such as job loss or absence, or significant changes in work conditions, as well as the loss of daily social activities and hospitalizations, are critical areas where the ECSS-20 could reveal significant impacts on mental and physical health (Bahamondes-Rosado et al., 2023; Pérez-Villalobos et al., 2023). It is also advisable to explore how the ECSS-20 functions across different cultures and countries, as stress and its perception can vary considerably among different environments and populations (Tasnim et al., 2024; Tonon et al., 2024). The COVID-19 pandemic, as a natural experiment, has provided a unique context to better understand these phenomena (Gormley, 2024; Ruggeri et al., 2024). Therefore, adapting and validating the ECSS-20 in diverse cultural and environmental stress contexts could enrich our understanding of the interactions between the environment, stress, and mental health, thereby broadening the practical applications of the scale in designing targeted interventions and effective public health policies.

The instrument’s multidimensional structure, internal consistency, gender invariance, and evidence associated with related measures support its validity and usefulness. The ECSS-20 is a valuable tool for investigating and further understanding the effects of confinement on the population’s mental health. Future research could explore its applicability in different contexts and populations to strengthen understanding of its psychometric properties and utility in assessing confinement situations.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The studies involving humans were approved by Comité ético científico Universidad de Tarapacá. The studies were conducted in accordance with the local legislation and institutional requirements. The participants provided their written informed consent to participate in this study.

JS-P: Writing – review & editing, Writing – original draft, Validation, Project administration, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. RF-U: Writing – original draft, Validation, Supervision, Software, Resources, Methodology, Funding acquisition, Formal analysis, Data curation, Conceptualization. GS-P: Writing – original draft, Validation, Supervision, Project administration, Methodology, Formal analysis, Data curation, Conceptualization. JF: Writing – original draft, Investigation. KA-C: Writing – review & editing, Writing – original draft, Visualization, Validation, Project administration, Methodology, Investigation, Formal analysis, Data curation, Conceptualization.