New observations on the perceptual mechanisms at the rear of very long-term alterations in people’s neighborhood mobility: Evidence from COVID-19 in China

The necessity and frequency of travel has not returned to pre-pandemic concentrations

Figure 3a displays adjustments in 5 perception dimensions and self-report frequencies at 6 areas across the four levels. Following the outbreak of the epidemic, phase2 travel was restricted by guidelines, and the self-described journey frequency dropped significantly, which is regular with many studies. As the epidemic has been brought under regulate from the next to the fourth phase, residents’ perceived possibility has dropped appreciably, and residents’ controllability and willingness have enhanced. Higher perceived risk was involved with lower willingness, necessity, and self-noted journey frequency (see Supplementary Tables S3-S8 on-line).

With the liberalization of policies, the frequency of travel has risen from stage2 to stage4. The necessity of 6 places decreased in stage2 and increased from stage2 to stage4. But in stage4, when the confirmed cases have been cleared and travel is not restricted by the policy, the necessity and frequency are lower than stage1. Compared with stage 1, the common necessity and frequency loss from stage 2 to stage 4 is shown in Fig. 3b. The difference in stage4 is still significant, implying that the changes in necessity and frequency are asymmetric. Although the pandemic has returned, the necessity and frequency of travel has not.

In addition, the perception differences among the six regions also show certain regularities. Indoor leisure venues are considered the highest risk venues, with the lowest controllability, willingness, necessity and frequency. Shopping places are the most necessary places.

This is similar to many previous studies.Study Divides Social Distancing Behaviors in Pandemics into Policy-Induced and Voluntary Behaviors^{twenty four}. And people still maintain voluntary social distancing behaviors after the epidemic disappears, such as reducing the frequency of travel²⁵choose a location with less risk²⁶choose a safer means of transportation²⁷Risk perception influences adherence to COVID-19 social distancing restrictions²⁸.

Necessity is the main factor directly affecting travel intentions under persistent threats

We use the GLMM model to explore the influencing factors of willingness change. We found that the willingness to travel to the six places changes significantly at different stages. As shown in Table 2, necessity was the main factor affecting willingness (value> 25%, p < 0.001). Risk perception is also an important factor affecting the willingness to travel to the six places. Perceived risk had a significant negative impact on willingness, except for shopping venues. Attention to policy has a positive effect on intention. Controllability affects the public's willingness to go to life service places.

Demographic factors influencing travel intentions vary widely. Age significantly affects residents’ willingness to go to restaurants, leisure places (indoor and outdoor), and life service places. Household size influences willingness to eat restaurants and outdoor recreation. Willingness for leisure places (indoor and outdoor) and residential places are also affected by the vaccine. Sports venues are less affected by demographic factors.Demographic factors found in previous studies also varied in their impact on perception^29,30. However, many studies have also found no significant impact of demographic factors on travel intentions³¹Therefore, based on the instability of demographic factors, we did not consider the influence of demographic factors in the subsequent SEM analysis, but focused on the relationship between perception and behavior, which does not mean that the influence of demographic factors is not important.

New observations on the role of risk perception in linking government risk management and public behavior

Figure 4 shows that the influence of controllability and focus on necessity has shifted from being largely ignored to being properly considered. This creates the effect of these two factors, amplifying the impact on frequency of visits not only by attention but also by necessity. This suggests the development of a functional relationship between government risk management and public risk perception. Public attention to government policy has a consistent and significant positive effect on perceived risk, but the effect is smaller when the pandemic is contained. While the impact of public belief in the controllability of COVID-19 on perceived risk has changed fundamentally. With the implementation of the zero-COVID policy, the impact of COVID-19 controllability beliefs on perceived risk has shifted from positive to negative.

It also demonstrates changes in the ways in which public risk perceptions influence their behaviour. Although risk perception always plays a decisive role in individuals’ travel decisions, the pathways of impact differ between a pandemic outbreak and a zero-COVID environment. However, the effect of perceived risk on willingness has shifted from a direct effect to an indirect effect when the environment is zero COVID. Both the effect of risk perception on necessity and the effect of necessity on willingness are strong. Furthermore, we find that a unit’s risk perception has a stronger blocking effect on willingness in a zero-COVID environment than during the pandemic.

New observations on risk perception pathways: a candidate theory for long-term voluntary social distancing

The correlation analysis of perceived risk, necessity and willingness changes from stage2 to stage4 (Figure 4). The mechanism for determining the frequency of visits has changed. The necessity of the three stages has an impact on the willingness. From stage2 to stage4, the influence of risk perception on willingness decreases significantly, and the influence of risk perception on necessity increases significantly. It can be seen that the influence of risk perception on visit willingness has changed from a direct influence to an indirect influence of necessity.

After the pandemic resumed, the impact of perceived risk changed. In stage2 and stage3, risk has a significant impact on willingness. Willingness depends more on perceived risk. However, in Phase 4, while the impact is still there, it is lower than during the pandemic. The relationship between necessity and risk has changed 2-4 times compared to before the pandemic. Willingness is not directly affected by risk but by necessity, which has a significant effect on willingness at all three stages. Necessity in the post-pandemic period will be determined more by risk than in the pre-epidemic phase. The results show that the effect of perceived risk on necessity is lagged and asymmetric. One reason is that perceived risk has a limited impact on necessity. If the magnitude and mechanism of the impact of risk on perceived risk remain unchanged, people’s willingness to visit will not rebound even if the risk is reduced to its original position. This asymmetry dictates that while perceived risk has recovered, frequency has not fully recovered. Therefore, even if the perceived risk of a person in Stage 4 is reduced, the effect on frequency is bound to be amplified.

It offers a novel explanation for persistent voluntary social distancing. Since the start of the coronavirus pandemic, statistics on community mobility have revealed permanent and profound changes in the psychological dynamics behind individual behavior. There is evidence that visitor numbers to businesses and workplaces have yet to return to pre-pandemic levels despite the government lifting mandatory social distancing orders, even in areas with no COVID cases.The literature confirms that risk perception determines behavior under the threat of a possible accident¹⁵. However, the motivation for preventing people from visiting commercial establishments and areas of the workplace with no COVID cases remains a mystery. The theory of changing roles in risk perception offers a candidate theory for durable voluntary social distancing.

Our findings suggest that it is the mechanism by which public risk perception determines travel willingness that changes that causes travel frequency to fail to recover in an environment free of COVID cases. The impact of risk perception is even multiplied as the role of risk perception shifts from direct influencer to indirect influencer. On the contrary, if risk perception has always directly affected travel behavior, then changes in travel behavior with perceived risk should be symmetrical. For example, we compared the frequency of stage 3 and stage 4 in two scenarios: the real scenario in which the influence of risk perception changes from direct to indirect, and the counterfactual scenario mechanism stage in which risk perception always directly determines will. (Fig. 5) In the counterfactual scenario, each unit increase in perceived risk resulted in a 0.62-fold decrease in stage 3 weekly frequency. In the realistic scenario, the frequency decreased by 0.69 times per week, which is 13% more than in the counterfactual scenario. In stage4, the difference between counterfactual and real scenarios is even amplified. A one-unit increase in perceived risk directly leads to a 0.13-unit decrease in necessity, which in turn leads to a 0.43-per-week decrease in frequency in real scenarios. In the counterfactual scenario, a one-unit increase in perceived risk only reduced weekly frequency by a factor of 0.19 at stage 4. A unit increase in risk perception leads to a 2.3-fold reduction in trips due to a change in the mechanism by which risk perception affects frequency.

Thus, the risk-perceived effect of preventing travel is amplified by shifting from direct to indirect effects. Moderate risk perceptions would lead to a significant reduction in travel during COVID-zero periods, which would lead to voluntary social distancing, compared to during the pandemic outbreak.