Methodology: How the GAIA Study Was Conducted
Ipsos, on behalf of The Virsa Foundation conducted a comprehensive survey between September 29 – November 29, 2021 to understand the relationship between Americans' dietary habits and how they fared during the COVID-19 pandemic in 2021.
Study Design & Sample
For this survey, a sample of 7,816 adults aged 18+ from the continental U.S., Alaska, and Hawaii, was interviewed online in English using the Ipsos KnowledgePanel©. An additional sample of 6,810 adults aged 18+ from the continental U.S., Alaska, and Hawaii, was interviewed online in English using survey links distributed by select partners to their networks, which was conducted between October 15—November 28, 2021. In total, 14,626 adults were surveyed.
Sampling and what these figures describe
The GAIA Study drew on two samples. The first was a probability-based sample of 7,816 adults from the Ipsos KnowledgePanel®, which is designed to represent the US adult population and carries a margin of sampling error of ±1.3 percentage points at the 95% confidence level, taking into account a design effect of 1.35.
The second was a convenience sample of 6,810 adults recruited through open survey links distributed by more than 50 partner organisations to their own networks, largely through the social media accounts and mailing lists of plant-based advocates. A convenience sample has no known selection probabilities, so it carries a credibility interval of ±3.6 percentage points rather than a margin of sampling error.
In total, 14,626 adults were surveyed. Combined figures describe the 14,626 people who took part in this study. They are not projectable to the US adult population, and no margin of sampling error applies to them.
This matters most for figures about plant-based eating. Plant-based participants were recruited almost entirely through partner and advocacy networks. Among participants who had COVID-19, the probability-based panel contained too few plant-based respondents to support comparison, so all diet-group comparisons in this study describe the partner sample.
Each respondent entering the survey was screened for having had, or suspected to have had, COVID-19 between January 2020 and October 2021. Regardless of COVID-19 response, all respondents provided demographic data, diet type followed, food consumed regularly, and a checklist of medical conditions.
Target Audience & Data Collection
The survey focused on individuals exposed to COVID-19, including both confirmed cases and those who suspected infection.
Data was collected on multiple key factors, including:
Demographics (age, gender, income, education, race, region).
Dietary habits (Standard American Diet, Whole Food Plant-Based, Vegan, etc.).
COVID-19 experiences (severity of symptoms, medical interventions, mental health impact).
Overall health and well-being, including stress levels, resilience, and burnout.
2,375 U.S. adults had or suspected they had COVID-19.
Sampling Error & Weighting
The GAIA Study drew on two samples. The first was a probability-based sample of 7,816 adults from the Ipsos KnowledgePanel®, which is designed to represent the US adult population and carries a margin of sampling error of ±1.3 percentage points at the 95% confidence level, taking into account a design effect of 1.35.
The second was a convenience sample of 6,810 adults recruited through open survey links distributed by more than 50 partner organisations to their own networks, largely through the social media accounts and mailing lists of plant-based advocates. A convenience sample has no known selection probabilities, so it carries a credibility interval of ±3.6 percentage points rather than a margin of sampling error.
Consumer Segmentation
To better understand Americans' relationships with food, health, and stress, Ipsos categorized respondents into five distinct consumer segments: Health-Conscious, At a Crossroad, Resilient, Unhealthy with Comorbidity, Healthy & Affluent.
These segments provide key insights into how different groups experience and approach diet, health, and wellness.
Definitions
How diet groups were determined. Participants described their own diet, and were separately asked how often they ate each of 23 food types. The food frequency questions were built into the questionnaire by design, agreed with Ipsos, so that self-described diet labels could be checked against reported eating behaviour. The two measures agree closely: participants describing a whole-food plant-based or vegan pattern reported never eating meat, dairy, poultry, eggs, fish or shellfish, while those describing a Standard American Diet reported eating each of them between two or three times a week and almost daily. Both measures are self-reported.
The Standard American Diet (SAD) In this study, "Standard American Diet (SAD)" refers to participants who described their diet as "I don't follow a specialized diet." These self-descriptions were checked against the study's food frequency questions: these participants reported eating meat, dairy, poultry and eggs between two or three times a week and almost daily.
A Whole Food Plant-Based (WFPB) approach is defined as someone who does not eat any food derived from animals and avoids packaged or processed food. This exclusions include dairy, eggs, and honey.
Veganism, on the other hand, is defined as someone who does not consume any food or non-food product derived from animals but may consume processed foods. More generally, for some, being vegan can also be an ethical or philosophical identity, while a WFPB approach can be a lifestyle or approach to nutrition.
Vegetarian: A vegetarian diet is a plant-based diet that may include dairy products and/or eggs, but excludes meat
Pescatarian: A dietary pattern that includes fish and seafood, but excludes meat and poultry.
Flexitarian: A dietary pattern that is primarily vegetarian but occasionally includes meat, poultry, or fish.
Learn More About the GAIA Study Findings
Explore how different dietary patterns affect physical and mental health outcomes during global health crises.