Perceived Severity of the Coronavirus Disease 2019: An International Comparative Analysis

(1)

An International Comparative Analysis Margherita Angioni

Carlo Bo University of Urbino, Italy

R zvan Mihai B canuă ă

University of Bucharest, Romania

Fabio Musso

Carlo Bo University of Urbino, Italy

Abstract

The Coronavirus disease 2019 (COVID-19) which started in China in December 2019, has rapidly spread all over the world. Italy was the first European country to experience the outbreak in mid-February 2020. The virus has been spreading at different speed and timing in other European states, including Romania, which declared a state of emergency on 16th March. As there are no vaccines available, governments had toface the emergency by implementing lockdown in order to reduce the number of infections. The aim of this paper is to analyse the severity perceived by citizens regarding COVID-19 infection, through a comparative analysis between Romanians and Italians. Drawing on the theories of Health Behaviour, the perceived severity was measured through 8 items, subsequently reduced through an exploratory factorial analysis that allowed to identify two factors defined as

“Emotional reaction” and “Perceived consequences”. For each of the two factors, the correlation was measured both with the demographic variables (gender, age, level of education) and with other variables considered relevant (possibility of home working, perceived level of information on preventive measures, and self-reported adoption of preventive behaviour). The citizens who answered the online questionnaire were 1126 in Romania and 742 in Italy. Although the two countries were in different stages of the infection, and with different political actions implemented by the two governments, results showed numerous similarities in severity perception. Practical implications emerged for designing intervention programs by local and national governments.

Keywords: COVID-19, Perceived severity, Emotional reaction, Perceived consequences, Romanians, Italians Riassunto. Gravità percepita dell’epidemia da Coronavirus. Un’analisi comparativa internazionale

La malattia da coronavirus (COVID-19) iniziata in Cina a dicembre 2019, si è rapidamente diffusa in tutto il mondo, con l’Italia che è stato il primo Paese europeo a registrare un focolaio a metà febbraio 2020. Il virus si è diffuso con velocità e tempi differenti negli altri Stati europei, fra cui la Romania che ha dichiarato lo stato di emergenza il 16 marzo. Non essendo disponibile un vaccino, i governi hanno dovuto fronteggiare l’emergenza attuando il cosiddetto lockdown per ridurre il più possibile i contagi. Lo scopo di questo articolo è quello di analizzare la gravità percepita dai cittadini riguardo all’infezione da COVID-19, attraverso una analisi comparativa fra la popolazione rumena e quella italiana. Attingendo dalle teorie comportamentali relative alla salute (Health Behaviour Theories), è stata misurata la gravità percepita attraverso 8 elementi, ridotti poi a due in seguito a una analisi esplorativa fattoriale: la

“reazione emotiva” e le “conseguenze percepite”. Per ciascuno dei due è stata misurata la correlazione sia con variabili demografiche (genere, età, istruzione) sia con altre variabili rilevanti (possibilità di lavorare da casa, percezione del livello di informazione sulle misure preventive e adozione di comportamenti di prevenzione). Gli intervistati, attraverso un questionario on-line, sono stati 1126 in la Romania e 742 in Italia. Nonostante i due Paesi fossero in stadi differenti di diffusione del contagio e con differenti azioni pubbliche messe in atto, i risultati hanno evidenziato numerose similitudini. Ne sono derivate significative implicazioni pratiche per politiche di intervento sia a livello locale che nazionale.

Parole chiave: COVID-19, gravità percepita, reazione emotiva, conseguenze percepite, rumeni, italiani DOI: 10.32049/RTSA.2020.2.18

1. Introduction

Initially described as a pneumonia-like disease with unknown origins by the Chinese authorities who informed World Health Organisation China Country Office on 31^st

(2)

December 2019 (WHO, 2020a), the Coronavirus disease (COVID-19) has been officially classified as a pandemic on 11^thMarch 2020 (WHO, 2020b).

Caused by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), as it was latter named by the World Health Organisation (WHO, 2020c), the pandemic has its origins in the city of Wuhan from the People’s Republic of China (Lu, Stratton and Tang, 2020), early studies linking its origins with a local wet market (Wang, Tang and Wei, 2020; Wu et al., 2020). Preliminary results indicate that the virus has a natural origin (Andersen et al., 2020) and likely to have been transmitted from bats (Lillie et al., 2020; Xu et al., 2020; Ji et al., 2020), which were linked in the past with several other severe acute respiratory syndromes (Muller et al., 2005; de Wit et al., 2016; Bootz et al., 2017).

Due to its novelty, much is still unknown about its symptoms and means of transmission, however several emergency warning signs have been identified so far (CDC, 2020; WHO, 2020d), as well as the incubation period estimated at around 5-6 days and up to 14 days (WHO, 2020e). Although the WHO estimated a mortality rate of 2% (WHO, 2020f), individuals over 60 or suffering from affections like diabetes, cancer, or hypertension are considered at high-risk. Without an available vaccine in sight, people are advised to implement social distancing and self-quarantine if they suspect they had been in contact with an infected person and several countries decided to restrict movement in order to limit the spread of the disease. Nonetheless, whether the individuals will respect the authority’s recommendations depends on a large number of factors, including perceived severity.

Having presented the present context of the pandemic, we believe that the perceived severity will prove to be a key unit of measurement in determining the level of responsibility manifested by the individuals (Fishbein et al., 2001).

The aim of this paper is to analyse and explain the severity perceived by people facing Covid-19 infections, comparing the difference in terms of emotional reaction and perceived consequences between Romanians and Italians. In particular, the research intends to investigate whether and how, the gender, age, level of education, the possibility of working from home, the perceived level of information on preventive measures, the self-reported adoption of preventive behaviour, and the current state of health, impact on perceived

(3)

severity.

Worth noting is that the study was done in two countries at different stages of the pandemic, with Italy reaching its pandemic peak during the data gathering and Romania expecting to reach this point around the end of April (Voiculescu, 2020). We also feel obliged to mention that Romania did not face a nearly similar number of deaths (Ministerul Sănătăţii, 2020) compared with Italy (Italian Ministry of Health, 2020), when the two-time frames of the disease evolution are compared. These results are mostly attributed both to the fact that when the epidemic was discovered in Italy (21^stFebruary 2020), its severity was still unknown, and to the fact that in Romania very strict social distancing measures were imposed by the Romanian authorities as early as 16^thMarch when an emergency situation was declared (Romanian Parliament, 2020).

The structure of the rest of the paper is as follows: the next section illustrates the theoretical framework focusing on the concept of perceived severity in the main behavioural theories, section 3 presents data and the research method performed, and section 4 discusses the results obtained by comparison between Romania and Italy.

2. Study background: the concept of Perceived severity in Health Behaviour Theories

Perceived severity is defined as the psycho-social construct of the individual’s perceived likelihood to face a negative end-result depending on the actions taken at a specific moment (Rosenstock, 1974).

The origins of the perceived severity (also encountered as “perceived seriousness” in the literature) seem to be with the development of the Health Belief Model (HBM) in the 1960s in the United States, which aimed to understand the (lack of) efficacy for public healthcare programs (Janz and Becker, 1984). The model contains four dimensions: Perceived susceptibility (the perceived likelihood of facing negative effects on one’s health), Perceived severity (the construct covers the individual assumptions about the risk’s associated with the illness and includes the probability of facing medical as well as social negative outcomes),

(4)

Perceived benefits (capturing the gained benefits of following the recommended course of action and the likelihood of accepting it in relation with the understood danger), and Perceived barriers (the “costs” in regard with following the recommendation or health procedure, be it medical side effects, or other inconveniences).

The HBM is based on behavioural psychology, namely the conceptualization of two variables (Maiman and Becker, 1974; Janz and Becker, 1984): the importance of the outcome, and belief that a particular action will have a desirable end-result. According to this model, the probability of individuals choosing a specific course of action is reliant on the dimensions presented above, but also takes into consideration the cues to action (Rosenstock, 1974), which appears to be a structural necessity in the decision making process and its intensity highly dependent with individual context. Later, studies also showed that motivation is influenced interdependently by perceived probability and perceived severity (Weinstein, 2000), which are normally hard to observe by the researchers.

Furthermore, optimistic biases must also be taken into consideration when discussing health risks appraisals (Weinstein and Klein, 2015; Druică, Cosma and Ianole-Călin, 2020), with individuals holding the erroneous assumption that the likelihood of negative events is higher among others than themselves.

Protection motivation theory (PMT) is another framework developed for understanding human behaviour correlated with protective measure, the dimension of perceived severity playing a crucial role. Developed in 1975 by the American psychologist Ronald Rogers, the PMT is based on the assumption that the: «three crucial components of a fear appeal to be (a) the magnitude of noxiousness of a depicted event; (b) the probability of that event’s occurrence; and (c) the efficacy of a protective response» (Rogers, 1975), thus, the decided action is correlated with the perceived severity of the situation or behaviour, the vulnerability of the individual and the degree of counter-action which aims to ameliorate the situation. However, as Rogers notes: «fear appeals have been found to differ in their interest value, seriousness, importance and amount of concern elicited» (Rogers, 1975).

A third framework which worth to be mentioned is that developed in the 1990s by Kim

(5)

Witte. Based on the existent literature, including the one cited above, Witte (1992) argues that there are several reasons why fear appeals, «understood as persuasive messages designed to scare people by describing the terrible things that will happen to them if they do not do what the message recommends». He fails to provide consistent results and forwards several reasons for this: a terminological misunderstanding when two distinct terms are concerned, a lack of clarity when interpreting in reactions of the subject, and a lack of consistent in-depth analysis between threat and efficacy.

Another important aspect highlighted by the theoretical models is that demographic and socio-psychological variables may influence perceptions and, thus, indirectly influence health-related behaviour (Glanz, Rimer and Viswanath, 2008). Several studies in perceived severity (Kasmaei et al., 2014; Constant et al., 2005; Nau et al., 2005; Hunt et al., 1980) showed that variables as age, gender, educational level, and current health status have (or do not have) an indirect effect on behaviour by influencing the perception of severity.

In this study, in addition to the aforementioned demographic and socio-psychological variables, we felt it was important to take into consideration some variables that are lacking in the literature. Given the peculiarity of the pandemic situation, we believed that the possibility to work from home, the perceived level of information on preventive measures, and the self-reported adoption of preventive behaviour were useful for measuring perceived severity. In more detail, the possibility to work from home could give individuals the belief of continuing to live an almost normal life as the time that people must mandatorily spend at home is occupied with daily commitments. The level of information on preventive measures is an important factor for individuals because it can affect the management of the mood and fear of people in an emergency that they are not used to. Finally, the self-reported adoption of preventive behaviours in a situation of perceived seriousness is crucial for understanding the reaction of citizens, their sense of responsibility and their respect for the rules imposed by governments.

(6)

3. Data and research method

In this section, the research method is illustrated. In order to provide a clear explanation of the statistical analysis performed, the pre-processing of the data and the methodology are shown step by step.

3.1 Data pre-processing

Data collection was carried out through the administration of an online questionnaire, adopting a convenience sampling methodology (Kitchenham and Pfleeger, 2002) via Facebook and LinkedIn, and via e-mail and WhatsApp, from 14th March 2020 to 9th April 2020. This methodology, also known as availability sampling (Leiner, 2016), is frequently adopted for online questionnaires as it is fast, maximises the time-cost trade off and increases the size of samples, being respondents readily available (Schmidt and Hollensen, 2006). According to Heckathorn (2011), as the sample expands – through social networks in our case – it reaches an equilibrium that is independent of the convenience sample from which it started. Therefore, if the sample size reaches a large enough threshold value, it does not matter if the initial sample was non-random (Heckathorn, 2011).

1126 responses came from Romanians and 742 from Italians.

The questionnaire consisted of 15 questions: three for demographic aspects (Gender, Age, Level of education); eight for Perceived Severity (SEV1, SEV2, SEV3, SEV4, SEV5, SEV6, SEV7, SEV8) measured with a 7 point Likert type scale in terms of level of disagreement/agreement; one for the Possibility to work from home; one for Perceived level of information on preventive measures; one for Self-reported adoption of preventive behaviour, and one for the Health status.

Table 1 shows the range of values and the measurement scales of these variables.

(7)

Variables Range of values Measurement scale

Demographic

Gender Male, Female Nominal

Age 14, 15, 16, 17, ... Ordinal

Level of education

Primary or secondary education,

University or post- university education

Ordinal

Perceived severity

(SEV)

1. The thought of getting infected with

Coronavirus scares me 1 – 7 (Likert scale) Interval 2. When I think about Coronavirus,

my heart beats faster 1 – 7 (Likert scale) Interval 3. I am afraid even to think of

Coronavirus 1 – 7 (Likert scale) Interval

4. The problems that I would experience as a result of getting ill from Coronavirus would last for a long time

1 – 7 (Likert scale) Interval

5. Getting sick from Coronavirus would threaten my relationship with important people in my life

(boyfriend/girlfriend, team- mates, or parents)

6. Getting sick from Coronavirus

would threaten my work performance 1 – 7 (Likert scale) Interval 7. If I suffered from Coronavirus my

whole life would change 1 – 7 (Likert scale) Interval 8. If I got ill from Coronavirus I

would suffer consequences years from now

Possibility to work from home Yes, No Nominal

Perceived level of information on

preventive measures 1 – 10 (Likert scale) Interval

Self-reported adoption of preventive

behaviour 1 – 10 (Likert scale) Interval

Health Status

Lower than other people, The same as other people, Better than other people

Ordinal

Tab. 1 - Range of values and measurement scales of the variables

3.2 Hypotheses and methodology

Following data pre-processing, the statistical analysis was performed. After extracting the

(8)

subsets for Romania and Italy, the graphs of item distributions to compare the two countries were analysed. The visual inspection suggested that there were differences between countries in each case. However, since the visual inspection was not enough, the Wilcoxon Rank Sum Test was used to check whether the differences were statistically significant. This non-parametric test was chosen due to the fact that the distributions were not normally distributed. The hypotheses are presented below:

H1 There is a statistically significant difference between countries in the response to the sentence

«The thought of getting infected with Coronavirus scares me» (SEV1)

H2 There is a statistically significant difference between countries in the response to sentence

«When I think about Coronavirus, my heart beats faster» (SEV2)

«I am afraid even to think of Coronavirus» (SEV3)

«The problems that I would experience as a result of getting ill from Coronavirus would last for a long time» (SEV4)

“Getting sick from Coronavirus would threaten my relationship with important people in my life (boyfriend/girlfriend, team-mates, or parents)” (SEV5)

«Getting sick from Coronavirus would threaten my work performance» (SEV6)

«If I suffered from Coronavirus my whole life would change» (SEV7)

«If I got ill from Coronavirus I would suffer consequences years from now» (SEV8)

In order to identify whether the 8 variables that measure Perceived severity (SEV1, SEV2, SEV3, SEV4, SEV5, SEV6, SEV7, SEV8) could be reduced to fewer variables, a correlational analysis was performed. Since the correlations between the elements were high, pursuing data reduction made sense. Thus, subsequently, a Parallel Analysis (PA) with promax rotation was conducted to identify the number of components to be extracted.

After identifying two factors (Emotional reaction and Perceived consequences) that emerged from the Exploratory Factorial Analysis (EFA), and analysing the differences between factors within each country, further possible correlations were investigated. In more detail, the Wilcoxon Rank Sum Test was used to investigate the correlation, in each country,

(9)

between factor scores and the variables “Gender”, “Level of education”, and “Possibility to work from home”. The hypotheses are shown below:

H09 The emotional reaction depends on the gender of Romanian respondents

H10 The emotional reaction depends on the level of education of Romanian respondents

H11 The emotional reaction depends on the possibility to work from home of Romanian respondents

H12 The emotional reaction depends on the gender of Italian respondents

H13 The emotional reaction depends on the level of education of Italian respondents

H14 The emotional reaction depends on the possibility to work from home of Italian respondents H15 The perceived consequences depends on the gender of Romanian respondents

H16 The perceived consequences depends on the level of education of Romanian respondents H17 The perceived consequences depends on the possibility to work from home of Romanian

respondents

H18 The perceived consequences depends on the gender of Italian respondents

H19 The perceived consequences depends on the level of education of Italian respondents

H20 The perceived consequences depends on the possibility to work from home of Italian respondents

The Spearman’s Rank Correlation Coefficient was used to investigate the correlation, in each country, between factor scores and the variables “Age”, “Perceived level of information on preventive measures”, and “Self-reported adoption of preventive behaviour”.

The hypotheses are presented below:

H21 The emotional reaction depends on the age of Romanian respondents

H22 The emotional reaction depends on the perceived level of information on preventive measures of Romanian respondents

H23 The emotional reaction depends on the self-reported adoption of preventive behaviour of Romanian respondents

H24 The emotional reaction depends on the age of Italian respondents

H25 The emotional reaction depends on the perceived level of information on preventive measures of Italian respondents

H26 The emotional reaction depends on the self-reported adoption of preventive behaviour of Italian respondents

H27 The perceived consequences depends on the age of Romanian respondents

(10)

H28 The perceived consequences depends on the perceived level of information on preventive measures of Romanian respondents

H29 The perceived consequences depends on the self-reported adoption of preventive behaviour of Romanian respondents

H30 The perceived consequences depends on the age of Italian respondents

H31 The perceived consequences depends on the perceived level of information on preventive measures of Italian respondents

H32 The perceived consequences depends on the self reported adoption of preventive behaviour of Italian respondents

Lastly, after a preliminary visual inspection, the non-parametric Kruskal-Wallis Rank Sum Test and the Bonferroni Post-hoc Test were used to analyse the correlation between the factor scores and the variable “Health status”. The hypotheses are the following:

H33 There is a statistically significant difference across the health status categories of Romanian respondents and the emotional reaction

H34 There is a statistically significant difference across the health status categories of Italian respondents and the emotional reaction

H35 There is a statistically significant difference across the health status categories of Romanian respondents and the perceived consequences

H36 There is a statistically significant difference across the health status categories of Italian respondents and the perceived consequences

The hypotheses are summarized in Figure 1.

(11)

Fig. 1 – Conceptual model

4. Findings and discussion

Data analysis was performed with R software, version 3.4.3, considering a 0.05 significance level for all analyses.

4.1 Sample profile

Table 2 shows the main frequencies of the sample. The sample consists of 24.51% male and 75.49% female for Romania, and 38.14% male and 61.86% female for Italy. Data about Age were collected as continuous variable. The age of respondents is between 16 and 82

(12)

years old for Romanians and between 14 and 79 years old for Italians. Regarding the level of education, both the majority of Romanians and Italians are highly educated (67.76% and 65.09% respectively).

Romania Italy

n. % n. %

Gender

Male 276 24.51 283 38.14

Female 850 75.49 459 61.86

Min Max Min Max

Age 16 82 14 79

n. % n. %

Level of education

Primary or secondary education 363 32.24 259 34.91

University or post-university education 763 67.76 483 65.09

Tab. 2 - Sample profile

4.2 Descriptive statistic

Table 3 shows the minimum, first quartile, median, mean, third quartile, and maximum for each of the eight items of perceived severity (SEV).

Min 1^st Qu. Median Mean 3^rd Qu. Max

SEV1 1.000 2.000 4.000 3.908 6.000 7.000

SEV2 1.000 1.000 2.000 2.823 4.000 7.000

SEV3 1.000 1.000 2.000 3.001 4.000 7.000

SEV4 1.000 2.000 3.000 3.369 5.000 7.000

SEV5 1.000 2.000 4.000 3.911 6.000 7.000

SEV6 1.000 2.000 4.000 4.055 6.000 7.000

SEV7 1.000 1.000 3.000 3.114 4.000 7.000

SEV8 1.000 1.000 2.000 2.730 4.000 7.000

Tab. 3 - Descriptive statistic

(13)

4.3 Perceived severity plots and Wilcoxon Rank Sum Test

The Kernel density plots in Figure 2 compare Romania and Italy in terms of Perceived severity distributions. The Kernel density curve allows observing the shape of distribution more closely, considering a total area equal to one. This visual inspection suggests that there are differences between countries in each item.

As can be observed, Italians seem to be more worried than Romanians about the contraction of the Coronavirus (SEV1, SEV2, SEV3). Distributions relating to threats to relationships with important people, effects on work performance, and duration of consequences (SEV4, SEV5, SEV6, SEV7, SEV8) are asymmetric for both countries, but with very different peaks.

Fig. 2 - Perceived severity plots - Romania and Italy

However, as mentioned above, the visual inspection is never enough, therefore we use the

(14)

Wilcoxon Rank Sum Test to check whether the differences are statistically significant (Table 4).

Hypotheses Items W p-value Statistical

significance

H1 SEV1 358744 .000 accepted

H5 SEV5 423132 .632 rejected

Tab. 4 - Wilcoxon Rank Sum Test

Results in Table 4 show that the only item with no country differences is SEV5 which corresponds to the statement “Getting sick from Coronavirus would threaten my relationship with important people in my life (boyfriend/girlfriend, team- mates, or parents)” (H5, p-value = .632), thus H5 is rejected. All other hypotheses (H1, H2, H3, H4, H6, H7, H8, p-value < .05) are accepted. This means that there are statistically significant differences between Romanians and Italians for the items SEV1 (“The thought of getting infected with Coronavirus scares me”), SEV2 (“When I think about Coronavirus, my heart beats faster”), SEV3 (“I am afraid even to think of Coronavirus”), SEV4 (“The problem that I would experience as a result of getting ill from Coronavirus would last for a long time”), SEV6 (“Getting sick from Coronavirus would threaten my work performance”), SEV7 (“If I suffered from Coronavirus my whole life would change”), and SEV8 (“If I got ill from Coronavirus I would suffer consequences years from now”).

4.4 Advanced analysis

The aim of the advanced analysis is to identify whether the 8 variables measuring Perceived severity can be reduces to a lower number of variables. We begin with a

(15)

correlational analysis (Table 5).

ROMANIA ITALY

SEV1 SEV2 SEV3 SEV4 SEV5 SEV6 SEV7 SEV8 SEV1 SEV2 SEV3 SEV4 SEV5 SEV6 SEV7 SEV8

SEV1 1.00 SEV1 1.00

SEV2 .723 1.00 SEV2 .699 1.00

SEV3 .674 .810 1.00 SEV3 .832 .773 1.00

SEV4 .535 .538 .512 1.00 SEV4 .547 .470 .513 1.00 SEV5 .390 .362 .375 .444 1.00 SEV5 .390 .361 .411 .405 1.00 SEV6 .326 .279 .273 .411 .566 1.00 SEV6 .291 .244 .313 .423 .522 1.00 SEV7 .460 .459 .470 .554 .605 .546 1.00 SEV7 .422 .360 .437 .528 .575 .646 1.00 SEV8 .356 .370 .387 .575 .431 .385 .647 1.00 SEV8 .399 .379 .402 .569 .459 .512 .710 1.00

Tab. 5 - Correlational analysis

The correlations among items are high, so pursuing data reduction makes sense.

In order to measure the internal consistency, that is, how closely related a set of items are as a group, the Cronbach's Alpha is measured. By measuring the internal consistency of our data we find a high Cronbach’s Alpha values (= .88 for each country) that cannot be increased by dropping items. As known, a high value for Alpha does not imply that the measure is unidimensional, so we use the Exploratory Factor Analysis that is a good method to check dimensionality.

4.4.1 Parallel analysis: Emotional reaction and Perceived consequences

After examining the Kaiser's rule (eigenvalues >1) and producing the scree plots, we performed the Parallel analysis both for Romania and Italy. As the data is not normally distributed, we extracted the factors using the principal axis method and the promax rotation.

In both cases, we set the cut-off for factor loadings as .400. The reliability of both models is good: TLI = .926 for Romania and TLI = .955 for Italy; the RMSEA indices are around .90, slightly higher than the recommended value. Lastly, the cumulative variance explained in the case of Romania is 61%, and in the case of Italy is nearly 65%.

(16)

Table 6 shows the results of Parallel analysis both for Romania and Italy.

ROMANIA ITALY

PA1 PA2 PA1 PA2

SEV1 .737 SEV1 .888

SEV2 .991 SEV2 .835

SEV3 .885 SEV3 .965

SEV4 .474 SEV4 .445

SEV5 .728 SEV5 .573

SEV6 .730 SEV6 .795

SEV7 .839 SEV7 .937

SEV8 .654 SEV8 .749

Tab. 6 - Results of Parallel analysis

Given the items in each factor, we label factor 1 as “Emotional reaction” (SEV1, SEV2, SEV3), and factor 2 as “Perceived consequences” (SEV4, SEV5, SEV6, SEV7, SEV8).

Figure 3 shows the plots of the new factors by comparing Romania and Italy. The Kernel density plots regarding Emotional reaction and Perceived consequences show a different distribution for Romanians and Italians. In the first plot, the data distributions for the two countries seem quite similar, however a small peak for Italy is highlighted. This means that, on average, the emotional reaction of the respondents of the two countries is mainly similar:

both the majority of the Romanian and Italian respondents, indeed, responded with low scores that correspond to a greater level of disagreement with the proposed statements.

(17)

Fig. 3 - Emotional reaction and Perceived consequences between Romania and Italy

In the second plot, distributions are very different with a strong peak for Romania. In this case, the responses of the Italian respondents regarding the perceived consequences are distributed more homogeneously than those of the Romanian respondents. The latter, indeed, tend to be more optimistic.

The Kernel density plots in Figure 4 allow us to better observe the differences between Emotional reaction and Perceived consequences within each country.

(18)

Fig. 4 - Emotional reaction and Perceived consequences within Romania and Italy

By comparing the factor scores within countries, the first plot in Figure 3 highlights a peak for Perceived consequences compared to Emotional reaction for Romania’s data.

Conversely, in the second plot, Emotional reaction shows a higher peak compared to Perceived consequences for Italy’s data. This suggests that the emotional reaction of most Romanians is not of concern, however even those who are concerned are anyway optimistic about the future consequences of the virus. The opinion of Italians is different, indeed, even those who declare that they are not worried, perceive in a more negative way the consequences of a possible contraction of the COVID-19 infection. This result may depend on the different stage of contagion that the countries were experiencing at the time of the interview.

4.4.2 Non-parametric analysis: Wilcoxon Rank Sum Test and Spearman’s Rank Correlation Coefficient

In order to identify any correlations, in this section the differences between the new factors Emotional reaction and Perceived consequences within each country are analysed.

(19)

We use the Wilcoxon Rank Sum Test to investigate the correlation (hypotheses from H9 to H20), in Romania and in Italy, between factor scores and the variables “Gender”, “Level of education”, and “Possibility to work from home” (Table 7).

Emotional reaction Perceived consequences

Romania Italy Romania Italy

H p-value Sig. H p-value Sig. H p-value Sig. H p-value Sig.

Gender H9 .004  H12 .000  H15 .000  H18 .000 

Level of education H10 .880 X H13 .209 X H16 .268 X H19 .685 X

Work from home H11 .733 X H14 .791 X H17 .149 X H20 .746 X

= accepted; X= rejected Tab. 7 - Wilcoxon Rank Sum Test

The results in Table 7 highlight that the only statistically significant relation is with Gender (H9 p-value =.004; H12 p-value =.000; H18 p-value =.000) for both countries. This means that neither the Emotional reaction nor Perceived consequences depend on the level of education (H10, H13, H16, H19) and on the possibility to work from home (H11, H14, H17, H20).

Figure 5 shows that, both for Emotional reaction and Perceived consequences, women score higher than men. Although gender difference has been widely demonstrated in the medical and socio-psychological literature (Hess et al., 2000; Ross and Bird, 1994), in the particular case of Covid-19 this result is unexpected as data currently made known by all countries show that men are more affected by the virus than women (Chen et al., 2020; ISS, 2020).

(20)

Fig. 5 - Emotional reaction, perceived consequences, and gender in each country

Now, we test whether factors scores are correlated with the “Age”, the “Perceived level of information on preventive measures” (Information), and the “Self-reported adoption of preventive behaviour” (Prevention) (Table 8) using the Spearman’s Rank Correlation Coefficient.

(21)

H p-value Sig. H p-value Sig. H p-value Sig. H p-value Sig.

Age H21 .026  H24 .887 X H27 .986 X H30 .006 

Information H22 .025  H25 .016  H28 .000  H31 .163 X

Prevention H23 .179 X H26 .954 X H29 .029  H32 .012 

= accepted; X= rejected

Tab. 8 - Spearman’s Rank Correlation Coefficient

As shown in Table 8, Age is positively related to Emotional reaction in Romania (H21, rho = .0663), but it is unrelated in Italy (H24). For what concerns the Perceived consequences, Age is unrelated in Romania (H27), but is negatively related in Italy (H28, rho = -.1012).

The higher the level of Information related to preventive measures, the lower the Emotional reaction in both countries (H22, rho = -.0668; H25, rho = -.0885). Regarding the Perceived consequences, Information on preventive measures records a negative correlation in Romania (H28, rho = -.1190), but it is uncorrelated in Italy (H31).

Lastly, the Self-reported level of adoption does not seem related to Emotional reaction (H23 and H26). However, the higher the score for Perceived consequences, the higher the score for Self-reported preventive behaviour adoption (H29, rho =.0649; H32, rho = .0920).

4.4.3 Health status

In the questionnaire, the Health status was investigated with the question: “How do you assess your health status?”. Respondents could choose between “Lower than others people”,

“The same as other people”, or “Better than other people”.

From a preliminary visual inspection of the data (Figure 6), it is possible to note that respondents with a health status lower than others tend to score higher at the Emotional reaction. The same applies to Romania with respect to Perceived consequences. In Italy, however, those with perceived similar health status as others tend to score higher at

(22)

perceived consequences.

Fig. 6 - Health status plots – Romania and Italy

This is an unexpected result, but we know that visual inspection is not reliable, so we conduct statistical testing. Thus, in order to test whether there are differences, we apply the non-parametric Kruskal-Wallis Rank Sum Test.

χ² 4,9948 5,7891 15,767 11,826

df 2 2 2 2

p-value .082 .055 .000 .003

H H33 = X H34 = X H35 =  H36 = 

= accepted; X= rejected

Tab. 9 - Kruskal-Wallis Rank Sum Test for Health status – Romania and Italy

(23)

The results in Table 9 show that there are no statistically significant differences across Health status categories in terms of Emotional reaction (H33, p-value = .082 for Romania, and H34, p-value = .055 for Italy). Therefore, we only accept hypotheses relating to Perceived consequences (H35 and H36).

Lastly, the following Bonferroni post-hoc test will clarify which categories are different (Table 10).

Perceived consequences

Romania Italy

1 2 1 2

2 .077 - 2 1.000 -

3 .001 .015 3 .470 .002

Tab. 10 - Bonferroni post-hoc test – Romania and Italy

As shown in Table 10, in Romania the differences appear between category 2 (same level of health status as others) and 3 (better than others) as the p-value is .015. Also, there is a significant difference between category 1 (lower than others) and category 3 (p-value

= .001).

We conclude that the lower than health status, the higher the perceived consequences of infection but there are no significant differences between the perception of those with low health status (category 1) and those in the second category, with self-reported health status the same as other people (p-value = .077).

For what concerns the perceived consequences in Italy, the post-hoc test shows that the only statistically significant difference occurs between categories 2 and 3 (p-value = .002).

Those in category 2 scoring higher at perceived consequences than those in category 3.

5. Conclusions, limits, implications, and future research

Perception of severity, one of the fundamental factors of the Health Belief Model

(24)

developed in 1966 with the purpose to analyse health-promoting behaviour, has long been studied in the literature (Rosenstock, 1974; Janz and Becker, 1984; Weinstein, 2000;

Weinstein and Klein, 2015).

In this research, we have chosen to investigate the severity perceived by the population during the first phase of COVID-19, which was spreading worldwide. In this pandemic period, the aim was to analyse the perception of severity as a factor affecting people’s behaviour in relation to a series of variables, specifically gender, age, level of education, the possibility to work from home, the perceived level of information on preventive measures, and the self-reported adoption of preventive behaviours. The analysis was carried out by comparing the data collected simultaneously in two different European countries: Romania and Italy. The results obtained demonstrate several similarities between the two populations.

Although the data provided worldwide on the characteristics of COVID-19 patients tell us that the subjects most exposed to infection are men, in both countries the perceived severity both in terms of emotional reaction and perceived consequences records higher scores for women. This is certainly a fact that governments must take into account in their political and management decisions. Indeed, if it is true that mortality is more than double in the male gender, it is equally true that women are more exposed to the risk of being infected. Indeed, women in almost all the cultures of the world have always been the ones who run the house (Plant, 1997; Levinson, 2011): they buy groceries and take care of children and the elderly (Angioni and Musso, 2020). During this pandemic, supermarkets remain a fundamental place to go for food purchasing but at the same time, they are also the place where the risk of contagion is greatly elevated. In the same way, the risk of contagion is perceived as high even in contact with children - whom the television stations have often reported as possible healthy carriers of the disease - and with the elderly, the main victims of the virus. In this context, therefore, proper communication campaigns by public authorities should have been taken into account of these issues, particularly highlighting that women, despite having a lower predisposition of being infected, are more exposed to risks depending on behaviours and contacts occurred in daily activities.

Another interesting fact concerns age. While in Italy this factor does not seem to be

(25)

related to the emotional reaction, in Romania as the age increases the emotional aspect of perceived severity also increases, according to what previous studies reported about the relationship between age and optimism (Chowdhury et al., 2014). On the contrary, the age of Romanians does not impact on the perceived consequences, while in Italy it is young people who perceive the negative consequences of the infection more. It should be noted that in Italy, from the outset, the provisions given by the government focused on the social distancing referred in particular to the relationship between grandparents and grandchildren.

Young people have certainly suffered from this ban, amplifying the unease of the population. The communication focused on the necessary sacrifice, but the benefits of these sacrifices were probably not clearly communicated. In this area too, the practical implications suggest that public authorities' communication campaigns should be clearly addressed to population groups, making clearer the responsibility of individuals in ensuring collective behavior that could counter the spread of the epidemic.

Among the results obtained, it worth to be noted that, in this particular case, unlike the existing literature (Barron, Gamboa and Rodriguez-Lesmes, 2019; Ophir, 2019), the level of education does not affect the perceived severity in any of the countries analysed.

Communication once again becomes a central issue in the management of perceived severity as a determining factor of behaviour when analysing the relationship with the perceived level of information on preventive measures. The results show that in both countries as the perception of information on preventive measures increases, the emotional reaction of citizens decreases. In Romania, moreover, a high level of information lowered the perceived consequences (this result must be interpreted as positive indication since the related questions of items were placed using pessimistic terms). Thus, governments need to ensure updated scientific information and clear and transparent communication about their decision making. In this context, also mass media have a role, providing a connection between daily events and public perceptions (Lippmann, 1922). As Cohen (1963), and subsequently McCombs and Shaw (1972) stated in their agenda-setting theory, «press may not be successful much of the time in telling people what to think, but it has unexpected success in telling its readers what to think about. The world will look different to different

(26)

people, depending on the map drawn for them by writers, editors, and publishers of the paper they read».

Both Romanian and Italian citizens claim to adopt adequate prevention behaviour.

Indeed, the more the perception of the consequences from COVID-19 increases, the greater the commitment made by citizens to respect the anti-contagion rules. So, in this first phase, the prevention campaigns carried out by governments seem to have hit the target. However, it must be remembered that if the road to a return to normality is long, citizens are going to suffer increasingly. The risk is to go towards a relaxation of the measures that could once again raise the spread curve of the virus, frustrating all the sacrifices made so far. The suggestion is to implement sustainability-oriented policies (Musso, Esposito and Angioni, 2019), with new communication strategies, in order to keep in the long term citizens’

attention on the importance of the adoption of preventive measures also in the following phases.

The main limitation of this research is that the data were collected in two countries that are at different times and speeds of the spread of the infection. However, the results studied demonstrate numerous similarities in the perception of severity by Romanian and Italian citizens.

Furthermore, from the results that emerged, a fundamental aspect that we are keen to suggest for future research is to consider in depth the communication campaigns carried out by governments, including press conferences, legislative acts, and posts on the institutional profiles of social networks and analyse how these affect the perceived severity while studying citizens’ reactions and comments.

Reference List

Andersen K.G., Rambaut A., Lipkin W.I., Holmes E.C., Garry R.F. (2020). The proximal origin of SARS-CoV-2. Nature medicine, 26, 4: 450. DOI: 10.1038/s41591-020-0820-9.

Angioni M., Musso F. (2020). New perspectives from technology adoption in senior

(27)

cohousing facilities. The TQM Journal, 32, 3: 1. DOI: 10.1108/TQM-10-2019-0250.

Barron K., Gamboa L.F., Rodriguez-Lesmes P. (2019). Behavioural Response to a Sudden Health Risk: Dengue and Educational Outcomes in Colombia. The Journal of Development Studies, 55, 4: 620. DOI: 10.1080/00220388.2018.1425392.

Bootz A., Karbach A., Spindler J., Kropff, B., Reuter N., Sticht H., Mach M. (2017).

Protective capacity of neutralizing and non-neutralizing antibodies against glycoprotein B of cytomegalovirus. PLoS pathogens, 13, 8: 1. DOI: 10.1038/nm.3985.

Centers for Disease Control and Prevention - CDC (2020). Symptoms of Coronavirus.

Washington: CDC. Retrieved from: https://www.cdc.gov/coronavirus/2019-ncov/

symptoms-testing/symptoms.html (19/04/2020).

Chen N., Zhou M., Dong X., Qu J., Gong F., Han Y., Qiu Y., Wang J., Liu Y., Wei Y., Xia J., Yu T., Zhang X, Zhang L. (2020). Epidemiological and clinical characteristics of 99 cases of 2019 novel coronavirus pneumonia in Wuhan, China: a descriptive study. The Lancet, 395, 10223: 507. DOI: 10.106/S0140-6736(20)30211-7.

Chowdhury R., Sharot T., Wolfe T., Düzel E., Dolan R.J. (2014). Optimistic update bias increases in older age. Psychological medicine, 44, 9: 2003. DOI:

10.1017/S0033291713002602.

Cohen B. (1963). The press and foreign policy. New York: Harcourt.

Constant A., Castera L., de Ledinghen V., Couzigou P., Bruchon-Schweitzer M. (2005).

Psychosocial factors associated with perceived disease severity in patients with chronic hepatitis C: relationship with information sources and attentional coping styles.

Psychosomatics, 46, 1: 25. DOI: 10.1176/appi.psy.46.1.25.

de Wit E., van Doremalen N., Falzarano D., Munster V.J. (2016). SARS and MERS: recent insights into emerging coronaviruses. Nature Reviews Microbiology, 14, 8: 523. DOI:

10.1038/nrmicro.2016.81.

Druică E., Cosma, A., Ianole-Călin R. (2020). Frequency distributions concerning Romanians’ attitude towards preventive behavior – the Coronavirus case. RG project:

“Covid – 19 in data - Romania”. Retrieved from: https://www.researchgate.net/project/

Covid-19-in-data-Romania (05/06/2020).

(28)

Fishbein M., Triandis H.C., Kanfer F.H., Becker M., Middlestadt S.E., Eichler A. (2001).

Factors influencing behavior and behavior change. In Baum A., Revenson T.A. Singer, J.E., eds., Handbook of Health Psychology. Mahwah, NJ, US: Lawrence Erlbaum Associates Publishers.

Glanz K., Rimer B.K., Viswanath K. (2008). Health behavior and health education: theory, research, and practice. Hoboken, NY: John Wiley & Sons.

Heckathorn D.D. (2011). Comment: Snowball versus respondent-driven sampling.

Sociological methodology, 41, 1: 355. DOI: 10.1111/j.1467-9531.2011.01244.x.

Hess U., Senécal S., Kirouac G., Herrera P., Philippot P., Kleck R.E. (2000). Emotional expressivity in men and women: Stereotypes and self-perceptions. Cognition & Emotion, 14, 5: 609. DOI: 10.1080/02699930050117648.

Hunt S.M., McKenna S.P., McEwen J., Backett E.M., Williams J., Papp E. (1980). A quantitative approach to perceived health status: a validation study. Journal of Epidemiology & Community Health, 34, 4: 281. DOI: 10.1136/jech.34.4.281.

ISS (2020). Epidemia Covid-19. Aggiornamento nazionale 30 marzo 2020 ore 16.00. Roma:

Istituto Superiore di Sanità. Retrieved from: https://www.epicentro.iss.it/coronavirus/bol lettino/Bollettino-sorveglianza-integrata-COVID-19_30-marzo-2020.pdf (18/04/2020).

Italian Ministry of Health (2020). Monitoraggio 19 aprile 2020. Rome: Ministry of Health.

Retrieved from: http://www.salute.gov.it/imgs/C_17_notizie_4530_0_file.pdf (19/04/

2020).

Janz N.K., Becker M.H. (1984). The health belief model: A decade later. Health Education Quarterly, 11, 1: 1. DOI: 10.1177/109019818401100101.

Ji W., Wang W., Zhao X., Zai J., Li X. (2020). Cross‐species transmission of the newly identified coronavirus 2019‐nCoV. Journal of medical virology, 92, 4: 433. DOI:

10.1002/jmv.25682.

Kasmaei P., Shokravi F.A., Hidarnia A., Hajizadeh E., Atrkar-Roushan Z., Shirazi K.K., Montazeri A. (2014). Brushing behavior among young adolescents: does perceived severity matter. BMC Public Health, 14, 1: 8. DOI: 10.1186/1471-2459-14-8.

(29)

Kitchenham B., Pfleeger S.L. (2002). Principles of Survey Research. Part 5: Populations and Samples. Sotware Engineering Notes, 25, 5: 17. DOI: 10.1145/571681.571686.

Leiner D.J. (2016). Our Research's Breadth Lives on Convenience Samples. A Case Study of the Online Respondent Pool “SoSci Panel”. Studies in Communication Media, 5, 4:

367. DOI: 10.5771/2192-4007-2016-4-36769-134.

Levinson D.J. (2011). The season of a woman’s life: a fascinating exploration of the events, thoughts, and life experiences that all women share. New York: Ballantine Books.

Lillie P.J., Samson A., Li A., Adams K., Capstick R., Barlow G.D., Easom N., Hamilton E., Moss P.J., Evans A., Ivan M. (2020). Novel coronavirus disease (Covid-19): the first two patients in the UK with person to person transmission. Journal of Infection, 80, 5: 578.

DOI: 10.1038/s41586-020-2012-7.

Lippmann W. (1922). Public Opinion. New York: Harcourt.

Lu H., Stratton C.W., Tang Y.W. (2020). Outbreak of Pneumonia of Unknown Etiology in Wuhan China: the Mystery and the Miracle. Journal of Medical Virology, 92, 4: 401.

DOI: 10.1002/jmv.25678.

Maiman L.A., Becker M.H. (1974). The health belief model: Origins and correlates in psychological theory. Health Education Monographs, 2, 4: 336. DOI:

10.177/109019817400200404.

McCombs M., Shaw D. (1972). The agenda-setting function of mass media. Public Opinion Quarterly, 36, 2: 176. DOI: 10.1086/267990.

Ministerul Sănătăţii (2020). Comunicate de presa. Decese 446 – 451. Bucharest: Ministerul Sănătăţii. Retrieved from: http://www.ms.ro/2020/04/19/decese-446-451/ (20/04/2020).

Muller M.P., Tomlinson G., Marrie T.J., Tang P., McGeer A., Low D.E., Detsky A.S., Gold W.L. (2005). Can routine laboratory tests discriminate between severe acute respiratory syndrome and other causes of community-acquired pneumonia? Clinical Infectious Disease, 40, 8: 1079. DOI: 10.1086/428577.

Musso F., Esposito G.F., Angioni M. (2019). Per un modello di valutazione della sostenibilità istituzionale delle politiche di sviluppo locale. Rivista Trimestrale di Scienza dell’Amministrazione, 4. DOI: 10.32049/RTSA.2019.4.01.

(30)

Nau D.P., Ellis J.J., Kline-Rogers E.M., Mallya U., Eagle K.A., Erickson S.R. (2005).

Gender and perceived severity of cardiac disease: evidence that women are “tougher”.

The American Journal of Medicine, 118, 11: 1256. DOI: 10.1016/j.amjmed.2005.08.006.

Ophir Y. (2019). The Effects of News Coverage of Epidemics on Public Support for and Compliance with the CDC-An Experimental Study. Journal of Health Communication, 24, 5: 547. DOI: 10.1080/10810730.2019.1632990.

Plant S. (1997). Zeros and ones: Digital women and the new technoculture. London:

Doubleday.

Rogers R. (1975). A Protection Motivation Theory of Fear Appeals and Attitude Change.

The Journal of Psychology, 91, 1: 93. DOI: 10.1080/00223980.1975.9915803.

Romanian Parliament (2020). Acte ale Parlamentului României. Bucharest: Romanian Parliament. Retrieved from: https://www.mai.gov.ro/wp-content/uploads/2020/03/Hot.- Parlamentului-RO-nr.-3-2020-%C3%AEncuviin%C8%9Bare-instituire-stare-de-urgen%

C8%9B%C4%83-1.pdf (20/04/2020).

Rosenstock I.M. (1974). Historical origins of the health belief model. Health Education Monographs, 2, 4: 328. DOI: 10.1177/109019817400200403.

Ross C.E., Bird C.E. (1994). Sex stratification and health lifestyle: consequences for men's and women's perceived health. Journal of Health and Social Behavior, 35, 2: 161. DOI:

10.2307/2137363.

Schmidt M.J., Hollensen S. (2006). Marketing research: An international approach.

Harlow, UK: Prentice Hall.

Voiculescu L. (2020). Ministrul Nelu Tătaru: Populația să aibă răbdare 7-10 zile. Suntem aproape de un vârf al pandemiei. Republica, April 19. Retrieved from:

https://republica.ro/ministrul-nelu-tataru-populatia-sa-aiba-rabdare-7-10-zile-suntem-apro ape-de-un-varf-al-pandemiei (20/04/2020).

Wang W., Tang J., Wei F. (2020). Updated understanding of the outbreak of 2019 novel coronavirus (2019‐nCoV) in Wuhan, China. Journal of Medical Virology, 92, 4: 441.

DOI: 10.1002/jmv.25689.

Weinstein N.D. (2000). Perceived probability, perceived severity, and health-protective

(31)

behavior. Health Psychology, 19, 1: 65. DOI: 10.1037/0278-6133.19.1.65.

Weinstein N.D., Klein W.M.P. (2015). Health Risk Appraisal and Optimistic Bias. In Weight JD., ed., International Encyclopedia of the Social & Behavioral Sciences, 698. DOI:

10.1016/B978-0-08-097086-8.25012-5.

WHO (2020d). Report of the WHO-China Joint Mission on Coronavirus Disease 2019 (COVID-19). Geneva: WHO. Retrieved from: https://www.who.int/docs/default- source/coronaviruse/who-china-joint-mission-on-covid-19-final-report.pdf (19/04/2020).

WHO (2020e). Coronavirus disease 2019 (COVID-19). Situation Report – 73. 2 April.

Retrieved from: https://www.who.int/docs/default-source/coronaviruse/situation-reports/

20200402-sitrep-73-covid-19.pdf?sfvrsn=5ae25bc7_2 (19/04/2020).

WHO (2020f). Novel coronavirus press conference at United Nations of Geneva. Geneva:

WHO. 29 January. Retrieved from: https://www.who.int/docs/default-source/corona viruse/transcripts/who-audio-script-ncov-rresser-unog-29jan2020.pdf (19/04/2020).

Witte K. (1992). Putting the fear back into fear appeals: The extended parallel process model. Communication Monographs, 59, 4: 329. DOI: 10.1080/03637759209376276.

World Health Organization - WHO (2020a). Pneumonia of unknown cause – China. 5 January. Retrieved from: https://www.who.int/csr/don/05-january-2020-pneumonia-of- unkown-cause-china/en (19/04/2020).

World Health Organization - WHO (2020b). WHO Director-General's opening remarks at the media briefing on COVID-19 - 11 March 2020.Retrieved from: https://www.who.int/

dg/speeches/detail/who-director-general-s-opening-remarks-at-the-media-briefing-on-cov id-19---11-march-2020 (19/04/2020).

World Health Organization - WHO (2020c). Naming the coronavirus disease (COVID-19) and the virus that causes it. Retrieved from: https://www.who.int/emergencies/

diseases/novel-coronavirus-2019/technical-guidance/naming-the-coronavirus-disease-(co vid-2019)-and-the-virus-that-causes-it (19/04/2020).

Wu F., Zhao S., Yu B., Chen Y.M., Wang W., Song Z.G., Hu Y., Tao Z.W., Tian J.H., Pei Y.Y., Yuan M.L., Zhang Y.L., Dai F.H., Liu Y., Wang Q.M., Zheng J.J., Xu L., Homes E.C., Zhang Y.Z. (2020). Author Correction: A new coronavirus associated with human

(32)

respiratory disease in China. Nature, 580: E7. DOI: 10.1038/s41586-020-2008-3.

Xu X., Chen P., Wang J., Feng J., Zhou H., Li X., Hao P. (2020). Evolution of the novel coronavirus from the ongoing Wuhan outbreak and modeling of its spike protein for risk of human transmission. Science China Life Sciences, 63, 3: 457. DOI: 10.1007/s11427- 020-1637-5.