Authenticity Mediates the Relationship between Risk Perception of COVID-19 and Subjective Well-Being: A Daily Diary Study

Abstract

The present research investigated whether risk perception of COVID-19 relates to subjective well-being and the mediating role of authenticity in this association. We conducted a 12-day daily diary study with 133 undergraduates (M_age = 19.9 years, SD = 1.27 years; 64 females). Participants self-reported risk perception of COVID-19, authenticity, and subjective well-being every day. Results revealed that (1) risk perception of COVID-19 was negatively related to subjective well-being at the interindividual level; (2) authenticity mediated the relationship between risk perception of COVID-19 and subjective well-being at the interindividual level but not at the intraindividual level. In general, findings suggested that risk perception of COVID-19 is negatively related to subjective well-being only at the interindividual level, and authenticity plays a mediating role in this relationship. The finding suggested that keeping authenticity is a good strategy for avoiding the disruption caused by COVID-19. Longitudinal studies on samples with a broader age range, larger sample size, and extended sociodemographic background, as well as experimental studies, should be conducted to explore the causal relationship among interested variables that the current research has not detected.

1. Introduction

The Corona Virus Disease 2019 (COVID-19) is a tremendous risk event that influences individuals’ physical health, behavior, and mental health worldwide. Most people follow the protective advice and rules of governments, taking measures such as social distancing, and mask-wearing to protect themselves and others. Still, others do not take such behavior seriously [1]. One of the most predictive factors of this difference is people’s risk perception of the COVID-19 (RPC). RPC is the affective and cognitive reflection of COVID-19 events and is perceived as the key driver of those mental and behavioral outcomes [2,3]. The reduction of well-being is frequently explored due to risk perception recently [4,5]. For instance, Krok and Zarzycka examined 226 southern polish health care workers using the questionnaire method. The result illustrated that RPC is a significant predictor of psychological well-being [5]. However, the influence of RPC on individuals’ subjective well-being (SWB) has rarely been explored. Moreover, a few studies used cross-sectional data to explore their relationship, ignoring that RPC may cause the fluctuation of well-being within one person across days. Furthermore, the mechanism of the influences is not fully unfolded.

To solve the above problems, the current study intends to use the daily dairy study method to examine the relationships among the three variables at two levels. Without a theoretical foundation and statistical validation, generalizing experimental research findings from one level to another is improper and could result in incorrect conclusions (such as ecological fallacy) [6]. Additionally, each relationship has a specific meaning at different levels. For example, the negative relationship between RPC and SWB at the interindividual level means a person with a higher trait RPC has a lower trait SWB than another person’s SWB who has a lower trait RPC. Meanwhile, at the intraindividual level, the negative relationship between RPC and SWB means that, for a given person, when their state RPC is high, their state SWB is also low. For the rationale above, it is plausible to separate hypotheses on the intraindividual and interindividual levels.

1.1. RPC and SWB

There are two viewpoints about the stability of risk perception. Some studies found that risk perception is stable (trait risk perception) and co-variates with demographic characteristics and personalities, such as gender and optimism [7,8]. For instance, the experimental study found that although people changed their attitude about the nuclear power plant, their risk perception remained the same [9]. As a trait-like variable, risk perception is significantly related to negative emotions, such as anxiety, which are also believed to be indicators of SWB [10]. For instance, Ding and colleagues conducted an online investigation of the relationship between risk perception and depression with 1115 Chinese individuals. The results show that cognitive and affective risk perception is positively related to depression [11]. Xu and colleagues conducted a questionnaire survey of 408 Chinese adult students. The results showed that environmental risk perception has a direct negative relationship with their SWB [12]. Self-determination theory, which emphasizes the role of basic psychological needs, namely autonomy, competence, and relationships in individuals’ well-being, can explain the above findings [13]. Specifically, the high RPC limits an individual’s daily activities, which can weaken autonomy and relationships that are important for well-being. From the above experiment, studies, and the self-determination theory, we propose the hypotheses as follows.

Hypothesis 1.

RPC has a negative relationship with SWB at the interindividual level.

Other studies empathize with the malleability of risk perception, namely state risk perception. Risk-related factors, such as knowledge of COVID-19 and involuntary exposure to RPC, can induce a fluctuation in RPC at the within-person level [14]. Similarly, SWB can also have the characteristics of state-like and trait-like [15]. The state SWB can be influenced by daily incidents, such as emotional and cognitive factors [16]. Besides, from previous experimental studies of risk perception on the subject, there has been a similar relationship pattern of risk perception on outcomes at interindividual and intraindividual levels. For the above reasons, the state-like RPC may affect one’s daily SWB.

Hypothesis 2.

RPC has a negative relationship with SWBat the intraindividual level.

1.2. The Mediating Role of Authenticity

According to earlier studies, authenticity is a trait-like concept that refers to how closely one acts under one’s true self. Authenticity is defined as acting according to one’s inner thoughts and feelings [17]. According to studies, numerous dispositional traits have been linked to trait authenticity [17,18]. According to attachment theory [19,20], those with attachment security have a dispositional feature that can encourage the development of adaptive coping with threats and stressors. In contrast, those with attachment insecurity tend to magnify threats and stressors, including risk perception. According to the rationale above, RPC as a personal trait may be associated negatively with authenticity.

According to the self-determination theory, authenticity reflects a high degree of autonomy. It is one of the vital dispositional facilitators of the fundamental psychological needs that might increase subjective well-being [21]. According to prior research, authenticity as a personal quality has also been linked to increased psychological and subjective well-being, including happiness and self-acceptance, and decreased anxiety, depression, and stress [22,23]. Based on these arguments, we proposed the following hypothesis.

Hypothesis 3.

Authenticity mediates the negative relationship between RPC and SWB at the interindividual level.

In recent studies, authenticity is also regarded as a state, which is the instant feeling of the degree of alignment with one’s true self [24]. Furthermore, experimental investigations support this viewpoint [25]. Several situational factors have been shown to influence authenticity. For instance, a library of experimental studies illustrated those psychological states, such as positive moods, autonomy, and attachment security, had increased state authenticity. Negative moods and attachment insecurity have decreased state authenticity [26,27].

Attachment theory also provides a framework for thinking about the relationship between risk perception and authenticity at the intraindividual level, in that attachment security is also a state-like variable [28]. Specifically, people exposed to security situations are more inclined to induce a secure attachment, making them more open and authentic. In contrast, insecurity situations, such as risk, can hide an individual’s security attachment which in turn causes an individual to be inauthentic [24]. Based on attachment theory, it is plausible to speculate that an individual’s sense of authenticity can be reduced by many antecedent variables, such as external judgment, the feeling of threat, risk, and other adverse events [28]. Many experimental studies also suggested that the variance of authenticity can be influenced by contextual factors, such as culture and the internet [29,30]. From the above statement, the prevalence of COVID-19 in China put everyone into a threatening context. The RPC is also an insecurity factor that fluctuates daily and may induce in-authenticity. Thus, it is plausible to assume that state RPC has a negative relationship with state authenticity, which is critical for state SWB. As a result, the relationship between state RPC and SWB might be explained by state authenticity. The following hypothesis has been put forward.

Hypothesis 4.

Authenticity mediates the negative relationship between RPC and SWB at the intraindividual level.

1.3. The Current Study

From the studies mentioned above, few studies have examined the complex relationships among RPC, authenticity, and SWB at both interindividual and intraindividual levels. Moreover, given that the number of infected cases changes daily, the RPC may change throughout the day. It is necessary to detect this changing relationship of the above variables. Based on the above consideration, we conducted a daily diary study to clarify the role of RPC and the authenticity of Chinese college students’ subject well-being in their everyday life. The first aim is to explore the negative association between RPC and subject well-being. The second aim is to verify the mediating role of authenticity in the relationship between RPC and SWB. Although the association directions may be similar within two levels according to the above theory and experimental studies, the association magnitude among variables may differ at different levels. Thus, we separated the hypotheses into intraindividual and interindividual levels. The conceptual model of this study is provided in Figure 1.

2. Materials and Methods

2.1. Participants and Procedure

The institution’s research ethics committee approved our study. Participants were informed of the purpose of the study and completed informed consent. They volunteered to participate and acknowledged that they were free to withdraw at any time. We conducted this study through an online survey platform in the spring semester (April 2021).

With the convenience sampling method, we recruited participants from five universities in three provinces of China (Hunan, Hubei, and Shandong province). None of them had physical or mental problems. The questionnaire also collected sociodemographic information (sex, age, language, education level). The recruited participants distributed the online questionnaire to classmates or other friends who study at universities. Specifically, students were informed of the purpose of the investigation. They were then asked to complete online surveys (<5 min), which included a total of 25 items (in Chinese) concerning all the study variables. All items were rated along a 6-point scale, ranging from 1 = not true at all, to 6 = very true. Before launching the survey, researchers previewed all online materials, and all questions had to be answered. In total, 217 undergraduates aged 18–22 years from five college participated in the study. All the participants speak mandarin and all of them are Han Chinese. Finally, 133 (61%) participants (M_age = 19.9 years, SD = 1.27; 64 females) provided complete data across all 12 days, and another group partially completed the questionnaire. The payment was distributed randomly on “WeChat” (a social networking software widely used in China), ranging from one to two dollars. Finally, 1596 data samples were collected. We conducted chi-square tests to evaluate these two groups’ significant differences in demographic characteristics. We found no significant differences (p > 0.05) between participants and those who did not finish the questionnaire on any variables, which illustrated that the missing data were random.

2.2. Measurement

Measurement of RPC. Following the psychometric paradigm of risk perception [31], the current study adjusted the Scale for Hurricane Risk, which is well established and has good validity (8 items, Likert scale from one to five) and can generalize to other domains of risk [32]. We changed the word” hurricane” to” COVID-19 disease” and added the world “today” at the beginning of the sentence for all the original items. We adjusted their responses on a six-point Likert scale, ranging from one to six, which is usually used in China. The adjusted scale includes affective and cognitive perception dimensions, to assess daily levels of RPC. An example of an item is: “Today, how do you rate the danger of COVID-19 disease?” The higher the average scores, the more risk perception they perceived. The daily Cronbach’s alpha range was from 0.75 to 0.88.

Measurement of authenticity. Referring to the research of Thomaes and colleagues [33], a daily authenticity scale with three items was used for the study. An example of the item is “Today, I was my true self.” The responses were collected on a Likert scale ranging from 1 = strongly disagree to 6 = strongly agree. The scale has been certificated to have good validity [33]. The daily Cronbach’s alpha range was from 0.72 to 0.93.

Measurement of SWB. Referring to the previous study [33], we choose emotion and life satisfaction as indicators of SWB [34,35]. In addition, to fully capture the participants’ everyday SWB, we summarized the emotions most frequently experienced by Chinese college students through interviewing method, and selected the emotions most frequently encountered every day. According to the saturation principle of interviewing method, 16 college students were interviewed, and 8 kinds of emotions were finally collected. One item was selected from the Satisfaction with Life Scale (SWLS) [35]. A nine items scale was finally composed for the measurement of daily SWB with the instruction: “Please choose the number that best fits your situation based on your experience today”. The nine items included five positive emotional adjectives (interesting, happy, valuable, fulfilling, and encouraging), three negative emotional adjectives (lonely, hopeless, and unfortunate), and a ninth item, “How satisfied are you with your life today.” The questionnaire was a Likert scale, ranging from one to six. Negative items are reverse scored. The daily Cronbach’s alpha range was from 0.83 to 0.89. We also included age and gender as control variables in the model, for predicting authenticity and well-being.

2.3. Data Analysis

Because all the data were collected using the attitude questionnaire, common method bias may exist. Firstly, we conducted Harman’s single factor test for estimating common method bias [36]. Secondly, we used Mplus Version 8.3 (Los Angeles, CA: Muthén & Muthén, USA) to test if the variables can aggregate to the interindividual level by calculating the ICCs of variables [37]. Thirdly, we found the Pearson correlation coefficients of variables and conducted a mixed-effect model analysis using MLmed macro Beta 2 in IBM SPSS statistics version 25.0 (Armonk, NY: IBM Corp, USA) for statistical analyses. To test the hypothetical model, parameters were estimated through REML estimation [38]. In the mixed-effect model, the variables of interest are hypothesized to exist at both the within-person and between-person levels for exploring the within and between-person variation.

3. Results

3.1. Harman’s Single Factor Test, Results of Intraclass-Coefficients (ICCs) and Correlations

We loaded all the variables to a single factor with no rotation using exploratory factor analysis (EFA) to test the common method bias for each test point. The single factor explains 19.3% to 22.6% of the variance (no more than 40%), which indicated that the data have no series common method bias [36]. The statistical magnitude ICCs were computed for the three variables. ICC1 can be interpreted as the proportion of variance explained by interindividual; ICC2 illustrates the reliability of the aggregation from intraindividual level to interindividual level; usually, more than 0.7 is acceptable. As expected, there were considerable intraindividual changes and acceptable reliability in the interindividual level for RPC (ICC1 = 0.78, ICC2 = 0.98), authenticity (ICC1 = 0.53, ICC2 = 0.93), and SWB (ICC1 = 0.47, ICC2 = 0.92). Bliese suggests ICC1 > 0.05 as a conventional criterion for aggregation [39]. Thus, all the variables can be aggregated to a between-person level. The group-mean centered score of all variables represents the intraindividual level score; the group-mean score of all variables represents the interindividual level score (the 12 days of data collected from one person who worked as a group). Pearson’s correlation coefficients of tested variables at both levels showed that RPC, authenticity, and SWB correlated with each other significantly at both levels (p < 0.01) (see

Table 1. Estimated correlation matrix for the latent variables.

Variables	1	2	3	4
1 Risk perception of COVID-19		−0.26 **	−0.23 **
2 Authenticity	−0.30 **		0.40 **
3 Subjective well-being	−0.24 **	0.43 **
4 Age	0.00	−0.02	0.25 **

Note: Up diagonal are the correlations at the intraindividual level—level 1; under diagonal are the correlations at the interindividual level—level 2; ** p < 0.01. Gender differences are not significant in variables 1, 2, and 3 with t test. The t scores were 0.06 (p > 0.05), 2.02 (p > 0.05), 1.65 (p > 0.05), respectively.

).

3.2. The Effect of RPC and Authenticity on SWB

MLmed macro Beta 2 in IBM SPSS statistics version 25.0 was used to account for the nesting of data, that is, daily data were nested in person. We assessed the direct impact of RPC on SWB and the mediating role of authenticity at both levels. Age and gender were induced as co-variables. In the regression model, all the variables collected daily are supposed to aggregate to the interindividual level (level 2), the intercept is random, and the slope is fixed. In the direct model, SWB regressed on RPC at both levels.

The regression coefficients showed that there was a significant impact of RPC on SWB at the interindividual level (level 2) (B = −0.98, p < 0.01), but insignificant at the intraindividual level (level 1) (B = 0.05, p > 0.05). Hypothesis 1 was supported, while hypothesis 2 was not supported.

At the interindividual level (level 2), RPC had a significant effect on authenticity (B = −0.28, p < 0.01). However, the effect of RPC on authenticity was insignificant at the intraindividual level (level 1).

At the interindividual level (level 2), authenticity had a significant effect on SWB (B = 1.98, p < 0.01). In the intraindividual level (level 1), the regression coefficient of SWB on authenticity was also significant (B = 1.68, p < 0.01). The regression coefficients of the model are shown in

Table 2. The regression models of all variables.

Path	B	SE	t	95% CI
Level 1
Risk perception of COVID-19 → Authenticity	0.00	0.04	0.06	[−0.07, 0.08]
Authenticity → Subjective well-being	1.68	0.14	11.75 **	[1.40, 1.96]
Risk perception of the COVID-19 → Subjective well-being	0.05	0.21	0.24	[−0.36, 0.47]
Level 2
Risk perception of COVID-19 → Authenticity	−0.28	0.07	−3.67 **	[−0.42, −0.13]
Gender → Authenticity	−0.23	0.264	−0.88	[−0.67, 0.20]
Age → Authenticity	−0.02	0.06	−0.28	[−0.14, 0.11]
Authenticity → Subjective well-being	1.98	0.42	4.81 **	[1.16, 2.79]
Risk perception of the COVID-19→ Subjective well-being	−0.98	0.36	−2.78 **	[−1.68, −0.27]
Gender → Subjective well-being	−0.81	1.41	−0.57	[−3.58, 1.96]
Age → Subjective well-being	0.94	0.30	3.13	[0.35, 1.52]

Note: B = unstandardized regression coefficient; ** = p < 0.01; SE = standard error of B; t = t test values of B; 95% CI = 95% certificate interval of B.

.

3.3. The Mediating Effect of Authenticity

The Monte Carlo samples method procedure was applied to examine the mediating effects of authenticity. Specifically, if the confidence interval generated by the Monte Carlo sampling includes zero, there are no mediating effects, and vice versa [30]. The results show that the mediating role of authenticity on the relationship between RPC and SWB is significant at interindividual level (level2) (B1 * B2 = −0.54, 95% CI = [−0.97, −0.23], p < 0.05) (see

Table 3. The mediating role of authenticity at two levels.

	B1 * B2	SE	z	95% MC CI
Level 1
Risk perception of COVID-19 → Authenticity → Subjective well-being	0.00	0.07	0.06	[−0.13, 0.14]
Level 2
Risk perception of COVID-19 → Authenticity → Subjective well-being	−0.54	0.26	−2.16 *	[−0.97, −0.23]

Note: B1 * B2 = coefficient of indirect path; SE = stander error of B1 * B2; z = z test values of B1 * B2; * p < 0.05; 95% MC CI = 95% confidence intervals gained by Monte Carlo samples method.

), but not the case in intraindividual level (level 1). Hypothesis 3 was supported, while hypothesis 4 was not supported according to the above results.

4. Discussion

Consistent with previous studies [5,40], our study also found that trait risk perception negatively affects trait SWB. The current study also tested the mediating role of authenticity at the interindividual level for the first time. The results that RPC was negatively associated with SWB at the interindividual level supported the view of the point proposed by attachment theory. The mediating role of trait authenticity between trait RPC and SWB also supports the statement of self-determination theory argued in our introduction. The results from the intraindividual level are not always in line with the hypotheses. Further studies are needed to explore these inconsistencies. Only state authenticity is significantly related to state SWB, which retested the robust positive role of authenticity on SWB [21,41].

Our findings expanded the understanding of the influence of RPC on SWB at both between-person and within-person levels. Findings also shed light on the influence of RPC on individuals’ SWB and its mechanisms at the trait level. The inconsistent results between interindividual and intraindividual levels indicated that the interindividual to intraindividual generalizability of the relationship of these variables would result in mistaken conclusions [6].

4.1. Relationship between RPC and SWB

Researchers found a negative relationship between RPC and SWB at the interindividual level. This is like previous studies conducted with diverse cultural backgrounds and participants [5]. On the one hand, the findings supported the self-determination theory that unsafe mental factors will harm basic psychological needs, such as autonomy, which in turn harm well-being. Our results supported attachment theory, emphasizing the critical role of insecurity factors on an individual’s mental health and well-being [42]. On the other hand, when we used multilevel regression to further explore the predicted effect of RPC on SWB, the regression coefficient of SWB on RPC was insignificant at the intraindividual level. This finding illustrated that the daily changes of RPC cannot influence their daily SWB. In other words, the state RPC does not influence state SWB. There are possible reasons for this result as follows. The first reason may be that the COVID-19 was relatively stable and under control during the data collecting period in most areas of China. In this period, the RPC of most participants was kept at a low level, which could not influence SWB. The second reason may be that multicollinearity exists among independent variables, weakening the explanation power of RPC. The third reason can be implicated in the determination theory. For the above reason, the influence of RPC on SWB at the intraindividual level should be explained cautiously.

4.2. Relationship between RPC and Authenticity

The effect of RPC on authenticity was significant at the interindividual level, which suggested that as trait-like variables, people with high RPC are more inclined to have low authenticity. According to self-determination theory, a safe external environment can provide autonomous support, which is essential for the expression of authenticity [21]. The current findings supported the above theory to some degree. However, it is not the case at an intraindividual level. Given that little study focused on the relationship between risk perception and authenticity, it is hard to find a related theoretical explanation or experimental support. Therefore, the reason for that finding requires further investigation.

4.3. Relationship between Authenticity and SWB

There are significant positive relationships between authenticity and SWB at both interindividual and intraindividual levels. These findings align with most studies, illustrating that authenticity is a vital predictor of SWB [43,44]. These findings also align with the perception that authenticity is essential to healthy psychological development [45], given that our result illustrated that authenticity contributes to individuals’ SWB at the intraindividual level.

4.4. The Mediation Effect of Authenticity

The mediation analysis showed that authenticity mediated the relationship between RPC and SWB at the interindividual level, which is the most critical finding in our study. Specifically, high trait RPC may be associated with low trait authenticity, and the limit of trait authenticity is related to low trait SWB. When authenticity was added to the regression model, the direct association of RPC and SWB became insignificant, meaning trait authenticity may be the key reason for the close relationship between trait RPC and trait SWB. This finding is like many previous studies in which authenticity played the mediator role in the relationship between antecedence variables and SWB [46,47,48]. The finding also connected attachment theory with self-determination theory, in that the former theory emphasizes that trait authenticity is closely related to attachment insecurity and may be negatively associated with trait authenticity. The latter theory focuses on the role of authenticity on basic psychological needs, such as autonomy which are essential to SWB.

However, this mediation effect did not exist at the intraindividual level. In other words, for the same person, the daily change of PRC (state RPC) does not influence state authenticity and SWB. This result was unexpected, and the reason for this finding may be as follows. One explanation is that the variation of RPC derived from the intraindividual level can be divided into random error, and systematic change [49], and the small proportion of systematic change caused the magnitude of people to differ from themselves is smaller compared to people who differ from each other in our study. The small variation of RPC at the intraindividual level may make the mediation effect of authenticity unable to unfold. Another explanation is that the mechanisms involved at the interindividual level differ from those involved at the intraindividual level, which will be further examined in future studies.

4.5. Implications

Several theoretical implications of the current study should be noted. First, it integrated attachment theory and self-determination theory to effectively explain the relationship between RPC, authenticity, and SWB. Second, it expanded the antecedence factors of SWB to the field of people’s negative expectations (risk perception). Authenticity as an important trait-like and state-like variable benefits people’s SWB. Meanwhile, the current study also shed light on practice. First, people should remember that their risk perception of COVID-19 can weaken their SWB. Thus, they must ensure that risk perceptions correspond to reality and avoid amplifying them. Second, for people in a risky situation, maintaining their authenticity is a desirable choice to protect their mental health and SWB.

5. Limitations and Future Direction

Several limitations to this study that warrant further discussion, and the results of our study, should be interpreted cautiously. First, the current study sample is limited to Chinese undergraduates and has a limited sample size. It is necessary to explore individuals from different ages and cultural backgrounds to explore the generalization of the findings. Second, the data were collected during a stable period of the COVID-19, which resulted in RPC having limited variety. This may cause the study to find it hard to detect the actual relationship among all variables at the intraindividual level. Third, although the daily diary method was used to explore the effect of RPC on SWB and the mediator role of authenticity, the autoregression of PRC and authenticity were not taken into consideration because of the limitation of the multilevel method we conducted. The lag effect of independent variables on SWB has not yet been detected. Finally, the results of our study are limited to the exploration of correlation, in the absence of a control/control group or treatment that can examine causal effects. Experimental manipulation of RPC can be conducted to examine the causal relationship between variables. Although our study collected dense longitudinal data to explore the dynamic relationship between interested variables, the major hypothesis was not supported at the intraindividual level. Thus, more longitudinal studies should be conducted to examine the stability of the current study.

6. Conclusions

In conclusion, the current study discovered that RPC damages people’s SWB. Furthermore, at the interindividual level, authenticity mediated the link between RPC and SWB. Specifically, individuals with a high level of trait RPC can reduce their trait authenticity and, in turn, their levels of trait SWB. The findings expanded the theory of authenticity and SWB. The above findings also shed light on the practical work of improving people’s SWB during the COVID-19 period, such as reducing COVID-19 risk perception and enhancing authenticity.