COVID-19 Beliefs, Behaviors & Norms Survey

About

Here you can access aggregate data from the COVID-19 Beliefs, Behaviors & Norms Survey. There are two ways to access the data. First, downloading the data via simple online form, Second, there is a REST API for programmatically accessing the data. The data provided through these two endpoints contains of summary statistics of the dataset. For the full dataset containing individual level responses, please request the data through this link.

Please sign up for the mailing list for aggregate data users.


Online download

The first way to download aggregated data is through the form below. This returns a CSV file aggregated by gender, age group and country.








Through the API

The second way to get data is through the API. The API allows a better way of querying and slicing the data. There are five parameters used for query,age, gender, country, signal, and output. country is a required parameter.

ParametersDescriptionFormat
ageage range for the datacould be one of all,20-30,31-40,41-50,51-60,61-70, or 71-80, default: all
gendergender of the usersmale,female,other,or unknown, default: all
countryuse to query data for a country (two letter ISO code) Should be one of , all,AF,AO,AR,AU,AZ,BD,BO,BR,CA,CI,CL,CM,CO,DE,DZ,EC,EE,EG,ES,FR,GB,GE,GH,GT,HN,ID,IN,IQ,IT,JM,JP,KE,KH,KR,KZ,LK,MA,MM,MN,MX,MY,MZ,NG,NL,NP,PE,PH,PK,PL,PT,RO,SD,SG,SN,TH,TR,TT,TW,TZ,UA,UG,AE,US,UY,VE,VN,ZA. A two letter ISO code, see this link for reference. Note: When country is set to all, the default for wave is all.
signalSpecific signal you want from the data (defined below)Could be one of all,vaccine_accept,mismatch_index,community_risk_index,measures_taken,locations_would_attend,distancing_familiarity,effect_mask,news_sources,news_mediums,health,retail,restaurants,or worship. Multiple signals can be requested at once using e.g. {vaccine_accept,mismatch_index,restaurants}, default: all.
waveThe survey was conducted in waves (two week periods). Weights are computed for users in a single wave once all the data is collected. wave1 is July 6 - July 20. wave2 is July 21 - August 3, wave3 is Aug 3 - Aug 16, 2020, wave4 is Aug 17 - Aug 30, 2020, wave5 is Aug 31 - Sep 13, 2020, wave6 is Sep 14 - Sep 27, 2020, wave7 is Sep 28 - Oct 11, 2020, wave8 is Oct 12 - Oct 25, 2020 wave9 is Oct 26 - Nov 8, 2020 and so on. The survey is expected to continue until the end of 2020. Note: this parameter only applies for a subset of countries (US,BR,MX,IN,GB,JP,DE,IT,AR,ID,CO,TR,FR,EG,VN,PH,PK,BD,RO,PL,TH,MY,NG)wavei where i is 1,2,3,..., orall, default: all
us_stateUS state (2 letter code. Use this link for the mapping. Should be one of CA,TX,FL,NY,PA,OH,IL,WA,GA,NC,VA,MI,MA,NJ,IN,MO,AZ,TN,CO,MD,MN,WI,OR,KY,OK,AL,SC,NV,UT,LA,KS,AR,IA,CT,NM,ID,MS,WV,ME,NE,MT,NH,RI,DE,HI,SD,AK,ND,VT,WY, default: None
india_stateIndia state (2 letter code. Use this link for the mapping.) Should be one of MH,UP,WB,KA,DL,TN,GJ,BR,RJ,TG,KL,AP,MP,HR,OR,JH,PB,AS,UT,JK,CT,HP,CH,GA,MN,TR,ML,AR,NL,SK,MZ,PY,DD,AN,LD, default: None
timeseriesboolean parameter which iterates over all waves and returns data for each wave.not set by default, works if set to true
outputformat of the output json, default: json

 

Signal Definitions

This section provides the details on how the signals are obtained and what they mean.

Signal nameDescriptionExample output value
vaccine_acceptResponses to the question If a vaccine for COVID-19 becomes available, would you choose to get vaccinated?. (-1 in the response means the user saw the question but did not answer.)"vaccine_accept": { "weighted": { "Yes": 0.676, "-1": 0.029, "Don't know": 0.184, "No": 0.111 } }
mismatch_indexAverage of the disagreement in personal beliefs of a user and their beliefs of others. This is based on reponses to the questions: Q_individual How important is it for you to take actions to prevent the spread of COVID-19 in your community? (possible answers: Extremely important, Very important, Moderately important, Slightly important, Not important at all) and Q_societal How important do other people in your community think it is to take actions to prevent the spread of COVID-19?. We first discretize the results (20 = Not important at all and 100 = Extremely important) and take the difference between a users responses Q_individual - Q_societal. mismatch_index is the mean of the distribution of the selected population. Interpretation: A value close to zero indicates agreement in beliefs in the society, a higher value indicates disbelief. "mismatch_index": {"weighted": 19.031}
community_risk_indexBased on response to the question: How dangerous do you think the COVID-19 risk is to your community? with possible answers: {Not at all dangerous, Slightly dangerous, Moderately dangerous, Very dangerous, Extremely dangerous}. The measure returns the fraction of users who respond to a particular option. "community_risk_index": { "weighted": { "Very dangerous": 0.366, "Moderately dangerous": 0.242, "Slightly dangerous": 0.139, "Not at all dangerous": 0.036, "-1": 0.004, "Extremely dangerous": 0.213 }}
measures_takenBased on response to the question: What measures have you taken to prevent infection from COVID-19 in the past week? (multiple choices can be selected). The measure returns the fraction of users who answered Yes/No for each choice. "measures_taken": { "avoid_touching_face": { "weighted": { "Yes": 0.963, "No": 0.037 } }, "getting_the_flu_vaccine": { "weighted": { "Yes": 0.839, "No": 0.161 } }, "meter_distance": { "weighted": { "Yes": 0.658, "No": 0.342 } }, ... }
locations_would_attendBased on response to the question: Which of the following businesses, locations, or events would you visit or attend in the coming two weeks if they were operating at full capacity?. The measure returns the fraction of users who answered Yes/No for each choice. "locations_would_attend": { "health_care_facilities": { "weighted": { "Yes": 0.277, "No": 0.723 } }, "parks_and_beaches": { "weighted": { "Yes": 0.254, "No": 0.746 } }, "places_of_employment": { "weighted": { "Yes": 0.141, "No": 0.859 } }, ... }
distancing_familiarityBased on response to the question: Which of the following best describes your familiarity with the term 'physical distancing' during the COVID-19 pandemic? with possible answers: {I have not heard of it, I have heard of it but do not know what it means, I have heard of it and have some idea of what it means, I have heard of it and know what it means}. The measure returns the fraction of users who respond to a particular option. "distancing_familiarity": { "weighted": { "I have heard of it and have some idea of what it means": 0.125, "I have heard of it and know what it means": 0.788, "I have heard of it but do not know what it means": 0.045, "I have not heard of it": 0.042 } }
effect_maskBased on response to the question: How effective is wearing a face mask for preventing the spread of COVID-19? with possible answers: {Extremely effective, Very effective, Moderately effective, Slightly effective, Not effective at all}. The measure returns the fraction of users who respond to a particular option. "effect_mask": { "weighted": { "Extremely effective": 0.485, "Moderately effective": 0.125, "Not effective at all": 0.011, "Slightly effective": 0.037, "Very effective": 0.342 } }
news_sourcesBased on response to the question: In the past week, from which of the following, if any, have you received news and information about COVID-19? Select all that apply. with possible answers: {government_health_authorities:Government health authorities or other officials, journalists:Journalists, local_health_workers:Local health workers, clinics, and community organizations, ordinary_people_i_dont_know_personally:Ordinary people I don't know personally, ordinary_people_i_know_personally:Ordinary people I know personally, politicians:Politicians, scientists:Scientists, doctors, and health experts, world_health_organization:World Health Organization (WHO)}. The measure returns the fraction of users who selected each of these choices. "news_sources": { "government_health_authorities": { "weighted": { "No": 0.545, "Yes": 0.455 } }, "journalists": { "weighted": { "No": 0.417, "Yes": 0.583 } }, "local_health_workers": { "weighted": { "No": 0.726, "Yes": 0.274 } },... }
news_mediumsBased on response to the question: In the past week, from which of the following, if any, have you received news and information about COVID-19? Select all that apply. with possible answers: {online_sources: Online sources (websites, apps, social media), messaging_apps:Messaging apps / SMS/ text messaging, newspapers:Newspapers, television: Television, radio: Radio}. The measure returns the fraction of users who selected each of these choices. "news_mediums": { "messaging_apps": { "weighted": { "No": 0.804, "Yes": 0.196 } }, "newspapers": { "weighted": { "No": 0.461, "Yes": 0.539 } }, "online_sources": { "weighted": { "No": 0.164, "Yes": 0.836 } },... }
health,retail,restaurants,worshipBased on response to the question(s): Compared to a [location] with no precautions at all, are you more or less likely to visit a restaurant with the following precautions?, where location could be health,retail,restaurants,worship. Possible answers of this question are: {More likely to visit, Doesn't affect my actions, Less likely to visit}. The measure returns the fraction of users who selected each of these choices. "restaurants": { "everyone_has_to_pass_a_temperature_check": { "weighted": { "Doesn't affect my actions": 0.175, "More likely to visit": 0.506, "-1": 0.21, "Less likely to visit": 0.109 } }, "operating_at_limited_capacity": { "weighted": { "Doesn't affect my actions": 0.151, "More likely to visit": 0.361, "-1": 0.106, "Less likely to visit": 0.383 } }, ... }

 

API Request examples

Method: GET

Example 1. Retrieve the data for a specific signal (say, locations people would attend locations_would_attend) in a specific country, age group for all genders using the http://covidsurvey.mit.edu:5000/query end point

This could be done using the link http://covidsurvey.mit.edu:5000/query?age=20-30&gender=all&country=US&signal=locations_would_attend would return the response:




Using the link http://covidsurvey.mit.edu:5000/query?age=20-30&gender=all&country=US&output=csv&signal=locations_would_attend would return the same response as a csv file.

Example 2. Retrieve the data for multiple signals (say, vaccine_accept, mismatch_index, and community_risk_index) in all countries, for a specific age group and gender

This could be done using the link http://covidsurvey.mit.edu:5000/query?age=31-40&gender=female&country=all&signal={vaccine_accept,mismatch_index,community_risk_index} would return the response:





Method: Python

Method: R

Citing

We hope this API is useful to different research groups. If you use the API and would like to cite it, please use:

Global survey on coronavirus beliefs, behaviors, and norms: Technical Report

Credit: Website design taken from the UMD COVID-19 survey API.