Developing and validating the risk perceptions and behavioral responses questionnaire for COVID-19 (Risk PreceptCOVID-19): an application of the extended parallel process model

Background: In order to understand how people perceive, process, and respond to COVID-19 risk information, a valid and reliable instrument is needed. This study aimed to develop an instrument to measure the risk perceptions of people about COVID-19 based on the Extended Parallel Process Model (EPPM). Methods: This cross-sectional study was conducted in Tabriz, Iran. First, a comprehensive literature review was conducted to develop the initial draft of Risk-percept COVID-19. Then, in order to obtain a pre-nal version of the instrument, face and content validity were conducted. After that, construct validity (exploratory and conrmatory), internal consistency using test-retest reliability and Cronbach’s alpha of the instrument were measured. Results: Forty two item were initially generated. During the content validity, it was reduced to 29 items. The exploratory factor analysis (EFA) revealed four factors (perceived response ecacy, perceived threats (severity and susceptibility), perceived self-ecacy, and defensive responses (denial, avoidance and reactance) that accounted for 49.97 of the observed variance. The conrmatory factor analysis (CFA) designated a model with appropriate tness for the data. Cronbach’s alpha coecient for the dimensions ranged from .697 to .793, and the Intraclass Correlation Coecient (ICC) ranged from .71 to .80 within acceptable ranges. Conclusion: the current ndings show that the Risk-percept COVID-19 is a valid and reliable instrument to assess the cognitions and perceptions of adults.


Background
Corona virus disease (Covid- 19), a global pandemic, spread in many countries rapidly [1]. World health organization declared that the most important way to prevent and slaw dawn the transmission of COVID-19 is to reduce person-to-person transmission of COVID-19 among general population. Well informing and raising the knowledge of general population is an effective strategy to mitigate and control the current pandemic [2]. It is believed that awareness might persuade people to take appropriate actions toward control COVID-19 [3]. Indeed, in order to in uence people's behavior using persuasive messages applying fear appeals is recommended [4].
The extended parallel process model (EPPM) suggests that health risk messages produce two cognitive appraisals including an appraisal of the e cacy of the recommended response and an appraisal of the threat [5]. The model suggests that a fear appeal might work if certain conditions exist including response e cacy, self-e cacy, and threats (susceptibility and perceived severity). The model suggests two distinct processes in response to fear appeals; a danger control process and a fear control process [6].
In order to understand how people perceive, process, and respond to COVID-19 risk messages, a valid and reliable instrument is needed. Thus this study aimed to develop an instrument to measure the risk perceptions of people about COVID-19 based on the EPPM. We thought this would allow applying the questionnaire in both epidemiological and interventional studies and also could provide an opportunity for future research to take necessary actions toward COVID-19 control and prevention. The EPPM has many strong properties that make it useful in guiding many decisions of public communication campaigns.

Study design and data collection
This cross-sectional study was conducted in Tabriz, Iran. An online questionnaire was developed for general population. The Ethics Committee of Tabriz University of Medical Sciences approved the study (No: IR.TBZMED.REC.1398.1307). We obtained informed consent from all participants.

Item generation
The extended parallel process model was used for developing the risk perception and behavioral responses [5] to COVID-19. A comprehensive review of the literature was conducted to create an item pool [7][8][9][10][11][12][13][14][15][16][17][18][19][20][21]. Members of the research team (PR, SS, and ZS) reviewed the data and generated 150 items. After screening, in total 42 items were selected and divided into four dimensions including perceived response e cacy, perceived self-e cacy, perceived susceptibility and severity. The response e cacy dimension included seven items measuring beliefs about the effectiveness of the recommended response for averting COVID-19. Perceived self-e cacy included nine items t measuring beliefs about people's ability to perform the recommended response to avert COVID-19. Perceived susceptibility included eleven items measuring beliefs about people's risk of experiencing COVID-19. Perceived severity included ve items measuring beliefs about the magnitude of COVID-19. Each item is rated on a 5-point Likert-scale (1 = strongly disagree, 2 = disagree, 3 = neutral, 4 = agree, and 5 = strongly agree). Content validity Qualitative and quantitative content validity were applied for Risk-Percept QVID-19. A panel of six experts including three health education specialists, a specialist physician in infectious diseases, an epidemiologist and a psychologist evaluated the content validity of the Risk-Percept COVID-19. They assessed the questionnaire in the terms of wording, item allocation, grammar, and scaling. In the quantitative phase, the content validity index (CVI) and content validity ratio (CVR) were tested. Clarity, simplicity, and relevance of the items were measured in CVI assessment [22].

Face validity
Both qualitative and quantitative methods were used to measure the face validity of the Risk-Percept COVID-19. In the qualitative process, 8 young adults were asked to assess each item of the Risk-Percept COVID-19 to indicate whether they felt it was di cult or ambiguous to answer the items. According to the participants' viewpoint, the di cult and vague items were revised. In the quantitative stage, the impact score (frequency × importance) was calculated to determine the percentage of participants who recognized items as important on a 5-point Likert scale. Items were found suitable whether they had an impact score equal to or greater than1.5 [23]. Overall three items had an impact score equal or lower than 1.5, and 42 items had an impact ranging from 1.8 to 5. Therefore, the rst form of instrument contained 42 items.

Psychometric properties evaluation of the Risk-Percept COVID-19
In order to test the psychometric properties of the Risk-Percept COVID-19, a cross-sectional study was carried out in Tabriz, Iran. Participants were recruited from online applications and announcements (Telegram, WhatsApp, and Instagram) that ask people to take part in the study. The announcements asked people to take part in the study for control and prevention of COVID-19 and full-ll the study questionnaire. Those interested in participating in the study complete the online questionnaire. In order to estimate an optimal sample size, 5-10 respondents in to one variable ratio was used [24]. The participants had to be 15 years of age or older and having consent to participate in the study.
Data collection was conducted in March 21 to 28, 2020. The high incidence of COVID-19 in Tabriz and the crucial need for a questionnaire to assess educational and interventional programs led to it's selection as study design.

Statistical analysis
The characteristics of the participants were summarized as numbers, percentages, or means with standard deviations, where appropriate. Several statistical methods were performed to evaluate the psychometric properties of the Risk-Percept COVID-19 as follows.

Exploratory factor analysis
Exploratory factor analysis (EFA) was applied to identify the main factors of the Risk-Percept COVID-19. According to the number of items in the Risk-Percept COVID-19. A principal component analysis (PCA) with varimax rotation was used to extract the main factors. The Kaiser-Meyer -Olkin (KMO) test and Bartlett's test of sphericity were applied to evaluate the adequacy of the sample for the factor analysis [25]. The factors with an eigenvalue of more than 1 was considered acceptable for factor extraction, and a scree plot was performed to specify the number of factors. Factor loadings equal to or greater than .40 were considered acceptable [26].

Reliability
Internal consistency Cronbach's alpha coe cient was used to measure the internal consistency of each item, the whole scale and each subscale of the Risk-Percept COVID-19. Alpha values of .70 and above were considered acceptable [27].

Test-retest reliability
Test-retest reliability was performed to assess Risk-Percept COVID-19's stability by estimating the intraclass correlation coe cient (ICC). Two weeks after the rst online compellation of the questionnaire, twenty people were asked to full out and either to send them back. ICC values of .40 and above are considered acceptable [28].

Overall risk perception: linear transition scoring
The sum of responses across the 29 items was calculated (responses are scored; strongly disagree = 1; disagree = 2; neutral 3; agree = 4; strongly agree = 5). Total scores ranged from 29 to 145. The sum of scores for 29 answered items were calculated. We used linear transformation to convert to 0-100 score by subtracting the lowest raw score from the possible raw sum score divided by 116 (the highest score minus the lowest score (145 − 29)) and multiplied by 100. Lower scores indicating low risk perception, and higher scores indicating a better risk perception.

Reliability
In order to measure reliability, Cronbakh's alpha was calculated separately for the Risk-percept COVID-19 as a whole and for each dimension of the Risk-percept COVID-19. Cronbakh's alpha coe cient for the Risk-percept COVID-19 was .741 and ranged from .697 to .793 for the subscale. No items of the questionnaire were deleted in this phases. Furthermore, test-retest analysis was conducted to test the stability of the questionnaire. The results are shown in Table 3. Self-e cacy 7 items (8,9,11,12,13,14) .760 .72 4 Threats 7 items (16,17,28,27,26,10,29) .697 .71

Discussion
In this study, we developed a risk perception scale to measure cognitive appraisal of adults for COVID-19 threat. This study well evaluated the psychometric properties, validity, and reliability of the Rick-percept COVID-19 scale. This EPPM based study provide a measure for evaluating the items with four elements among Iranian adults. Most importantly, we wanted to make a self-report instrument of risk perception available for Iranian research, allowing for future interventional and cross-cultural assessments. A fourdomain was emerged through EFA. Also, a CFA showed acceptable the t of data. As such, the nal Rickpercept COVID-19 contained 29 items, with seven items representing response e cacy, eight items for susceptibility, seven items for self-e cacy, and eight items for severity). Our results indicate that the Rickpercept COVID-19 has sound psychometric properties, good structural validity with acceptable internal consistency across the scale and subscale of the instrument.
During the outbreak of COVID-19, it is important that how people perceive the risk of COVID-19 threat and its effects on their protective behavior. According to EPPM, when people receive a message they initiate two cognitive appraisals; an appraisal of the threat (severity and susceptibility) and an appraisal of the response e cacy. Cognitive appraisal of the threat in our scale was measured with the items such as "Corona is a threat, I think about it" or "I am at risk for getting Corona" (perceived susceptibility) or "Corona is a lethal threat." (perceived severity). Thus, if people have low perceived susceptibility or/and low perceived severity, they won't respond to the COVID-19 threat. In contrast, if people think about the COVID-19 threat and percept that they are vulnerable to the threat, it might lead to serious harm, thus, they scared and motivated to healthy practice against the threat. More the stronger perceived the threat, the greater the fear and the persuasive to have protective behavior [29,30].
When people believe that they are able to perform the recommended response, for example, "It is easy for me to use disinfectants" or "It is easy for me to access masks and disinfectants", they show high perceived self-e cacy, and they believe that doing health recommendations is effective in preventing Coronavirus" or "staying at home" (recommended response) works in averting the coronavirus threat, they have high perceived response e cacy. It's seems that intensive people's perceptions of Coronavirus threat and e cacy motivate them to control the danger [31]. In contrast, if people doubt their ability to perform the recommended response for averting Coronavirus threat (low perceived self-e cacy), they doubt whether the recommended response effectively prevent the threat. Therefore, if they have low perceived response e cacy, they believe that there's no use in controlling the danger, thus, rather than controlling the danger, they turn around defensive strategies such as "I believe that health staff have highly exaggerated this disease." When people block further thoughts about the Coronavirus threat (avoidance response: "when Television or Radio talks about Corona, I ip the channel", or if they refuse that it is possible they can experience the Coronavirus threat (denial response: "God protects me, even if I don't follow the health principles of Corona.", or if they try to pretense that the source of the message deceit people (reactance responses: "I believe that this is a government plan, they are trying to amuse people"), defensive responses occurs [32].

Strengths
According to EPPM, the key concept of this instrument is to determine people's perception of threat and e cacy toward the Coronavirus pandemic threat and recommended response. Then, the Risk-percept COVID-19 helps researcher to explore the cognitive and emotional mechanisms underlying acceptance or rejection of COVID-19 prevention campaign messages. Interpretation of the instrument's score and the association of it's dimensions indicate that emotion fear is associated with fear control responses (avoidance, denial and reactance) and perceptions or cognitions about recommended response are associated with danger control responses [33]. More speci cally, this EPPM based instrument speci es two cognitive process in response to threating COVID-19 information; danger control or fear control responses [34].

Conclusion
In general, the current ndings show that the Risk-percept COVID-19 is a valid and reliable instrument to assess the cognitions and perceptions of adults. Further studies from different cultural and economic backgrounds are suggested to nd stronger psychometrics properties.