Introduction

Population restrictions are among the most widely used non-pharmaceutical interventions (NPIs) implemented by governments around the globe to stem the spread of the SARS-CoV-2 virus, and its disease Covid-19. These restrictions affect completely different areas of everyday life such as sports, cultural offerings, wearing masks, or travel restrictions. In addition, different restrictions apply in different countries and states. Thus, this study examines the acceptance of these measures in the population. Furthermore, personal factors such as age, gender, and income and their impact on people’s view on the pandemic are examined.

Literature review

Some studies examined the measures against Covid-19 and their impact on the evolution of the pandemic. Vermeulen et al. (2021) used an agent-based simulation to identify the influence of different measures on the evolution of the pandemic. In their study, micro- as well as macro-level measures contribute to its control. Heiden and Buchholz (2020) also examined the impact of different parameters such as the immunity of the population and measures against Covid-19 on the evolution of the pandemic. Also here, the measures could slow down the evolution.

In contrast, Lippold et al. (2022) used a simulation with differential equations. Their results indicate that by cancelling major events, closing schools, and restricting contacts, the daily infection rate could be successfully slowed down and an uncontrolled evolution of the pandemic could be prevented. The results of Howard et al. (2021) confirmed that wearing masks reduces the transmissibility per contact in both laboratory and clinical contexts. Vrugt et al. (2021) used a simulation to show that face masks and hygiene measures could compensate for contact restrictions.

The studies above showed that measures are helpful to contain the pandemic. However, this does not necessarily mean that the population is happy with them. Martinez-Garcia et al. (2021) used an online survey in Spain to examine the factors that influence the unwillingness to be confined during the COVID-19 pandemic. Here, the unwillingness to be in confinement increased over time and women were less likely able to sustain long-term confinement. Furthermore, their results indicate that the acceptance of confinement was mainly influenced by psychological factors at the beginning of the pandemic, while these were replaced by economic factors toward the end of their study period. Similarly, Rosman et al. (2021) figured out that concerns about the virus were positively related to the acceptance of the measures. At least in part, concerns about economic consequences correlated negatively with the acceptance of the measures. They could also partially support their statement that concerns about the virus are decreasing over time, while concerns about the economic consequences are not increasing over time.

According to the results of Collins et al. (2021), political identity is more important than personal threats when assessing the pandemic. The study by Gollwitzer et al. (2021) indicated that the duration of the lockdown affected respondents’ reactions much more than the intensity or flexibility.

Brown et al. (2023) used a nationally representative sample of 496 participants during the first lockdown in the UK, and here the people were less likely engaged in healthful behavior when they perceived uncontrollable risks to their health. Kundu et al. (2021) conducted an online-based cross-sectional survey among 1765 adults in Bangladesh. Here, the respondent’s knowledge of Covid-19 was high (14.49 on a scale 0..17) and their COVID-19 knowledge score was significantly associated with a lower likelihood of negative attitudes. Schmidt et al. (2023) used an online survey with 1570 participants in Germany. Here, the overall COVID-19-related health literacy was high with, an average value of 37.4 (scale 0.50). In addition, COVID-19-related health literacy and knowledge about COVID-19 were slightly lower among men, migrants, people with low subjective social status, and those with low education. Governmental requirements and recommendations were rated as more effective by women, older people, and individuals with a chronic illness. DiClemente et al. (2022) collected data on COVID-19-related knowledge, awareness and adoption of preventive practices, depression and anxiety, stress, pessimism, and tobacco and alcohol use. In their study, COVID-19 knowledge and protective practice awareness were high, and COVID-19-related anxiety and depression were associated with increased drinking and smoking.

Also, the seasonal variation in transmissibility modulates the pandemic (Neher et al. 2020). Other topics related to Covid-19 research include, e.g., Dullen et al. (2020) who have found that one month of lockdown reduces the annual economic performance by 1%, in which half of it is reduced by the measures and the other half by reduced demand and disrupted supply chains. Furthermore, Radon et al. (2020) studied infection rates in the population by blood tests.

Methodology and hypothesis

The survey was conducted in Germany from 26 December 2020 to 14 February 2021 and was distributed via the university’s website and Facebook advertisements. A total of 1484 adults completed the survey. The data was evaluated with SmartPLS 3.0 and Bootstrapping was performed with 5000 samples. Figure 1 shows the main measurement model with the first 13 hypotheses. They are listed in Appendix 1 and are summarized in Table 2. The respective items (questions) of the latent variables are shown in Table 1. If no item is shown, the item equals its latent variable. All questions were measured on a quasi-metric Likert scale 1...5 (e.g., disagree, disagree on trend, neither agree nor disagree, agree on trend, agree; or: measures are exaggerated, exaggerated on trend, neither exaggerated nor appropriate, appropriate on trend, appropriate).

Fig. 1
figure 1

Measurement model including the first 13 hypotheses. + and – signs indicate a positive or negative correlation, respectively

Full size image
Table 1 Measures and their acceptance in the population on a scale 1...5
Full size table

In this model, hypotheses 10b, 11b, 12b, and 13b are moderation effects and presume that the statements “the vaccination has not been sufficiently tested” (10b), “the vaccination is not necessary for me” (11b), “the vaccination is not necessary for all people” (12b), and “the vaccination should be given to risk patients first” (13b) only impact the approval of vaccinations if, additionally, Covid-19 is rated as dangerous.

The measurement items of the latent variables are summarized in the results section in Table 1. For the descriptive statistics, all items are shown, including those removed from the main measurement model. Formative items were removed from the model following the recommendations of Hair et al. (2021): first, all formative items with ViF > 5 were gradually eliminated due to the collinearity. These were negative testing for flights (ViF = 14), mandatory masks when shopping (ViF = 10), closed theatres, musicals, ballet, opera (ViF = 6.5), mandatory masks at doctor’s practises (ViF = 5.6), closed cinemas (ViF = 5.4), no alcohol at Christmas markets (ViF = 5.4), negative testing when traveling abroad (ViF = 5.2), and mandatory masks at train stations (ViF = 5.2). Afterwards, further items were removed if their weight was not significant at the 95% level (p > 0.05) and, additionally, their loading was <0.5. Regarding the Covid-19 measures, only the item “no father’s presence when giving birth” was removed (p = 0.4, loading = 0.4). Further items to be removed here belonged to the “affectedness” (wearing masks (p = 0.9, loading = 0.3); limited travel possibilities (p = 0.6, loading = 0.4)) and “personal restrictions” (limited travel possibilities (p = 0.7, loading = 0.4); limited visiting of people in hospitals (p = 0.7, loading = 0.4)).

After evaluating the main measurement model, the impact of the respondent’s age, gender, education, city size, state/country, employment situation, household income, information sources, and personal experiences with Covid-19 on their opinions were evaluated using ANOVAS and t-tests. For this, the factor values of the latent variables were calculated with the loadings of their items. The corresponding hypotheses are shown in the results section and Appendix 1.

Results

Descriptive statistics

The descriptive results, showing the average approval of the respective statements, are summarized in Table 1. The statements are grouped according to the structural equation model shown in Fig. 1. Additionally, the loadings and weights with their respective latent variables are shown, and whether the variable was removed from the model due to low values.

The highest approval was for “quarantine for infected people,” “disinfecting hands when entering doctor’s practises, schools and shops” and for “wearing masks” at old people’s and nursing homes, doctor’s practises and in public transport vehicles. In contrast, the lowest approval was found for curfews 20–5 h, prohibited visits in hospitals, closed playgrounds, and prohibited father’s presence when giving birth. These results are obvious as the approved measures show a relatively high contribution to the pandemic control with low limitations.

Structural equation model

The results of the structural equation model are shown in Fig. 2 and Table 2. Further values are shown in Appendix 2, while the evaluation of the latent variables with their reflective or formative items is shown in Appendix 3.

Fig. 2
figure 2

Influence of the employment situation on the approval that Covid-19 is a dangerous disease

Full size image
Table 2 The first 13 hypotheses and their coefficients. Findings are supported if the algebraic sign of the path coefficient matches the causality of the hypothesis and, additionally, f2 > 0.02
Full size table

It can be seen that the following hypotheses are supported: the opinion that Covid is dangerous increases the support for all measures (H1) and decreases the opinion that the measures are worse than Covid (H2). Likewise, the opinion that the measures are worse than Covid reduces the support for all measures (H3). Furthermore, an increasing approval that the measures are a restriction for all people (H4) and the respondents themselves (H5) increase the opinion that the measures are worse than Covid itself. This indicates that personal restrictions have a higher impact than the restrictions for other people. In contrast, being more affected by the measures (H6) and feeling an increasing impact on one’s own life situation, fitness, psyche, and job situation (H7) do not increase the opinion that the measures are worse than Covid itself.

An increasing approval that Covid is dangerous also increases the fear of the disease (H8) and the approval of vaccinations (H9). Increasing opinions that the vaccines have not been adequately tested (H10) and that vaccination is not necessary for the person themselves (H11) decrease the support of vaccinations and there is also a moderating effect of the classification of Covid as dangerous on this dependency. In contrast, the opinion that the vaccination is not necessary for all people reduces the approval of vaccinations (H12+), but there is no moderating effect on the classification of Covid as dangerous (H12–). Furthermore, the opinion that risk patients should be vaccinated first does not affect the approval of vaccinations. Also, no moderating effect with the classification of Covid as dangerous can be observed here (H13).

Hypotheses with ANOVAs and t-tests

In addition, t-tests and ANOVAs were used to check whether age and gender have an impact on the life situation, fitness, psyche, ratings if Covid is dangerous and the measures are worse than Covid, the fear of Covid, the ratings on personal restrictions and restrictions for others, the respondents affectedness, the impact on the person and the opinion on the measures against Covid (H14). No impact could be found here (all p > 0.05). Likewise, an influence of the respondent’s city size and state did not show an influence on the approval that Covid is a dangerous disease (H15; city size p = 0.147 and for all states p > 0.05).

In contrast, people with higher education increasingly support that Covid is a dangerous disease (H16). Those who left school without a full high school degree supported this statement on average with 2.54 while those who completed high school or university rated an average of 3.23 (scale 1...5; p < 0.000).

Also, the employment situation influences the agreement that Covid is a dangerous disease (H17), as shown in Fig. 2 (p < 0.000).

In contrast, an impact of the household income on the approval that Covid is a dangerous disease (H18) could not be verified (p = 0.117). However, the respondent’s sources on information show an impact on the approval that Covid is a dangerous disease (H19). The results are summarized in Fig. 3 and show that people reading alternative press in particular are sceptical that Covid is dangerous. The p-values are information from TV (p < 0.000), radio (p = 0.004), newspapers (0.007), internet except social networks (p = 0.013), social networks (p < 0.000), friends (p < 0.000), and alternative press (p < 0.000).

Fig. 3
figure 3

Impact of the information sources on the approval that Covid is a dangerous disease

Full size image

Likewise, people who suffered a financial loss from the measures increasingly state that the measures are worse than Covid itself (H20, p < 0.00). The results are summarized in Fig. 4.

Fig. 4
figure 4

Impact of a personal financial loss on the opinion that the measures are worse than Covid itself

Full size image

In contrast, the hypothesis that people who received financial support to compensate for a loss are less likely to state that the measures are worse than Covid itself (H20) could not be verified (p = 0.23).

The hypothesis that people knowing a person who has tested positive increasingly state that Covid is a dangerous disease (H22) could be verified (p < 0.00). Here, the approval values are 2.37 (people not knowing a person who has tested positive) and 3.14 (people knowing a person who has tested positive).

In contrast, the hypothesis that people who had contact with people who tested positive increasingly state that Covid is a dangerous disease (H23) could not be verified (p = 0.924).

The hypothesis that people who were in quarantine increasingly state that Covid is a dangerous disease (H24) could be verified (p < 0.00). (Without quarantine 2.74, with quarantine 3.47.)

In contrast, the hypothesis that risk patients increasingly state that Covid is a dangerous disease (H25) could not be verified (p = 0.552).

Summary

This study examined the impact of personal experiences with Covid-19 as well as personal factors on the perception of the Covid-19 pandemic. A higher level of approval was found for measures with a major contribution to containing the pandemic with low restrictions on the people. In addition, the evaluation of the structural equation model showed that people classifying the pandemic as dangerous show a higher level of agreement with all measures and also rate the measures as worse than the pandemic itself. It was also shown that personal experiences in particular such as a quarantine and personal restrictions changed the personal opinion. Age and gender, on the other hand, had no influence on the view of the pandemic.

Policy

These results could be used for the assessment of restrictions as it might be helpful to consider public opinion to increase policy acceptance.

Further investigations

Another area of investigation could be a similar survey to determine the evolution of people’s opinions after the pandemic. The influencing factors could also be considered here. Furthermore, a comparison between people’s opinions and simulations of the measures could lead to beneficial results.