{"id":301999,"date":"2026-05-07T10:00:20","date_gmt":"2026-05-07T14:00:25","guid":{"rendered":"https:\/\/alpha.pewresearch.org\/pewresearch-org\/?p=301999"},"modified":"2026-05-07T12:26:12","modified_gmt":"2026-05-07T16:26:12","slug":"wellness-influencers-methodology","status":"publish","type":"post","link":"https:\/\/alpha.pewresearch.org\/pewresearch-org\/data-labs\/2026\/05\/07\/wellness-influencers-methodology\/","title":{"rendered":"Methodology"},"content":{"rendered":"\n

This report aims to gain an understanding of social media influencers who focus on health and wellness topics and the experience of U.S. adults who get such information from these influencers. It uses two different methodologies, including an analysis of health and wellness influencers who regularly post on Instagram, TikTok or YouTube, as well as a nationally representative survey of U.S. adults conducted through Pew Research Center\u2019s American Trends Panel (ATP).<\/p>\n\n\n\n

Survey<\/h3>\n\n\n\n

Survey data in this report comes from two waves of the American Trends Panel (ATP), Pew Research Center\u2019s nationally representative panel of randomly selected U.S. adults. It includes data from ATP Wave 173, conducted June 9-15, 2025, and ATP Wave 182, conducted Oct. 20-26, 2025.<\/p>\n\n\n\n

Overview<\/h4>\n\n\n\n

Wave 173 of the American Trends Panel (ATP) was conducted June 9-15, 2025. A total of 5,023 panelists responded out of 5,737 who were sampled, for a survey-level response rate of 88%.<\/p>\n\n\n\n

The cumulative response rate accounting for nonresponse to the recruitment surveys and attrition is 3%. The break-off rate among panelists who logged on to the survey and completed at least one item is 1%. The margin of sampling error for the full sample of 5,023 respondents is plus or minus 1.6 percentage points.<\/p>\n\n\n\n

The survey includes an oversample<\/a> of non-Hispanic Asian adults in order to provide more precise estimates of the opinions and experiences of this smaller demographic subgroup. Oversampled groups are weighted back to reflect their correct proportions in the population.<\/p>\n\n\n\n

SSRS conducted the survey for Pew Research Center via online (n=4,842) and live telephone (n=181) interviewing. Interviews were conducted in both English and Spanish.<\/p>\n\n\n\n

To learn more about the ATP, read \u201cAbout the American Trends Panel<\/a>.\u201d<\/p>\n\n\n\n

Panel recruitment<\/h4>\n\n\n\n

Since 2018, the ATP has used address-based sampling (ABS) for recruitment. A study cover letter and a pre-incentive are mailed to a stratified, random sample of households selected from the U.S. Postal Service\u2019s Computerized Delivery Sequence File. This Postal Service file has been estimated to cover 90% to 98% of the population.[1. AAPOR Task Force on Address-based Sampling. 2016. \u201cAAPOR Report: Address-based Sampling<\/a>.\u201d] Within each sampled household, the adult with the next birthday is selected to participate. Other details of the ABS recruitment protocol have changed over time but are available upon request.[2. Email pewsurveys@pewresearch.org<\/a>.] Prior to 2018, the ATP was recruited using landline and cellphone random-digit-dial surveys administered in English and Spanish.<\/p>\n\n\n\n

A national sample of U.S. adults has been recruited to the ATP approximately once per year since 2014. In some years, the recruitment has included additional efforts (known as an \u201coversample\u201d) to improve the accuracy of data for underrepresented groups. For example, Hispanic adults, Black adults and Asian adults were oversampled in 2019, 2022 and 2023, respectively.<\/p>\n\n\n\n

Sample design<\/h4>\n\n\n\n

The overall target population for this survey was noninstitutionalized persons ages 18 and older living in the United States. It featured a stratified random sample from the ATP in which non-Hispanic Asian adults were selected with certainty. The remaining panelists were sampled at rates designed to ensure that the share of respondents in each stratum is proportional to its share of the U.S. adult population to the greatest extent possible. Respondent weights are adjusted to account for differential probabilities of selection as described in the Weighting section below.<\/p>\n\n\n\n

Questionnaire development and testing<\/h4>\n\n\n\n

The questionnaire was developed by Pew Research Center in consultation with SSRS. The web program used for online respondents was rigorously tested on both PC and mobile devices by the SSRS project team and Pew Research Center researchers. The SSRS project team also populated test data that was analyzed in SPSS to ensure the logic and randomizations were working as intended before launching the survey.<\/p>\n\n\n\n

Incentives<\/h4>\n\n\n\n

All respondents were offered a post-paid incentive for their participation. Respondents could choose to receive the post-paid incentive in the form of a check or gift code to Amazon.com, Target.com or Walmart.com. Incentive amounts ranged from $5 to $20 depending on whether the respondent belongs to a part of the population that is harder or easier to reach. Differential incentive amounts were designed to increase panel survey participation among groups that traditionally have low survey response propensities.<\/p>\n\n\n\n

Data collection protocol<\/h4>\n\n\n\n

The data collection field period for this survey was June 9 to 15, 2025. Surveys were conducted via self-administered web survey or by live telephone interviewing.\u00a0<\/p>\n\n\n\n

For panelists who take surveys online:<\/strong>[3. 3. The ATP does not use routers or chains in any part of its online data collection protocol, nor are they used to direct respondents to additional surveys.] Postcard notifications were mailed to a subset on June 9.[4. Postcard notifications for web panelists are sent to 1) panelists who were recruited within the last two years and 2) panelists recruited prior to the last two years who opt to continue receiving postcard notifications.] Survey invitations were sent out in two separate launches: soft launch and full launch. Sixty panelists were included in the soft launch, which began with an initial invitation sent on June 9. All remaining English- and Spanish-speaking sampled online panelists were included in the full launch and were sent an invitation on June 10.<\/p>\n\n\n\n

\"\"<\/a><\/figure>\n\n\n\n

Panelists participating online were sent an email invitation and up to two email reminders if they did not respond to the survey. ATP panelists who consented to SMS messages were sent an SMS invitation with a link to the survey and up to two SMS reminders.<\/p>\n\n\n\n

For panelists who take surveys over the phone with a live interviewer:<\/strong> Prenotification postcards were mailed on June 6. Soft launch took place on June 9 and involved dialing until a total of seven interviews had been completed. All remaining English- and Spanish-speaking sampled phone panelists\u2019 numbers were dialed throughout the remaining field period. Panelists who take surveys via phone can receive up to six calls from trained SSRS interviewers.<\/p>\n\n\n\n

Data quality checks<\/h4>\n\n\n\n

To ensure high-quality data, Center researchers performed data quality checks to identify any respondents showing patterns of satisficing. This includes checking for whether respondents left questions blank at very high rates or always selected the first or last answer presented. As a result of this checking, three ATP respondents were removed from the survey dataset prior to weighting and analysis.<\/p>\n\n\n\n

Weighting<\/h4>\n\n\n\n

The ATP data is weighted in a process that accounts for multiple stages of sampling and nonresponse that occur at different points in the panel survey process. First, each panelist begins with a base weight that reflects their probability of recruitment into the panel. These weights are then calibrated to align with the population benchmarks in the accompanying table to correct for nonresponse to recruitment surveys and panel attrition. If only a subsample of panelists was invited to participate in the wave, this weight is adjusted to account for any differential probabilities of selection.<\/p>\n\n\n\n

Among the panelists who completed the survey, this weight is then calibrated again to align with the population benchmarks identified in the accompanying table and trimmed at the 1st and 99th percentiles to reduce the loss in precision stemming from variance in the weights. Sampling errors and tests of statistical significance take into account the effect of weighting.<\/p>\n\n\n\n

\"\"<\/a><\/figure>\n\n\n\n

The following table shows the unweighted sample sizes and the error attributable to sampling that would be expected at the 95% level of confidence for different groups in the survey.<\/p>\n\n\n\n

\"\"<\/a><\/figure>\n\n\n\n

Sample sizes and sampling errors for other subgroups are available upon request. In addition to sampling error, one should bear in mind that question wording and practical difficulties in conducting surveys can introduce error or bias into the findings of opinion polls.<\/p>\n\n\n\n

Dispositions and response rates<\/h4>\n\n\n\n
\"\"<\/a><\/figure>\n\n\n\n
\"\"<\/a><\/figure>\n\n\n\n

Overview<\/h4>\n\n\n\n

Wave 182 of the American Trends Panel (ATP) was conducted from Oct. 20 to 26, 2025. A total of 5,111 panelists responded out of 5,866 who were sampled, for a survey-level response rate of 87%.<\/p>\n\n\n\n

The cumulative response rate accounting for nonresponse to the recruitment surveys and attrition is 3%. The break-off rate among panelists who logged on to the survey and completed at least one item is 1%. The margin of sampling error for the full sample of 5,111 respondents is plus or minus 1.7 percentage points.<\/p>\n\n\n\n

The survey includes an oversample<\/a> of non-Hispanic Asian adults and parents with children younger than 18 in order to provide more precise estimates of the opinions and experiences of these smaller demographic subgroups. These oversampled groups are weighted back to reflect their correct proportions in the population.<\/p>\n\n\n\n

SSRS conducted the survey for Pew Research Center via online (n=4,945) and live telephone (n=166) interviewing. Interviews were conducted in both English and Spanish.<\/p>\n\n\n\n

To learn more about the ATP, read \u201cAbout the American Trends Panel<\/a>.\u201d<\/p>\n\n\n\n

Panel recruitment<\/h4>\n\n\n\n

Since 2018, the ATP has used address-based sampling (ABS) for recruitment. A study cover letter and a pre-incentive are mailed to a stratified, random sample of households selected from the U.S. Postal Service\u2019s Computerized Delivery Sequence File. This Postal Service file has been estimated to cover 90% to 98% of the population.[5. AAPOR Task Force on Address-based Sampling. 2016. \u201cAAPOR Report: Address-based Sampling<\/a>.\u201d] Within each sampled household, the adult with the next birthday is selected to participate. Other details of the ABS recruitment protocol have changed over time but are available upon request.[6. Email pewsurveys@pewresearch.org<\/a>.] Prior to 2018, the ATP was recruited using landline and cellphone random-digit-dial surveys administered in English and Spanish.<\/p>\n\n\n\n

A national sample of U.S. adults has been recruited to the ATP approximately once per year since 2014. In some years, the recruitment has included additional efforts (known as an \u201coversample\u201d) to improve the accuracy of data for underrepresented groups. For example, Hispanic adults, Black adults and Asian adults were oversampled in 2019, 2022 and 2023, respectively.<\/p>\n\n\n\n

Sample design<\/h4>\n\n\n\n

The overall target population for this survey was noninstitutionalized persons ages 18 and older living in the United States. It featured a stratified random sample from the ATP in which non-Hispanic Asian adults and parents with children younger than 18 were selected with certainty. The remaining panelists were sampled at rates designed to ensure that the share of respondents in each stratum are proportional to its share of the U.S. adult population to the greatest extent possible. Respondent weights are adjusted to account for differential probabilities of selection as described in the Weighting section below.<\/p>\n\n\n\n

Questionnaire development and testing<\/h4>\n\n\n\n

The questionnaire was developed by Pew Research Center in consultation with SSRS. The web program used for online respondents was rigorously tested on both PC and mobile devices by the SSRS project team and Pew Research Center researchers. The SSRS project team also populated test data that was analyzed in SPSS to ensure the logic and randomizations were working as intended before launching the survey.<\/p>\n\n\n\n

Incentives<\/h4>\n\n\n\n

All respondents were offered a post-paid incentive for their participation. Respondents could choose to receive the post-paid incentive in the form of a check or gift code to Amazon.com, Target.com or Walmart.com. Incentive amounts ranged from $5 to $20 depending on whether the respondent belongs to a part of the population that is harder or easier to reach. Differential incentive amounts were designed to increase panel survey participation among groups that traditionally have low survey response propensities.<\/p>\n\n\n\n

Data collection protocol<\/h4>\n\n\n\n

The data collection field period for this survey was Oct. 20 to 26, 2025. Surveys were conducted via self-administered web survey or by live telephone interviewing.\u00a0<\/p>\n\n\n\n

For panelists who take surveys online:<\/strong>[7. The ATP does not use routers or chains in any part of its online data collection protocol, nor are they used to direct respondents to additional surveys.] Postcard notifications were mailed to a subset on Oct. 20.[8. Postcard notifications for web panelists are sent to 1) panelists who were recruited within the last two years and 2) panelists recruited prior to the last two years who opt to continue receiving postcard notifications.] Survey invitations were sent out in two separate launches: soft launch and full launch. Sixty panelists were included in the soft launch, which began with an initial invitation sent on Oct. 20. All remaining English- and Spanish-speaking sampled online panelists were included in the full launch and were sent an invitation on Oct. 21.<\/p>\n\n\n\n

\"\"<\/a><\/figure>\n\n\n\n

Panelists participating online were sent an email invitation and up to two email reminders if they did not respond to the survey. ATP panelists who consented to SMS messages were sent an SMS invitation with a link to the survey and up to two SMS reminders.<\/p>\n\n\n\n

For panelists who take surveys over the phone with a live interviewer:<\/strong> Prenotification postcards were mailed on Oct. 17. Soft launch took place on Oct. 20 and involved dialing until a total of four interviews had been completed. All remaining English- and Spanish-speaking sampled phone panelists\u2019 numbers were dialed throughout the remaining field period. Panelists who take surveys via phone can receive up to six calls from trained SSRS interviewers.<\/p>\n\n\n\n

Data quality checks<\/h4>\n\n\n\n

To ensure high-quality data, Center researchers performed data quality checks to identify any respondents showing patterns of satisficing. This includes checking for whether respondents left questions blank at very high rates or always selected the first or last answer presented. As a result of this checking, three ATP respondents were removed from the survey dataset prior to weighting and analysis.<\/p>\n\n\n\n

Weighting<\/h4>\n\n\n\n

The ATP data is weighted in a process that accounts for multiple stages of sampling and nonresponse that occur at different points in the panel survey process. First, each panelist begins with a base weight that reflects their probability of recruitment into the panel. These weights are then calibrated to align with the population benchmarks in the accompanying table to correct for nonresponse to recruitment surveys and panel attrition. If only a subsample of panelists was invited to participate in the wave, this weight is adjusted to account for any differential probabilities of selection.<\/p>\n\n\n\n

Among the panelists who completed the survey, this weight is then calibrated again to align with the population benchmarks identified in the accompanying table and trimmed at the 1st and 99th percentiles to reduce the loss in precision stemming from variance in the weights. Sampling errors and tests of statistical significance take into account the effect of weighting.<\/p>\n\n\n\n

\"\"<\/a><\/figure>\n\n\n\n

The following table shows the unweighted sample sizes and the error attributable to sampling that would be expected at the 95% level of confidence for different groups in the survey.<\/p>\n\n\n\n

\"\"<\/a><\/figure>\n\n\n\n

Sample sizes and sampling errors for other subgroups are available upon request. In addition to sampling error, one should bear in mind that question wording and practical difficulties in conducting surveys can introduce error or bias into the findings of opinion polls.<\/p>\n\n\n\n

A note about the Asian adult sample<\/h4>\n\n\n\n

This survey includes a total sample size of 559 Asian adults. The sample primarily includes English-speaking Asian adults and, therefore, may not be representative of the overall Asian adult population. Despite this limitation, it is important to report the views of Asian adults on the topics in this study. As always, Asian adults\u2019 responses are incorporated into the general population figures throughout this report.<\/p>\n\n\n\n

How family income tiers are calculated<\/h4>\n\n\n\n

Family income data reported in this study is adjusted for household size and cost-of-living differences by geography. Panelists then are assigned to income tiers that are based on the median adjusted family income of all American Trends Panel members. The process uses the following steps:<\/p>\n\n\n\n

    \n
  1. First, panelists are assigned to the midpoint of the income range they selected in a family income question that was measured on either the most recent annual profile survey or, for newly recruited panelists, their recruitment survey. This provides an approximate income value that can be used in calculations for the adjustment.<\/li>\n\n\n\n
  2. Next, these income values are adjusted for the cost of living in the geographic area where the panelist lives. This is calculated using price indexes published by the U.S. Bureau of Economic Analysis. These indexes, known as Regional Price Parities<\/a> (RPP), compare the prices of goods and services across all U.S. metropolitan statistical areas as well as non-metro areas with the national average prices for the same goods and services. The most recent available data at the time of the annual profile survey is from 2023. Those who fall outside of metropolitan statistical areas are assigned the overall RPP for their state\u2019s non-metropolitan area.<\/li>\n\n\n\n
  3. Family incomes are further adjusted for the number of people in a household using the methodology from Pew Research Center\u2019s previous work on the American middle class<\/a>. This is done because a four-person household with an income of say, $50,000, faces a tighter budget constraint than a two-person household with the same income.<\/li>\n\n\n\n
  4. Panelists are then assigned an income tier. \u201cMiddle-income\u201d adults are in families with adjusted family incomes that are between two-thirds and double the median adjusted family income for the full ATP at the time of the most recent annual profile survey. The median adjusted family income for the panel is roughly $77,800. Using this median income, the middle-income range is about $51,900 to $155,600. Lower-income families have adjusted incomes less than $51,900 and upper-income families have adjusted incomes greater than $155,600 (all figures expressed in 2024 dollars and scaled to a household size of three). A panelist is assigned \u201cno answer\u201d in the income tier variable if they did not provide all three pieces of information needed to calculate their tier (family income, household size and residential address).<\/li>\n<\/ol>\n\n\n\n

    Two examples of how a given area\u2019s cost-of-living adjustment was calculated are as follows: the Pine Bluff metropolitan area in Arkansas is a relatively inexpensive area, with a price level that is 19.7% less than the national average. The San Francisco-Oakland-Berkeley metropolitan area in California is one of the most expensive areas, with a price level that is 18.2% higher than the national average. Income in the sample is adjusted to make up for this difference. As a result, a family with an income of $40,200 in the Pine Bluff area is as well off financially as a family of the same size with an income of $59,100 in San Francisco.<\/p>\n\n\n\n

    Dispositions and response rates<\/h4>\n\n\n\n
    \"\"<\/a><\/figure>\n\n\n\n
    \"\"<\/a><\/figure>\n\n\n\n

    Analysis of health and wellness influencers<\/h3>\n\n\n\n

    To better understand the health and wellness landscape on social media, researchers conducted an in-depth analysis of influencers who discuss health and wellness topics to large U.S. audiences on Instagram, TikTok or YouTube.<\/p>\n\n\n\n

    In the social media analysis portion of this report, health and wellness influencers<\/strong> are defined as people (not organizations or brands) who regularly post health and wellness information on Instagram, TikTok or YouTube. To be included in this analysis, they must have at least 100,000 followers on one of those three social media sites, primarily discuss their content in English, and address a U.S. audience. For the purposes of this research, \u201chealth and wellness\u201d<\/strong> refers to the information or discussion of treatments or lifestyle practices to improve or maintain physical or mental health or wellness. This may include medical advice, personal health testimonials, commentary and other holistic guidance for how to live a healthy lifestyle.<\/p>\n\n\n\n

    The final analysis was conducted on a sample of 6,828 health and wellness influencers that were identified through a combination of targeted health and wellness keyword searches and site auditing.<\/p>\n\n\n\n

    Key terms search<\/h4>\n\n\n\n

    To focus our analysis on only influencers who discuss health and wellness topics, we started by developing a list of search terms that are relevant to wellness trends and current health discussions in the United States. We conducted a topic search for each key term on each of Instagram, TikTok, and YouTube using Modash<\/a>, a third-party social media marketing data platform.<\/p>\n\n\n\n

    From June to September 2025, five researchers brainstormed key terms across 14 health and wellness categories. These categories included physical health, mental health, diet and nutrition, beauty, supplements and vitamins, health philosophies, medical practices, women\u2019s health, natural health, fitness, cosmetic treatments and surgeries, dental, sleep, and health devices. The initial list of topic key terms was built based on a review of existing literature on health and wellness influencers, qualitative analysis of topics mentioned by popular health and wellness influencers, and input from research team members with expertise in health and science. The topic key term list was further refined with iterative tests of search results across Instagram, TikTok and YouTube on Modash. Since each site is slightly different, we adjusted some of our search terms to return the largest set of results on each site. We also took note of algorithmically recommended similar searches and considered these for inclusion in our search term list.<\/p>\n\n\n\n

    For each search, we applied filters to get accounts that are based in the U.S., are operated by people (rather than businesses or organizations), post primarily in English, and have at least 100,000 followers. In total, we conducted the influencer account search with 333 key terms on Instagram, 334 on TikTok and 270 on YouTube from Sept. 19 to Oct. 9, 2025.<\/p>\n\n\n\n

    These searches resulted in 22,444 Instagram accounts, 21,152 TikTok accounts and 2,552 YouTube accounts (46,148 accounts in total).<\/p>\n\n\n\n

    Identifying additional influencers from podcasts<\/h4>\n\n\n\n

    In the survey portion of this analysis, we asked U.S. adults about their experiences getting health and wellness information from social media influencers and<\/em> podcasts. Although this report does not include findings from sites other than Instagram, TikTok, and YouTube, podcasts are an important part of the health and wellness influencer landscape. To ensure that popular podcast hosts were included in our set of influencers, we collected the Spotify health and fitness podcast charts<\/a> (Top 50) and the Apple Podcasts health and fitness top shows<\/a> (Top 100) each day from Aug. 12 to Sept. 14, 2025. Then, members of the research team manually reviewed the top podcasts that appeared most frequently in the Spotify and Apple charts.<\/p>\n\n\n\n

    For each podcast, we checked whether the host of the podcast was a health and wellness influencer on Instagram, TikTok or YouTube. Separately, we also checked accounts for each influencer on the TIME100 Creators 2025<\/a> list. If we found that these podcast hosts and creators were health and wellness influencers but were not already included in our dataset of accounts, we collected their account data and added it to our dataset on Jan. 15 and 16, 2026. This process resulted in the addition of four influencer accounts \u2013 the vast majority of the accounts we found were already included in our dataset.<\/p>\n\n\n\n

    Account data collection<\/h4>\n\n\n\n

    For each Instagram, TikTok and YouTube account that appeared in our health and wellness search results, we collected key account information including the account display name, username, bio text (or description text for YouTube), follower count, and up to 10 most recent posts. For the recent posts, we collected text data but not audio data or transcripts. Instagram and TikTok account data was collected using the Modash Raw API<\/a>, which provides real-time data from public social media profiles, while YouTube account data was collected directly using the YouTube Data API<\/a>. We ran the account data collection pipeline for all 46,148 accounts as soon as they appeared in our search results from Sept. 19 to Oct. 9, 2025.<\/p>\n\n\n\n

    Removing inactive accounts<\/h4>\n\n\n\n

    To ensure we were only including accounts that were currently active, we used the account data to remove accounts from our dataset that had fewer than three posts in the last 90 days since account collection. After removing these inactive accounts, we had 35,325 accounts to analyze, including 18,581 from Instagram, 15,469 from TikTok and 1,275 from YouTube.<\/p>\n\n\n\n

    Health and wellness influencer verification<\/h4>\n\n\n\n

    Next, we verified whether each of the 35,325 accounts we had collected belonged to health and wellness influencers. We were primarily concerned with two criteria: Did they regularly discuss health or wellness information? And do these accounts belong to individuals, rather than organizations or businesses?’<\/p>\n\n\n\n

    To determine the extent to which these accounts had relevant content, researchers used the health and wellness definition provided above. Health and wellness information can include testimonials, commentary, advice, and product recommendations from conventional or alternative medicine as well as day-to-day healthy living. To be included, each account needed to discuss health and wellness topics in at least 30% of their 10 most recent posts. If an account appears to belong to a business or organization and not to individuals, then we did not label them as health and wellness influencers. After finalizing the health and wellness influencer definition, two members of the research team labeled a random sample of 722 accounts from our set of Instagram, TikTok and YouTube accounts. For each account, the coders determined whether the account belonged to a health and wellness influencer, as defined above. The coders reached 88% agreement (Cohen\u2019s kappa = 0.752), and all instances of disagreement between coders were resolved by a third member of the research team who was familiar with the health and wellness influencer definition.<\/p>\n\n\n\n

    \"\"<\/a><\/figure>\n\n\n\n

    OpenAI\u2019s GPT-4.1 mini model was then used to classify all 35,325 active social media accounts in our dataset. For each account, we prompted the model with a codebook that included details about our health and wellness influencer definition, along with the account\u2019s bio text, username, and text from up to 10 of their most recent posts. Refer to the prompt<\/a> for more information about model input and the health and wellness influencer definition. Based on the human-annotated validation dataset, the model achieved 79% accuracy and an F1 score of 0.826. Refer to the \u201cModel performance: Influencer identification\u201d table for detailed performance metrics.<\/p>\n\n\n\n

    Overall, the machine classifier\u2019s recall score was higher than its precision score, suggesting that the set labeled as health and wellness influencers included false positives. We found that false positives were particularly prevalent for YouTube accounts compared with the other two platforms. To mitigate this, one member of the research team who had previously been trained on identifying these influencers manually reviewed all YouTube accounts that were classified as health and wellness influencers. Based on this manual review, we removed any false positive YouTube accounts that were not health and wellness influencers from our dataset.<\/p>\n\n\n\n

    Based on this analysis, 8,428 accounts of our original set of 35,325 were classified as belonging to a health and wellness influencer.<\/p>\n\n\n\n

    Influencer account matching<\/h4>\n\n\n\n

    Our data collection process took place on an account-level basis, but many influencers have a presence across multiple sites. For example, one individual influencer with a consistent online persona might maintain popular accounts on both Instagram and TikTok. Therefore, to match influencers across sites, we used two different data sources: 1) links to other accounts found in the bio of each account we collected and 2) Modash influencer profile reports.<\/p>\n\n\n\n

    For each health and wellness influencer account in our dataset, we collected any links that were available in the account bio. For bios that included a link to an Instagram, TikTok or YouTube account, we checked whether the linked account was already available in our dataset. If the linked account was in our existing dataset, then the two accounts were marked as owned by the same influencer. If not, then the linked account was added to our dataset before being marked as owned by the same influencer as the original account. In cases where the linked account was previously labeled as not being a health and wellness account, the influencer was removed from our final dataset.<\/p>\n\n\n\n

    Instagram, TikTok and YouTube limit the number of links that a user can add to their bio. For this reason, some influencer bios include links to a landing page that aggregates many links related to that influencer, rather than linking to their other social media accounts in their bios directly. These sites, known as \u201cmicrosites,\u201d are common among content creators. To match more influencer accounts across sites, we collected data from microsites linked in account bios for the following microsite providers: Linktree, Stan Store, Beacons, hoo.be, Milkshake and Bio Sites. These were the most common sites used by the influencers in our dataset that were straightforward to collect with automated methods. We identified Instagram, TikTok and YouTube account links that were available on these microsites and marked them as belonging to the same influencer as the original account.<\/p>\n\n\n\n

    Finally, we also matched accounts across sites using influencer profile reports from Modash (the same social media data provider used for account search and data collection) using their Discovery API<\/a>. This report includes an influencer\u2019s contact details, which contains information on their known social media accounts. We used this information to further match and consolidate the accounts in our collection to the individual influencers that own them.<\/p>\n\n\n\n

    Data from additional accounts identified during the account consolidation process was collected from Jan. 22 to 27, 2026, using the account data collection process described above. In total, this account matching process identified 5,441 additional influencer accounts from the three social media sites that we studied.<\/p>\n\n\n\n

    At this stage, we conducted a secondary language filtering step to ensure that all remaining influencer accounts posted primarily in English. We used the langdetect Python package<\/a> to remove influencers for which more than 20% of recent posts across all accounts were in a language other than English.<\/p>\n\n\n\n

    These processes resulted in a final dataset of 6,828 influencers for analysis. From these influencers, we collected information on 12,800 accounts that consisted of:<\/p>\n\n\n\n