Because landline telephone only covered 182 million households in China [12], there are a large amount of population which cannot be accessed using traditional landline-lined based telephone surveys, while there are 854 million Chinese people get access to Internet in 2019 [13], we chose to use an online sample in the present study. Upon IRB approval, 4991 respondents were randomly recruited from a national online panel that consists of 1,075,809 Chinese adults from February 12, 2020 to February 14, 2020, during which period of time the number of the new coronavirus cases surpassed 10,000 in a single day and the total cases in China reached to 63940.
All the respondents completed informed consent form before filling out the survey. After data cleaning procedure, a total of 4991 respondents were included in the final sample. Table 1 shows the demographic characteristics and anxiety level of the respondents. Of the respondents, 18% had high school education or lower, 31.9% had associate degree, 47% had college degree, 5% had graduate degree. 49.6% were male and 50.4% were female. A majority of the respondents have a monthly income below 5,000 rmb.
Table 1
Respondents’ characteristics, perceived knowledge, social and geographical proximity, anxiety level (N = 4991)
Demographics | N | % |
Gender | | |
Female | 2514 | 50.4% |
Age | | |
18–30 years | 3203 | 64.2% |
31–40 years | 1246 | 25.0% |
41–50 years | 399 | 8.0% |
51–60 years | 126 | 2.5% |
≧ 61 years | 17 | 0.3% |
Education | | |
Primary or secondary school | 800 | 16.0% |
Associate degree | 1585 | 31.8% |
College | 2356 | 47.2% |
Graduate degree | 250 | 5.0% |
Monthly Income | | |
≦ 1000 rmb | 929 | 18.6% |
1001–2000 rmb | 361 | 7.2% |
2001–5000 rmb | 1707 | 34.2% |
5001–8000 rmb | 1498 | 30.0% |
8001–15000 rmb | 411 | 8.2% |
≧ 15001 rmb | 85 | 1.7% |
Anxiety Level | | |
Normal (< 50) Mild to moderate anxiety levels (50–59) Moderate to severe levels (60 to 69) | 3963 699 256 | 79.4% 14% 5.1% |
Severe anxiety levels (> 70) | 73 | 1.5% |
Social Proximity | | |
Yes | 293 | 5.9% |
Geographical Proximity | | |
Yes | 1608 | 32.2% |
Perceived Knowledge of COVID-19 | | |
Do not know COVID-19 at all Know a little bit Know some Know fair amount Know a great deal | 22 209 2193 2035 532 | 0.4% 4.2% 43.9% 40.8% 10.7% |
Risk perception. Respondents were asked to assess their perceived risk of “being affected by the new coronavirus” and “being affected by the new coronavirus compared to others,” which was measured on 5-point Likert-type scale ranging from 1 = very unlikely to 5 = very likely.
Anxiety level was measured by Chinese version of Zung Self-Rating Anxiety Scale (SAS-20)[14] that consisted 20 questions. Respondents were asked to estimate their anxiety level within a period of one or two weeks prior to taking the test. Sample items include: I feel afraid for no reason at all; I can feel my heart beating fast; I have nightmares. Responses was rated on a four-point Likert-type scale anchored at 1 (never or very rare), 2 (sometimes), 3 (often), and 4 (very often or always). The total SAS score may vary from 20 (no anxiety at all) to 80 (severe anxiety). Zung recommended converting the raw score to Index Scores (ranged between 25 and 100) by multiplying the raw score by 1.25. The cut-off points were used: an index score of 50 and above. The internal reliability (Cronbach alpha) is .84 (M = 42.05, SD = 10.07).
Social proximity was measured with the dichotomous question “Is there someone you know who contracted the new coronavirus,” with (0 = No; 1 = Yes). Geographical proximity was measured with direct dichotomous (yes/no) question “Are there reported cases of infection in your neighborhood or in your town?”
Media exposure. In order to assess the amount of media exposure to different COVID-19-related content, we synthesized 10 categories of COVID-19 information that have been frequently appeared in mass media and social media. Respondents were asked to indicate how much attentions they paid to the each of following information in media, including “how to prevent the new coronavirus,” “the number of cases of infection,” “news coverages of patients,” “ news coverage of doctors and nurses,” “news coverages of government officials,” “news coverages of scientists,” “donation information,” “life of ordinary people during the COVID-19 outbreak,” “information about returning to work/school,” and “analysis of the pandemic” (1 = Do not pay any attention at all; 5 = Pay a lot of attention).
Dada Analyses
The data was analyzed using SPSS version 25. Pearson correlation was conducted to examine the relationship between predictors and outcome. The level of significance was set at p < 0.05. Hierarchical ordinary least squares (OLS) regression analysis was used to assess the associations between anxiety, media exposure, social proximity, geographical proximity, and risk perceptions, adjusting for age, sex, education, and income.