Health During COVID-19: The Roles of Demographics, Information Access, and Government Policy

Abstract

This study highlights how socio-demographic, information, and government factors play different roles in people’s health during COVID-19 between Asian countries and non-Asian countries by employing the Hierarchical Linear Regression. This study finds that government factors play a more significant role in shaping wellness and happiness in Asian countries, whereas they have a stronger impact on health status in non-Asian countries. Second, information factors—such as knowledge about vaccines, medical professionals, and reliable sources—have a more substantial effect on health status in Asian countries, while they are more strongly associated with wellness and happiness in non-Asian contexts. Third, socio-demographic variables exert a greater influence on overall health outcomes in non-Asian countries compared to Asian countries. In particular, gender, occupation, socioeconomic placement, height, and weight consistently play a significant role across all health dimensions in non-Asian countries, whereas their impact varies across different health domains in Asian settings.

1. Introduction

The COVID-19 pandemic has profoundly transformed global perceptions of health, well-being, and governmental responsibility. As societies navigate the aftermath of this global crisis, understanding the multifaceted determinants of health outcomes—including not only physical health status but also wellness and happiness—has become more critical than ever.

The COVID-19 pandemic has profoundly disrupted the daily lives of individuals across all age groups, with particularly acute effects on children and families. Extended school closures, limited access to outdoor activities, and prolonged social isolation have interfered with children’s educational progress, emotional development, and mental well-being. For adults, the pandemic altered employment patterns, increased economic uncertainty, and reshaped daily routines through work-from-home arrangements and mobility restrictions (e.g., shutdowns, surveillance, and mask mandates) [1,2,3].

The multidirectional impact of COVID-19 on health extends far beyond acute respiratory infection. Medically, the virus has shown significant pathogenicity across multiple physiological systems. In the respiratory system, it affects both the lungs and upper airways, sometimes leading to structural changes in airway dimensions and chronic pulmonary impairment. Beyond physical consequences, the pandemic has triggered profound psychosocial disruptions across all layers of society [4,5,6].

The COVID-19 pandemic not only disrupted global health systems but also triggered a cascade of psychosocial challenges that continue to affect individuals and institutions long after the initial crisis. Prolonged periods of isolation, uncertainty, and loss led to widespread increases in anxiety, depression, and social withdrawal across populations, disproportionately affecting vulnerable groups such as the elderly, children, and economically disadvantaged individuals. In response, many governments and health systems intensified medical protocols, improving disease surveillance, vaccine distribution, and digital health services. However, these advancements are paralleled by significant transformations in other sectors, such as education. These multidimensional impacts of the pandemic underscore the importance of addressing health not only through medical solutions but also through integrated social and institutional responses [7,8,9,10,11,12].

The spread of false information about the COVID-19 pandemic—particularly regarding the safety and effectiveness of vaccines—significantly hindered public health efforts in many countries. As a result, public trust in scientific guidance and vaccination programs declined in certain populations, leading to increased vaccine hesitancy and delays in achieving adequate immunization coverage. Therefore, access to the right information played an important role in reducing COVID-19 infection [13,14,15].

The pandemic introduced an unprecedented reliance on state-led public health strategies—such as nationwide lockdowns, surveillance technologies, and mandatory mask mandates—while also amplifying the role of information flows, including vaccine literacy, trust in medical professionals, and exposure to reliable health information. Simultaneously, the crisis unearthed deep-seated inequalities rooted in socio-demographic characteristics such as income, gender, occupational stability, and physical health indicators like height and weight. These dimensions, often overlooked in conventional health policy frameworks, have come to the forefront in discussions surrounding health justice and pandemic resilience [16,17,18].

The cross-regional comparison between Asian and non-Asian countries is particularly significant. Previous literature suggests that cultural orientations, institutional trust, and governance styles diverge markedly across these contexts, potentially moderating the effectiveness and perception of health interventions [19,20,21].

Therefore, this study explores three broad categories of influence: (1) socio-demographic factors (e.g., gender, job, economic placement, height, and weight), (2) information factors (e.g., access to reliable health information, vaccine, and doctors), and (3) governmental factors (e.g., shut down, surveillance, and mask usage). To be specific, the objective of this study is to examine how socio-demographic factors, informational variables, and government interventions influenced individuals’ perceived health status, wellness, and happiness during the COVID-19 pandemic according to countries. By applying hierarchical linear regression models across Asian and non-Asian populations, the study aims to identify both shared and region-specific determinants of health outcomes and to assess how informational and political responses to the pandemic shaped public well-being across different cultural and structural contexts.

For the purpose, this study utilizes the International Social Survey Programme: ISSP 2021–Health and Health Care II by employing the Hierarchical Linear Regression (HLR). To the best of my knowledge, this study is the first article exploring the effects of socio-demographic, information, and government factors on health status, wellness, and happiness between Asian and non-Asian countries by employing econometric models.

2. Literature Review

Health outcomes refer to the measurable changes in health status that result from interventions, behaviors, environmental exposures, or policy decisions. They serve as key indicators of population well-being and are widely used to evaluate the effectiveness of healthcare systems, public health initiatives, and broader social determinants. Health outcomes encompass not only the absence of disease but also broader dimensions of physical, psychological, and social well-being [22,23,24].

Health outcomes are typically categorized into several dimensions. First, health status refers to the objective and subjective assessments of an individual’s physical condition. It includes the presence or absence of disease, disability, or functional limitations, as well as self-rated health, which serves as a powerful predictor of morbidity and mortality [25,26,27].

Wellness extends beyond the clinical definition of health and captures a holistic state of balance in physical, emotional, intellectual, and social aspects of life. It reflects a proactive approach to living that emphasizes health promotion, quality of life, and resilience in the face of stressors [28,29,30].

Happiness and life satisfaction reflect an individual’s overall evaluation of their life experiences. Increasingly recognized as critical health outcomes, they are influenced not only by personal health status but also by social relationships, environmental factors, and institutional trust [31,32,33].

A considerable body of research affirms the influence of socio-demographic variables—including gender, age, income, education, employment, and residential location—on health outcomes [34,35]. For example, Vigl et al. highlight that socio-demographic factors show a strong effect on health quality of life from a questionnaire-based survey in 676 adults [36]. Oleribe and Alasia report that there is a significant relationship between socio-demographic variables and family health based on a total of 608 people in rural primary health care centre in Katcha local government area of Niger State [37].

Access to reliable health information, including knowledge about vaccines, doctor qualifications, and general public health guidelines, significantly affects health-related behavior and perception [38,39,40]. For instance, Arora et al. show that patients who report greater troubles in accessing needed information experience lower emotional, functional, and social/family scores, with 225 women surveyed within 6 months of diagnosis [41]. Graves finds that disparities in health outcomes across populations and geographical regions can be affected by information systems based on integrative literature review [42].

Government interventions during COVID-19—such as lockdowns, digital surveillance, and mask mandates—are central to public health strategies, with varying impacts across regions [43,44,45]. For instance, Williams et al. show that there are five positive changes, such as more quality time with family, developing new hobbies, more physical activity, and better quality of sleep, based on 3342 adults in Scotland during the COVID-19 lockdown [46]. Adjodah et al. find that masks are positively associated with a reduction in COVID-19 cases and deaths through all 50 states and D.C. and 69 countries [47].

3. Research Methodology

This study utilizes the International Social Survey Programme: ISSP 2021–Health and Health Care II produced by Infrastructure and Research for Social Sciences–GESIS. The International Social Survey Programme (ISSP) is an ongoing international research initiative that conducts annual surveys on key topics relevant to the social sciences. Launched in 1984 by four founding countries—Australia, Germany, Great Britain, and the United States—the program has since expanded to include nearly 50 member nations worldwide. Designed to allow for both cross-national and longitudinal comparisons, the surveys are structured for replication over time. Each ISSP module addresses a particular theme, which is revisited at regular intervals to ensure consistency and comparability [48].

Table 1. Samples of countries.

Regions	Countries	N
Asian Countries	China	2689
	India	1683
	Israel	1187
	Japan	1453
	Philippines	1800
	Russian Federation	1597
	Thailand	1497
	Taiwan	1604
	Total	13,510
Non-Asian Countries	Austria	1546
	Australia	1050
	Switzerland	3349
	Czech Republic	1262
	Germany	1744
	Denmark	1672
	Finland	1002
	France	1584
	Croatia	1101
	Hungary	1008
	Iceland	1086
	Italy	1138
	Mexico	1001
	Netherlands	1269
	Norway	1518
	New Zealand	1135
	Poland	1098
	Slovenia	1020
	Slovakia	1013
	Suriname	1468
	United States of America	1146
	South Africa	2829
	Total	31,039
	All countries (Asia + non-Asia)	44,549

and Figure 1 show samples of countries in the ISSP data.

To be specific, the ISSP 2021 module provides a standardized set of survey items focusing on various dimensions of health and health care, including but not limited to self-rated health status, long-term illness or disability, access to medical care, satisfaction with national health systems, trust in doctors and hospitals, views on government responsibility for health care, and health-related behavior such as smoking and alcohol consumption. These variables are developed collaboratively through the ISSP drafting group and are intended to be implemented across all participating countries.

Second, regarding measurement equivalence, all participating countries are required to adhere to the ISSP’s standardized translation and fieldwork guidelines, which are strictly enforced to ensure cross-national comparability. This includes the use of a rigorous translation protocol (typically the TRAPD method: Translation, Review, Adjudication, Pretesting, and Documentation) that aims to preserve conceptual and semantic equivalence across different languages.

In summary, the ISSP 2021 dataset offers a broad and largely standardized set of health-related variables across countries, measured through instruments designed for high cross-national equivalence. Although minor deviations in item availability or mode of administration may exist, these are transparently documented and do not undermine the dataset’s overall reliability for comparative analysis. This consistent methodological framework allows researchers to confidently conduct both cross-national and longitudinal analyses using the ISSP 2021 Health module.

The dataset includes countries from both Asian and non-Asian regions; however, not all continents are equally represented. For instance, African, South American, and some Middle Eastern countries are underrepresented or absent, limiting the generalizability of the findings to a truly global scale. Nevertheless, the countries included in the ISSP 2021 provide a meaningful basis for comparative analysis between geopolitical and cultural regions—particularly between collectivist and individualist societies—thereby enabling nuanced interpretations of how demographic, informational, and governmental factors influenced health-related outcomes.

This study employs hierarchical linear regression (HLR) models to explore how socio-demographic, information, and government factors play an important role in physical, mental, and emotional health. The dataset employed in this study comprises individual-level responses nested within different groups and countries, which justifies the use of a hierarchical modeling approach. Given that participants are grouped by country, there is a strong theoretical and empirical basis to assume that individuals within the same country may share similar social, political, and information contexts—such as government responses to COVID-19, healthcare systems, or media environments—which could influence their health-related outcomes. Ignoring this nested structure by using a single-level ordinary linear regression model would risk violating the independence assumption and underestimating standard errors, thus potentially leading to biased statistical inferences.

By applying a Hierarchical Linear Regression framework, the study appropriately accounts for these intra-country dependencies while allowing cross-country comparisons between Asian and non-Asian contexts. This modeling choice also enhances the accuracy of coefficient estimates and the reliability of the findings, particularly given the multi-layered nature of health determinants during a global pandemic.

HLR is one of the regression models, which demonstrates if variables of interest exhibit a statistically significant amount of variance in the dependent variable after controlling for all other explanatory variables. To be specific, HLR is a regression model in which the variables are entered in each stage of the model. This study employs three-stage HLR models. The first block entered into the HLR has the socio-demographic variables. The second block entered into the HLR has information variables. The third block entered into the HLR includes government variables.

The dependent variables are health status, wellness, and happiness. In this study, each dependent variable is a single-item outcome variable, directly taken from the ISSP 2021 dataset. Because these dependent variables are not multi-item scales or indices, there is no internal consistency to assess, and thus Cronbach’s alpha is not applicable or required.

Independent variables in the Stage 1 models are gender, age, education, job, marriage, placement, the family number, children, urban areas, height, and weight. Information factors for vaccines, doctors, and reliable information are added in the Stage 2 models. Government factors for shut down, surveillance, and masks are added in the Stage 3 models.

Table 2. Descriptive statistics.

					Asia				Non-Asia
Category	Variable	Explanations	Minimum	Maximum	Mean	S.D.	Minimum	Maximum	Mean	S.D.
Socio-demographic factors	gender	female: 0 and male: 1	0	1	0.5	0.5	0	1	0.4	0.5
	age	under 20 s: 0, 20 s: 1, 30 s: 2, 40 s: 3, 50 s: 4, 60 s above: 5	0	5	3.1	1.5	0	5	3.3	1.5
	edu	under highschool: 0, undergraduate: 1, graduate: 2	0	2	0.7	0.8	0	2	0.9	0.9
	job	no: 0 and yes: 1	0	1	0.7	0.5	0	1	0.7	0.5
	marriage	no: 0 and yes: 1	0	1	0.7	0.5	0	1	0.7	0.4
	placement	no: 0 and yes: 1	0	10	5.5	1.8	0	10	6.0	1.7
	family	the number of family	1	16	4.0	2.0	1	15	2.9	1.5
	children	no: 0 and yes: 1	0	1	0.6	0.5	0	1	0.4	0.5
	urban	no: 0 and yes: 1	0	1	0.7	0.5	0	1	0.6	0.5
	height	logged height	4.1	5.3	5.1	0.1	4.1	5.3	5.1	0.1
	weight	logged weight	3.3	5	4.2	0.2	3.3	5	4.3	0.2
Information	vaccine	information on vaccinations	1	5	3.3	1.2	1	5	3.5	1.1
	docinfo	helped me understand doctor	1	5	3.4	1.8	1	5	3.0	1.1
	relinfo	reliable health information on internet	1	5	3.6	1.1	1	5	3.8	1.0
Government factors	shutdown	shut down businesses and places of employment	1	4	3.1	0.9	1	4	2.8	1.0
	surveillance	digital surveillance to track infected people	1	4	3.1	0.9	1	4	2.4	1.1
	masks	require people to wear face masks	1	4	3.5	0.8	1	4	3.1	1.0
Health outcomes	health status	the magnitude of health status	1	5	3.2	1.1	1	5	3.3	0.9
	wellness	the magnitude of wellness	1	5	2.1	1.1	1	5	2.1	1.1
	happiness	the magnitude of happiness	1	7	5.2	1.2	1	7	5.1	1.0

Note: High values are better on the scale. The minimum and maximum values represent the smallest and largest data points in a set, while the standard deviation (S.D.) quantifies the dispersion or spread of the data around the mean (average).

highlights descriptive statistics in this study. The final sample is 8010 (Asia: 2977 and non-Asia: 5033) after removing improper data, such as no answer or blank values. This study employs SPSS version 30 to conduct hierarchical linear regression analyses.

4. Results

This study firstly shows Variance Inflation Factor (VIF), which measures how much multicollinearity affects the variance of a regression coefficient. VIF is a widely accepted diagnostic tool used to detect the extent to which the variance of an estimated regression coefficient is increased due to multicollinearity. In this analysis, I calculated VIF values for each block of predictors—socio-demographic, internet, and governmental—before entering them into the HLR model.

In most cases, a VIF of 1 indicates no correlation, values of more than 4 or 5 are sometimes regarded as being moderate to high, and values of 10 or more being regarded as very high. The VIF of all variables is also near one, indicating that there are no multicollinearity problems in the models and among variables. Therefore, this study confirms that the interpretation of the regression coefficients is statistically valid and not distorted by high intercorrelations among predictors. This step is taken to ensure the robustness and reliability of the hierarchical regression results.

To assess the meaningful changes in Hierarchical Linear Regression (HLR) models across Stage 2 and Stage 3, we examine the ΔR2 (change in explained variance) and ΔF (F-change statistic). These values help determine whether adding a new set of variables significantly improves the model’s explanatory power beyond the previous stage. ΔR2 represents the increase in the proportion of variance explained by the model after including a new block of predictors. ΔF tests whether the ΔR2 is statistically significant, adjusting for degrees of freedom.

This study highlights the model fits of HLR models by exploring ΔR2 and ΔF. In Asia models, ΔR2 and ΔF of Stage 2 and 3 HLR models are statistically better than the Stage 1 models in wellness, whereas ΔR2 and ΔF of Stage 2 are statistically better than the Stage 1 models in health status, and ΔR2 and ΔF of and 3 HLR models are statistically better than the Stage 1 models in happiness (health status: Stage 2: ΔR2: 0.004 and ΔF: 4.118*** and Stage 3: ΔR2: 0.001 and ΔF: 1.042; wellness: Stage 2: ΔR2: 0.008 and ΔF: 8.297*** and Stage 3: ΔR2: 0.008 and ΔF: 8.288***; happiness: Stage 2: ΔR2: 0.000 and ΔF: 0.352 and Stage 3: ΔR2: 0.014 and ΔF: 14.902***).

In non-Asia models, ΔR2 and ΔF of Stage 2 and 3 HLR models are statistically better than the Stage 1 models in health status and wellness (health status: Stage 2: ΔR2: 0.003 and ΔF: 4.742 *** and Stage 3: ΔR2: 0.004 and ΔF: 6.970 *** and wellness: Stage 2: ΔR2: 0.015 and ΔF: 27.254 *** and Stage 3: ΔR2: 0.003 and ΔF: 4.809 ***). In contrast, ΔR2 and ΔF of Stage 2 HLR models are statistically better than the Stage 1 models in happiness (Stage 2: ΔR2: 0.002 and ΔF: 3.358 *** and Stage 3: ΔR2: 0.000 and ΔF: 0.741).

To be specific,

Table 3. Health status (Asia).

			Stage 1			Stage 2			Stage 3
Category	Variable	Standardized	t	VIF	Standardized	t	VIF	Standardized	t	VIF
Socio-demographic factors	gender	*** 0.071	3.510	1.291	*** 0.071	3.522	1.302	*** 0.070	3.483	1.302
	age	*** −0.054	−2.968	1.062	*** −0.052	−2.829	1.065	*** −0.051	−2.790	1.067
	edu	−0.023	−1.227	1.145	−0.027	−1.395	1.206	−0.028	−1.435	1.209
	job	−0.001	−0.074	1.066	−0.001	−0.066	1.072	−0.001	−0.070	1.073
	marriage	−0.006	−0.324	1.101	−0.006	−0.332	1.102	−0.005	−0.287	1.107
	placement	*** 0.201	11.039	1.063	*** 0.199	10.914	1.066	*** 0.198	10.813	1.071
	family	0.026	1.184	1.594	0.027	1.220	1.596	0.028	1.253	1.597
	children	* 0.043	1.910	1.627	* 0.039	1.722	1.636	* 0.038	1.681	1.644
	urban	** 0.042	2.268	1.093	** 0.042	2.290	1.102	** 0.043	2.303	1.105
	height	0.004	0.180	1.496	0.003	0.148	1.496	0.003	0.142	1.498
	weight	*** −0.149	−7.043	1.421	*** −0.144	−6.791	1.436	*** −0.145	−6.824	1.445
Information factors	vaccine				−0.009	−0.479	1.059	−0.010	−0.533	1.063
	docinfo				*** 0.054	2.987	1.043	*** 0.055	3.042	1.048
	relinfo				0.029	1.600	1.040	0.029	1.630	1.050
Government factors	shutdown							0.029	1.368	1.436
	surveillance							−0.013	−0.612	1.530
	masks							0.026	1.205	1.534
	adj-r2			0.069			0.072			0.072
	N			2977			2977			2977

Note: The notation *** p < 0.01 ** p < 0.05 * p < 0.1 represents statistical significance levels. It indicates the degree to which results are unlikely to have occurred by random chance, with *** showing the highest level of significance (the 99% confidence intervals), ** indicating a lower level of significance (the 95% confidence intervals), and * showing the lowest level of significance (the 90% confidence intervals). Cronbach’s alpha is not applicable to single-item measures because alpha assesses internal consistency among multiple items, requiring inter-item correlations.

shows that in the analysis of health status within Asian countries, both positive and negative associations are identified in Stage 1 of the hierarchical linear regression model. For example, gender (0.071) and placement (0.201) appear as statistically significant positive predictors. This suggests that men and individuals with higher social status are more likely to report better health. The effect of gender may reflect cultural norms that influence how men and women assess and report their health, while the role of placement underscores the well-documented link between higher socioeconomic status and improved health outcomes, including access to healthcare resources, healthier living conditions, and overall life stability.

Conversely, age (−0.054) and weight (−0.149) are statistically significant negative predictors of health status. As individuals age, it is not surprising that they tend to report declining health, given the biological realities of aging and the higher likelihood of chronic health conditions. Similarly, higher body weight is associated with lower perceived health, a relationship that may reflect both physical health risks and the influence of cultural expectations around body image in many Asian societies.

In Stage 2 of the model, socio-demographic coefficients remain relatively stable, while information factors begin to play a more pronounced role. In particular, doctor information (0.054) becomes a statistically significant positive predictor of health status. This suggests that access to and trust in medical professionals contributes to a greater sense of physical well-being. By contrast, Stage 3 reveals no statistically significant predictors among the added government or urban planning variables. This lack of effect suggests that structural or policy-level interventions do not significantly shape self-reported health status once socio-demographic and informational variables are taken into account.

Table 4. Health status (non-Asia).

			Stage 1			Stage 2			Stage 3
Category	Variable	Standardized	t	VIF	Standardized	t	VIF	Standardized	t	VIF
Socio-demographic factors	gender	* 0.072	4.345	1.543	*** 0.068	4.080	1.554	*** 0.067	4.035	1.554
	age	*** −0.049	−3.678	1.016	*** −0.048	−3.590	1.018	*** −0.045	−3.382	1.021
	edu	0.012	0.828	1.095	0.016	1.130	1.108	0.021	1.508	1.118
	job	*** 0.099	7.067	1.106	*** 0.099	7.046	1.107	*** 0.096	6.848	1.111
	marriage	** −0.031	−2.226	1.106	** −0.030	−2.160	1.108	* −0.027	−1.936	1.113
	placement	*** 0.184	13.227	1.085	*** 0.183	13.198	1.086	*** 0.189	13.564	1.098
	family	*** 0.070	4.035	1.711	*** 0.068	3.908	1.715	*** 0.067	3.827	1.720
	children	0.019	1.087	1.690	0.019	1.115	1.692	0.019	1.117	1.693
	urban	0.009	0.682	1.013	*** 0.010	0.752	1.017	0.011	0.822	1.019
	height	*** 0.063	3.673	1.670	*** 0.065	3.775	1.678	*** 0.059	3.430	1.687
	weight	*** −0.230	−14.141	1.480	*** −0.226	−13.918	1.486	*** −0.224	−13.813	1.488
Information factors	vaccine				*** 0.049	3.613	1.042	*** 0.048	3.492	1.046
	docinfo				−0.006	−0.457	1.057	−0.006	−0.415	1.062
	relinfo				0.008	0.615	1.046	0.005	0.400	1.051
Government factors	shutdown							−0.016	−0.880	1.802
	surveillance							0.010	0.578	1.565
	masks							*** 0.056	3.088	1.857
	adj r2			0.104			0.106			0.109
	N			5033			5033			5033

Note: The notation *** p < 0.01 ** p < 0.05 * p < 0.1 represents statistical significance levels. It indicates the degree to which results are unlikely to have occurred by random chance, with *** showing the highest level of significance (the 99% confidence intervals), ** indicating a lower level of significance (the 95% confidence intervals), and * showing the lowest level of significance (the 90% confidence intervals). Cronbach’s alpha is not applicable to single-item measures because alpha assesses internal consistency among multiple items, requiring inter-item correlations.

highlights that in the health status model for non-Asian countries, most socio-demographic factors—specifically gender (0.072), job status (0.099), placement (0.184), the number of family members (0.070), and height (0.063)—are statistically significant positive predictors of health status, indicating that men, employed individuals, those with higher socioeconomic standing, individuals living in larger households, and taller respondents are more likely to perceive their health positively. In contrast, age (−0.049), marital status (−0.031), and weight (−0.230) act as negative predictors, with older age, being married, and higher body weight associated with lower reported health. These findings reflect a complex interplay of biological, social, and cultural dynamics in shaping health perception, with weight in particular standing out as a strong negative influence.

In Stage 2 of the model, while the overall effect of socio-demographic variables remains largely consistent, additional factors emerge. Urban areas (0.010) become statistically significant, indicating that individuals living in urban environments report slightly better health than their rural counterparts, possibly due to greater access to medical services and infrastructure. Importantly, two information factors also enter the model: vaccine knowledge (0.049) and reliable information (0.008) shows a positive association with health status, suggesting that individuals who are informed about vaccines reliable information are more likely to feel healthy.

By Stage 3, the significance of urban areas and reliable information disappears, and mask usage emerges as a statistically significant positive predictor of health status (0.056). This shift suggests that tangible, government factors—like wearing masks—may become more relevant than socio-demographic and information factors once the model accounts for all levels of explanatory variables.

Table 5. Wellness model (Asia).

			Stage 1			Stage 2			Stage 3
Category	Variable	Standardized	t	VIF	Standardized	t	VIF	Standardized	t	VIF
Socio-demographic factors	gender	−0.025	−1.219	1.291	−0.022	−1.063	1.302	−0.023	−1.112	1.302
	age	−0.009	−0.456	1.062	−0.011	−0.597	1.065	−0.008	−0.418	1.067
	edu	0.022	1.134	1.145	0.014	0.697	1.206	*** 0.015	0.760	1.209
	job	*** −0.067	−3.588	1.066	*** −0.069	−3.709	1.072	*** −0.068	−3.653	1.073
	marriage	** 0.043	2.276	1.101	** 0.045	2.369	1.102	*** 0.049	2.618	1.107
	placement	* −0.034	−1.840	1.063	* −0.031	−1.663	1.066	** −0.037	−1.979	1.071
	family	*** 0.084	3.676	1.594	*** 0.080	3.528	1.596	*** 0.080	3.535	1.597
	children	−0.030	−1.311	1.627	−0.029	−1.261	1.636	−0.036	−1.570	1.644
	urban	*** −0.108	−5.717	1.093	*** −0.105	−5.541	1.102	*** −0.102	−5.408	1.105
	height	−0.034	−1.523	1.496	−0.032	−1.436	1.496	−0.030	−1.348	1.498
	weight	*** 0.048	2.213	1.421	** 0.045	2.090	1.436	* 0.037	1.729	1.445
Information factors	vaccine				*** 0.068	−3.657	1.059	*** 0.072	−3.899	1.063
	docinfo				−0.028	−1.507	1.043	−0.024	−1.303	1.048
	relinfo				*** 0.057	3.080	1.040	*** 0.050	2.708	1.050
Government factors	shutdown							0.010	0.451	1.436
	surveillance							0.009	0.391	1.530
	masks							*** 0.099	4.452	1.534
	adj-r2			0.026			0.033			0.040
	N			2977			2977			2977

Note: The notation *** p < 0.01 ** p < 0.05 * p < 0.1 represents statistical significance levels. It indicates the degree to which results are unlikely to have occurred by random chance, with *** showing the highest level of significance (the 99% confidence intervals), ** indicating a lower level of significance (the 95% confidence intervals), and * showing the lowest level of significance (the 90% confidence intervals). Cronbach’s alpha is not applicable to single-item measures because alpha assesses internal consistency among multiple items, requiring inter-item correlations.

highlights that in the wellness model for Asian countries, the Stage 1 model identifies both positive and negative predictors among socio-demographic variables. Positive predictors include marriage (0.043), number of family members (0.084), and weight (0.048), suggesting that being married, having larger families, and higher body weight are modestly associated with better wellness. These results may reflect cultural contexts in many Asian societies where familial support and marital stability are strongly linked to emotional and social well-being. The positive association between weight and wellness may also reflect differing cultural body norms, where higher weight is not necessarily stigmatized and may even be associated with affluence or health in certain contexts. On the other hand, several socio-demographic factors serve as negative predictors of wellness. These include job status (−0.067), placement (−0.034), and urban residence (−0.108). The negative coefficient for job and placement may reflect work-related stress or dissatisfaction that diminishes personal well-being. This finding could indicate that in certain Asian settings, increased social pressure or expectations accompanying higher placement might reduce subjective well-being. The negative effect of urban living further supports this interpretation, possibly pointing to the stressors of dense, competitive, and environmentally burdened urban environments.

In Stage 2 of the model, the socio-demographic effects remain largely stable, while new variables related to health information begin to show influence. Vaccine (0.068) and reliable information (0.057) emerge as positive predictors of wellness, suggesting that awareness and access to trustworthy information enhance individuals’ sense of well-being. In Stage 3, education (0.015) becomes statistically significant, implying that higher levels of formal education are associated with improved wellness. Additionally, masks become a strong positive predictor of wellness (0.099), underscoring the role of protective behavior in contributing to a sense of safety and well-being, particularly in the post-pandemic context.

Table 6. Wellness model (non-Asia).

			Stage 1			Stage 2			Stage 3
Category	Variable	Standardized	t	VIF	Standardized	t	VIF	Standardized	t	VIF
Socio-demographic factors	gender	*** −0.098	−5.724	1.543	*** −0.091	−5.341	1.554	*** −0.09	−5.304	1.554
	age	−0.006	−0.404	1.016	−0.006	−0.441	1.018	−0.008	−0.609	1.021
	edu	*** −0.045	−3.122	1.095	*** −0.052	−3.595	1.108	*** −0.056	−3.918	1.118
	job	*** −0.096	−6.637	1.106	*** −0.093	−6.515	1.107	*** −0.091	−6.367	1.111
	marriage	0.000	−0.021	1.106	0.000	−0.029	1.108	−0.003	−0.233	1.113
	placement	*** −0.121	−8.477	1.085	*** −0.121	−8.555	1.086	*** −0.126	−8.878	1.098
	family	0.005	0.274	1.711	0.009	0.528	1.715	0.011	0.633	1.720
	children	0.000	0.009	1.690	−0.002	−0.135	1.692	−0.003	−0.151	1.693
	urban	** 0.033	2.394	1.013	** 0.029	2.092	1.017	** 0.028	2.041	1.019
	height	*** −0.05	−2.836	1.670	*** −0.052	−2.943	1.678	*** −0.047	−2.666	1.687
	weight	*** 0.144	8.647	1.480	*** 0.135	8.157	1.486	*** 0.134	8.065	1.488
Information factors	vaccine				*** 0.102	7.324	1.042	*** 0.101	7.260	1.046
	docinfo				*** 0.056	4.004	1.057	*** 0.056	3.986	1.062
	relinfo				−0.003	−0.196	1.046	−0.005	−0.393	1.051
Government factors	shutdown							0.027	1.500	1.802
	surveillance							−0.013	−0.742	1.565
	masks							** 0.038	2.048	1.857
	adj-r2			0.053			0.068			0.070
	N			5033			5033			5033

Note: The notation *** p < 0.01 ** p < 0.05 represents statistical significance levels. It indicates the degree to which results are unlikely to have occurred by random chance, with *** showing the highest level of significance (the 99% confidence intervals), ** indicating a lower level of significance (the 95% confidence intervals). Cronbach’s alpha is not applicable to single-item measures because alpha assesses internal consistency among multiple items, requiring inter-item correlations.

demonstrates that in the wellness model for non-Asian countries, several socio-demographic variables are statistically significant, indicating both positive and negative associations with wellness. Among the positive predictors, weight (0.144) and urban residence (0.033) are associated with higher levels of self-reported wellness. The positive coefficient for weight may seem counterintuitive from a medical standpoint, but it may reflect subjective perceptions of body image or satisfaction with appearance and health in specific cultural or economic contexts. The positive effect of urban residence could be linked to better access to services, entertainment, and healthcare, which may enhance overall quality of life in non-Asian countries.

Conversely, several negative predictors significantly reduce wellness. These include gender (−0.098), education (−0.045), job status (−0.096), placement (−0.121), and height (−0.050). The negative coefficient for gender likely indicates that women may report lower levels of wellness compared to men, a trend that could reflect gendered differences in emotional health, caregiving burdens, or labor market inequalities. Surprisingly, higher education and higher socioeconomic placement are negatively associated with wellness, which may be interpreted as a reflection of heightened stress, professional pressure, or unmet expectations among individuals in more privileged positions.

In Stage 2, the impact of socio-demographic variables remains largely stable, and two information variables become statistically significant: vaccine (0.102) and doctor information (0.056). Both variables serve as positive predictors of wellness, indicating that individuals who are informed about vaccinations and who receive reliable information from medical professionals tend to report higher levels of well-being.

In Stage 3 of the model, no significant changes are observed for socio-demographic or informational factors, but masks become a statistically significant positive predictor of wellness (0.038). This result reflects the psychological benefit of engaging in preventive health behaviors. Individuals who adopt mask-wearing may experience lower anxiety regarding exposure to illness, contributing to an improved sense of personal safety and emotional stability.

Table 7. Happiness model (Asia).

			Stage 1			Stage 2			Stage 3
Category	Variable	Standardized	t	VIF	Standardized	t	VIF	Standardized	t	VIF
Socio-demographic factors	gender	0.002	0.074	1.291	0.002	0.093	1.302	0.001	0.026	1.302
	age	0.002	0.126	1.062	0.003	0.160	1.065	−0.001	−0.067	1.067
	edu	*** −0.076	−3.985	1.145	*** −0.078	−3.975	1.206	*** −0.082	−4.177	1.209
	job	0.007	0.368	1.066	0.007	0.359	1.072	0.003	0.160	1.073
	marriage	*** 0.052	2.757	1.101	*** 0.052	2.758	1.102	** 0.044	2.366	1.107
	placement	*** 0.166	8.995	1.063	*** 0.165	8.950	1.066	*** 0.169	9.194	1.071
	family	−0.036	−1.594	1.594	−0.036	−1.590	1.596	−0.033	−1.467	1.597
	children	0.013	0.562	1.627	0.012	0.506	1.636	0.019	0.839	1.644
	urban	*** 0.111	5.917	1.093	*** 0.111	5.905	1.102	*** 0.110	5.886	1.105
	height	0.013	0.600	1.496	0.013	0.594	1.496	0.010	0.440	1.498
	weight	*** −0.075	−3.503	1.421	*** −0.073	−3.422	1.436	*** −0.064	−2.978	1.445
Information factors	vaccine				−0.007	−0.364	1.059	−0.005	−0.266	1.063
	docinfo				0.015	0.841	1.043	0.010	0.536	1.048
	relinfo				0.007	0.387	1.040	−0.004	−0.247	1.050
Government factors	shutdown							*** 0.088	4.133	1.436
	surveillance							−0.002	−0.095	1.530
	masks							** 0.052	2.373	1.534
	adj-r2			0.048			0.047			0.060
	N			2977			2977			2977

Note: The notation *** p < 0.01 ** p < 0.05 represents statistical significance levels. It indicates the degree to which results are unlikely to have occurred by random chance, with *** showing the highest level of significance (the 99% confidence intervals), ** indicating a lower level of significance (the 95% confidence intervals). Cronbach’s alpha is not applicable to single-item measures because alpha assesses internal consistency among multiple items, requiring inter-item correlations.

exhibits that in the happiness model for Asian countries, in the Stage 1 model, several socio-demographic factors emerge as significant predictors of happiness. Specifically, education (−0.076) and weight (−0.075) are negatively associated with happiness, while marriage (0.052), placement (0.166), and urban area residence (0.111) have a positive association. These results suggest that being married, having a higher social position, and living in urban areas contribute to greater happiness. In contrast, individuals with higher levels of education or body weight report lower happiness, which may reflect pressures or health concerns associated with these characteristics.

In the Stage 2 model, there are no meaningful changes among socio-demographic factors, and none of the information variables show statistically significant effects on happiness. This indicates that in the context of Asian countries, happiness appears relatively insulated from health information exposure or literacy, at least within the structure of this model. In the Stage 3 model, the influence of socio-demographic and information factors remains largely stable, but two government variables become statistically significant. Specifically, shutdown policies (0.088) and masks (0.052) are both positively associated with happiness. These findings underscore the role of visible and decisive public health interventions in promoting a sense of safety and psychological well-being.

Table 8. Happiness model (non-Asia).

			Stage 1			Stage 2			Stage 3
Category	Variable	Standardized	t	VIF	Standardized	t	VIF	Standardized	t	VIF
Socio-demographic factors	gender	* 0.032	1.925	1.543	* 0.029	1.735	1.554	* 0.029	1.746	1.554
	age	0.014	1.037	1.016	0.014	1.068	1.018	0.014	1.030	1.021
	edu	0.018	1.262	1.095	0.021	1.477	1.108	*** 0.021	1.456	1.118
	job	*** 0.055	3.903	1.106	*** 0.054	3.866	1.107	*** 0.055	3.932	1.111
	marriage	*** 0.114	8.080	1.106	*** 0.114	8.106	1.108	*** 0.114	8.061	1.113
	placement	*** 0.253	18.105	1.085	*** 0.253	18.098	1.086	*** 0.252	17.932	1.098
	family	0.004	0.223	1.711	0.002	0.142	1.715	0.002	0.104	1.720
	children	0.028	1.596	1.690	0.028	1.624	1.692	* 0.029	1.656	1.693
	urban	−0.008	−0.619	1.013	−0.007	−0.539	1.017	−0.008	−0.579	1.019
	height	** −0.041	−2.388	1.670	** −0.040	−2.306	1.678	** −0.039	−2.236	1.687
	weight	** −0.032	−1.967	1.480	* −0.029	−1.789	1.486	* −0.030	−1.822	1.488
Information factors	vaccine				*** 0.037	2.704	1.042	*** 0.038	2.792	1.046
	docinfo				−0.017	−1.218	1.057	−0.018	−1.282	1.062
	relinfo				−0.008	−0.556	1.046	−0.007	−0.537	1.051
Government factors	shutdown							−0.017	−0.942	1.802
	surveillance							0.012	0.733	1.565
	masks							0.018	0.989	1.857
	adj-r2			0.096			0.098			0.098
	N			5033			5033			5033

Note: The notation *** p < 0.01 ** p < 0.05 * p < 0.1 represents statistical significance levels. It indicates the degree to which results are unlikely to have occurred by random chance, with *** showing the highest level of significance (the 99% confidence intervals), ** indicating a lower level of significance (the 95% confidence intervals), and * showing the lowest level of significance (the 90% confidence intervals). Cronbach’s alpha is not applicable to single-item measures because alpha assesses internal consistency among multiple items, requiring inter-item correlations.

highlights that in the non-Asia happiness model, in the Stage 1 model, several socio-demographic variables emerge as significant predictors of happiness. Positive associations are found with gender (0.032), job status (0.055), marriage (0.114), and placement (0.253), indicating that being male, having employment, being married, and holding a higher socioeconomic position contribute to greater happiness. Conversely, height (−0.041) and weight (−0.032) are negatively associated with happiness, suggesting that individuals with greater height or weight report slightly lower happiness, which may be related to health concerns or social perceptions in these contexts.

In the Stage 2 model, socio-demographic factors remain largely stable without major shifts, while vaccine knowledge emerges as a significant positive predictor (0.037) of happiness. This suggests that greater awareness and understanding of vaccination correlate with increased happiness, possibly reflecting a sense of security and confidence in health protections.

By the Stage 3 model, education (0.021) and children (0.029) become newly significant positive predictors of happiness, implying that higher educational attainment and family size contribute positively to subjective well-being in non-Asian countries. Interestingly, no government factors show significant influence on happiness in this stage, indicating that public health policies may play a less direct role in shaping happiness compared to socio-demographic and informational variables in non-Asian settings.

5. Discussion

This study assesses how socio-demographic, informational, and governmental factors associate with health status, wellness, and happiness in Asian and non-Asian countries, using three-stage hierarchical regression frameworks. The approach enables identification of which predictors retain significance after accounting for other categories of variables, providing a clearer view of the relative contribution of each set of factors.

For health status in Asia, none of the government-related variables remain significant in the final stage once socio-demographic and information factors are included. Gender (0.071) and placement (0.201) retain significance, indicating that individual-level characteristics explain more variance in perceived health status than the governmental measures considered here. This pattern suggests that, within the Asian sample, the presence or communication of specific government interventions does not substantially alter perceptions of health status once foundational demographic and informational influences are taken into account.

In the non-Asia health status model, mask usage remains significant (0.056) even after controlling for earlier-stage predictors. This finding points to a measurable statistical association between a specific government-led protective behavior and reported health status in these settings, contrasting with the results from Asia where no government variables are retained. While the data do not establish causality, the persistence of mask usage in the model indicates a stronger link between this visible measure and health perceptions outside Asia.

For wellness, the Asia model shows a positive association with mask usage (0.099) and education. The significance of mask usage suggests that adherence to protective behaviors links to higher perceived wellness, while the role of education may reflect how background resources and information-processing capacity contribute to maintaining well-being during a health crisis. In the non-Asia wellness model, mask usage (0.038), vaccine-related information (0.101), and doctor-related information (0.056) serve as significant predictors. The presence of multiple significant information-related variables in non-Asia highlights the importance of access to trusted medical information in shaping wellness perceptions, in contrast to the Asia model where only masks and education retain significance.

For happiness, the Asia model shows that shutdown measures (0.088) and mask usage (0.052) are both positively associated with reported happiness. This indicates that, within this dataset, certain government measures link with higher happiness rather than declines, suggesting that these interventions may coexist with a sense of safety or order. In contrast, the non-Asia happiness model does not retain any government variables as significant in the final stage, with socio-demographic and information factors accounting for the variation in happiness. This absence points to a reduced statistical link between government policy measures and happiness in these contexts once other variables are included.

6. Conclusions

This study examines the associations between socio-demographic characteristics, information-related factors, and government interventions with self-reported health status, wellness, and happiness during COVID-19 context. The analysis uses comparative data from Asian and non-Asian countries, drawn from the ISSP 2021–Health and Health Care II. Three-stage hierarchical regression frameworks isolate the unique statistical contribution of each set of variables, enabling identification of both cross-regional similarities and region-specific differences. Because the dataset is cross-sectional and based on self-reports, the results are interpreted as correlational and descriptive rather than causal.

Several key findings emerge. In the Asian models, government-related variables are more often retained as significant predictors for wellness and happiness than for health status. This pattern suggests that, within the statistical limits of the model, subjective well-being and happiness in these contexts align more closely with government interventions than perceived physical health. In non-Asian models, however, government variables—particularly mask usage—are more strongly associated with health status than with wellness or happiness, indicating a different alignment between public health interventions and health perceptions.

Information factors also display distinct regional associations. In Asian countries, these variables are more closely tied to health status, with vaccine information and reliable medical advice entering as significant predictors. In non-Asian countries, information variables more often link to wellness and happiness, implying that credible information contributes more directly to psychological well-being and life satisfaction in these contexts, though the mechanisms remain undetermined from the present data.

Socio-demographic variables maintain a consistent presence in both regional models, though the strength and scope of their effects vary. Gender, occupation, and education prove significant in several models, with economic placement emerging as a consistently strong predictor across all outcomes and regions. The persistent role of placement underscores its association with health satisfaction regardless of cultural or institutional setting.

These findings require cautious interpretation. First, all variables derive from self-reported survey responses, which capture perceptions rather than objective measures of health and well-being. Such perceptions may diverge from clinically verified conditions and vary systematically across demographic and cultural groups. Second, while the ISSP follows rigorous standards for survey design and translation, country-specific psychometric properties—such as reliability and construct validity—of the health-related items are not publicly reported, limiting certainty about measurement equivalence. Third, the dataset lacks measures of within-country variation in ideology, risk perception, or access to healthcare, all of which could influence the observed associations. Finally, the cross-sectional design precludes conclusions about causal direction or temporal sequencing.

Future research should address these limitations by incorporating longitudinal designs to track changes over time, combining survey data with objective health measures to validate perceptions, and including qualitative or attitudinal components to better understand how individuals process public health information and interventions. These steps enable stronger inferences about the processes underlying the observed associations. Thus, the present study provides a descriptive baseline for further, methodologically diverse investigations into post-pandemic health perceptions across varied cultural and policy contexts.