Psychological reactions to COVID-19: Survey data assessing perceived susceptibility, distress, mindfulness, and preventive health behaviors

The COVID-19 pandemic created a complex psychological environment for persons in America. A total of 450 USA MTurk workers completed measures of: (a) basic demographic characteristics; (b) health risk factors for COVID-19; (c) perceived susceptibility variables related to COVID-19; (d) COVID-19 preventive health behaviors; and (e) distress, physical symptoms, and quality of life measures. The surveys were completed between April 9, 2020 and April 18, 2020. This recruitment period corresponded to the first 2–3 weeks of lockdown in most of the USA. Follow-up surveys were completed by 151 of the USA participants between June 19, 2020 and July 11, 2020 (approximately 2 months after the first measurement). These data permit evaluation of relationships among demographic variables, COVID-19 stress and coping, COVID-19 preventive health behavior, and the role of mindfulness as a possible moderator of distress as well as a predictor of preventive health behavior. The availability of follow-up data permit longitudinal analyses that provide a stronger basis for causal inference.


a b s t r a c t
The COVID-19 pandemic created a complex psychological environment for persons in America. A total of 450 USA MTurk workers completed measures of: (a) basic demographic characteristics; (b) health risk factors for COVID-19; (c) perceived susceptibility variables related to COVID-19; (d) COVID-19 preventive health behaviors; and (e) distress, physical symptoms, and quality of life measures. The surveys were completed between April 9, 2020 and April 18, 2020. This recruitment period corresponded to the first 2-3 weeks of lockdown in most of the USA. Follow-up surveys were completed by 151 of the USA participants between June 19, 2020 and July 11, 2020 (approximately 2 months after the first measurement). These data permit evaluation of relationships among demographic variables, COVID-19 stress and coping, COVID-19 preventive health behavior, and the role of mindfulness as a possible moderator of distress as well as a predictor of preventive health behavior. The availability of follow-up data permit longitudinal analyses that provide a stronger basis for causal inference.
© 2020 Published by Elsevier Inc. This is an open access article under the CC BY license ( http://creativecommons.org/licenses/by/4.0/ )

Value of the Data
• Data provide information on demographic characteristics, risk for COVID-19, perceived risk, distress, moderators of distress, and preventive health behaviors that allow for theoreticallybased predictions of COVID-19 relevant outcomes. • Data can be used for structural equation modeling and complex model testing of preventive health behaviors and psychological coping with the COVID-19 pandemic. • Follow-up data were collected two months after the first measurement in the USA which permits measurement of change across time.

Data Description
The data are available as SPSS.SAV files. Both sets of data contain the constructs and measures variables listed in Table 1 . The demographic characteristics of the two datasets are presented in Table 2 . The descriptive statistics of the survey measures the two datasets are presented in Table 3 .

Experimental Design, Materials and Methods
The survey was developed by an international team of researchers in the early stages of the COVID-19 pandemic. The project was approved by the Bowling Green State University Institutional Review Board (#1562479-4). Amazon Mechanical Turk Workers were enrolled through CloudResearch. A total of 635 highly rated participants initially responded to the survey from April 9, 2020 to April 18, 2020. Participant data were deleted from the data set if any of the following were detected: (a) less than 75% of items completed ( n = 131), duplicate IP address ( n = 26), (b) failing 2 of three attention check items ( n = 27), or (c) an unusually long time to complete the survey ( n = 1). This resulted in a final total sample of 450 participants for the Time 1 dataset.
A follow-up survey announcement and link were sent to all 635 participants who responded to the first survey announcement. The follow-up announcement and link were posted on MTurk on June 19, 2020 and were kept active for four weeks. A total of 178 MTurk workers completed the follow-up survey (39% response rate relative to the valid 450 responses to the first survey). Participant data were examined and retained using the same methods described above for Time 1 data. Time 1 and Time 2 responses were matched using unique ID codes that participants generated at time 1, IP addresses, and demographic characteristics. Out of the 178 completed surveys, we were able to unambiguously match 151. Thus, the final sample size for the Time 1 -Time 2 dataset set is 151.
The survey constructs and measure descriptions are provided in Table 1 . The demographic characteristics of participants in the Time 1 dataset and the Time1 -Time2 dataset are summarized in Table 2 . The descriptive statistics of measures in the two datasets are presented in Table 3 . The geographic distribution of participants in the two datasets are presented in Figs. 1 and 2 .
We comment here on the representativeness of MTurk samples. Mturk samples can have characteristics that differ from other samples used in psychological and social science research. MTurk samples tend to be more representative of the USA population relative to convenience samples, undergraduate samples, and samples collected in university communities but less representative than carefully collected national probability samples [1 , 2] . However, Chandler and Shapiro [2] noted that national probability samples can be biased because they rely on telephone methods that over sample older and more conservative participants.
There are strengths to MTurk samples that are important for this dataset. Most importantly, during the COVID-19 lockdown, online data gathering was the only practical way to gather complex survey data from diverse populations and differing geographic locations. Additionally, we adopted "best practices" by screening for response quality, using attention checks, and making sure that signaling cues were not in the surveys [3 , 4] .

Measures
Measure Translation. An international team of researchers from the USA, China, and Thailand developed the survey. Measures that were originally in Mandarin were translated to English. The first author and USA researchers then independently collected data using Qualtrics and MTurk.
Demographics. Participants completed items that provided information about basic demographic characteristics: age, sex, race, marital status, religion, educational attainment, if they     had children, living arrangements, and employment (e.g., job type, hours per week, changes in job since COVID-19). Participants also reported whether they were experiencing any illnesses or taking any medications. The demographic questionnaire used USA census items for sex, race, and religion which were then combined with items used in previous mindfulness and health investigations [5 , 6] . Self-isolation. Participants reported whether they were engaged in self-isolation or quarantine (yes/no). They also reported the number of times they left their residence in a typical day.
COVID-19 testing and COVID-19 status. At Time 2, participants reported whether they had a nasal and/or antibody test for COVID-19. They also reported whether or not they were infected with COVID-19 with 5 items: (a) currently not infected but not tested, (b) currently not infected based on negative test result, (c) currently infected but not tested, (d) currently infected based on positive test, (e) infected in the past based on antibody test.
General perceived vulnerability to disease (PVD). The PVD is a 15-item scale designed to measure general perceptions of risk for illness [7] . Items are rated on a 5-point Likert scale ranging from "strongly disagree" to "strongly agree." A total score was calculated with higher scores indicating higher levels of perceived vulnerability.
Perceived susceptibility to COVID-19. A single item was constructed to assess perceived vulnerability to COVID-19. Perceived susceptibility items like this have been extensively used in the preventive health behavior literature. The item was worded "How likely is it that you will contract COVID-19?" Response options ranged from "no chance" to "certain" using a 5-point scale.
The use of a single-item perceived susceptibility measure has been well supported in the preventive behavior literature [8] .
Mindfulness -Five Facet Mindfulness Questionnaire (FFMQ-24). The FFMQ-24 is a 24-item scale designed to measure mindfulness in daily life [9] . It has five subscales: awareness, observe, describe, nonjudgment, and nonreactivity. Items are rated on a 5-point scale ranging from "never or rarely true" to "very often or always true." A total score was calculated for the FFMQ. Mean scores were calculated for each subscale (mean scores were used because the number of items varies across subscales). Higher scores indicate higher levels of mindfulness [9] .
Intolerance of uncertainty. The 12-item version of the intolerance of uncertainty scale [10] was used to assess psychological distress associated with ambiguity and unpredictability (e.g., "unforeseen events upset me greatly"). Items were rated using a 5-point scale ranging from "not at all characteristic of me" to "entirely characteristic of me." A total score was calculated with higher scores indicating more intolerance of uncertainty.
Preventive actions taken scale (PATS). T he PATS was developed in late January 2020 based on recommendations generated by then-available COVID-19 research findings. The original 12-item measure assessed the extent to which participants engaged in behaviors to prevent COVID-19 infection. Two items were removed because of limited relevance to USA populations. Factor analyses using Time 1 data indicated that two additional items could be removed resulting in an 8item measure. Further, a varimax factor analysis indicated that the survey measured two distinct sets of COVID-19 preventive actions: "Personal protective equipment use" (3 items) and "avoiding travel and contact with other people (5 items)." Items were rated using a 5-point scale that ranged from "does not apply at all" to "applies very much or most of the time." Total scores for the personal protective equipment use and avoiding travel and contact with other people. Higher scores indicated more personal protective equipment use.
Post-traumatic stress symptoms. The Impact of Events Scale -Revised was used to measure post-traumatic stress symptoms [11] . The scale contains 22 items (e.g. "I thought about it when I didn't mean to") that were responded to on a 5-point scale that ranged from "not at all" to "extremely." A total score was calculated with higher scores indicating more post traumatic symptoms.
Physical symptoms. The Patient Health Questionnaire is a 15-item scale that [12] was used to measure physical symptoms associated with stress. The item measuring menstrual symptoms was removed because is only applied to males. Items are rated using a 3-point scale that ranges from "not bothered" to "bothered a lot." A total score was calculated with the remaining 14 items. Higher scores indicating more symptoms.
Psychological Symptoms of Distress. The six negatively worded items of the General Health Questionnaire were used to measure anxiety, depression, strain, and loss of confidence [13] . The items used a 5-point scale that ranged from "much less than usual" to "much more than usual." A total score was calculated for this measure with higher scores indicating more distress.
Quality of Life. T en items were taken from the 26 item World Health Organization Quality of life Brief scale [14] . Items were responded to using 5-point scales with anchors such as "never" to "very often" and "very dissatisfied" to "very satisfied." The items measured: life, health, sleep, work, self, concentrate, everyday life, depression, enjoy life, life meaning. A total score was calculated for this measure with higher scores indicating better quality of life.

Procedure
An announcement was placed on Amazon Mechanical Turk on April 9, 2020. The announcement read "The COVID-19 situation is creating worldwide challenges. University researchers hope to gain important useful information about how people are reacting to COVID-19 and coping with COVID-19. The survey is intended to be taken by individuals who are at least 18 years old who reside in the United States. The survey should take around 20 min to complete. Follow-up surveys will be sent to participants in the future to measure possible changes across time. You will receive $1.00 for completing each survey." Interested participants were then directed the survey informed consent form. If the participant provided consent, they were linked to the survey. The survey contained 3 attention check items and 3 captcha items. If a participant skipped an item, they were asked if they intended to skip the item(s) before being able to move on to the next page of the survey. If the participant responded "yes" that they intended to skip an item, they could move on to the next page of the survey. If they responded "no" they were returned to the skipped item.
The follow-up survey link was sent to all participants who completed the first survey. Both surveys took about 20 min to complete. The surveys were completed anonymously.

Ethics Statement
The project conformed to APA ethical guidelines and was approved by the Bowling Green State University Institutional Review Board. Participants were provided with informed consent. Survey items were designed to provide information on demographic characteristics, health history, risk for COVID-19, psychological distress, and preventive health behaviors. None of the survey items presented participants with stress-inducing stimuli. Therefore, there was no foreseeable risk for participation. Participants could end participation at any time without penalty. All collected data were not linked to participant identifying information or identifiable information. Therefore, responses were anonymous which was designed to provide participants with assurance of a higher degree of confidentiality.

Declaration of Competing Interest
All authors declare that they have no known competing financial interests or personal relationships which have, or could be perceived to have, influenced the work reported in this article.
acknowledge the helpful support of the Duke Kunshan University and Duke University institutional review board who provided guidance on policies and procedures for international data collection.

Supplementary Materials
Supplementary material associated with this article can be found in the online version at doi: 10.1016/j.dib.2020.10 6 687 .