The media ecology of weight loss on TikTok, Instagram and Telegram
A networked media mapping of weight loss
Ryland Shaw, Paula Helm, Susan Zijp, Simon Bakker, Laura Caroleo, Sabine Niederer, Carlo De Gaetano, Jeroen de Vos, Luuk Ex, Maarten Groen
0. Summary of key findings
Media ecology of ‘weight loss’. Find the full pdf here
(warning: big file; may take some time to open!).
In recent years, scholars and journalists have raised concerns about potentially harmful weight loss content (e.g., self-medication, extreme diets/fitness and eating disorders) being presented to even underaged publics on social media (Gerrard 2018; Conger, Browning and Woo 2021; Davey, 2023). And where content moderation and fact-checking focus on the flagging, deleting and debunking of such misleading and at times toxic content, Philips and Milner (2021) have proposed a focus on media ecologies rather than individual harmful content and sources. In an attempt to operationalize this networked media approach, we have chosen to map the sensitive topic of weight loss as it appears on the social media platforms of TikTok
, Instagram and Telegram.
In an exploratory study of this topic across the three platforms, we make the act of mapping and tracing central to our inquiry. Analysing co-hashtag networks, (most-engaging) content, top URLs and users, we see the three platforms present themselves differently with some commonly shared characteristics.
is a motivational space where people support each other in their weight loss journey - sometimes for commercial benefits. Users invite others on a weight loss journey while showing their own ‘before’ and ‘after’ pictures, often supported by music, such as “No Role Modelz" by J. Cole. TikTok
tries to keep users safe, but often harmful content is only a hashtag away.
On Instagram, users similarly share their weight loss journey. However, the narrative seems more focused on displaying individual progress with less visibility of peer support. Posts promoting products and diets with sometimes questionable promises of rapid weight loss are prominent and occur next to tips and tricks on a healthy lifestyle.
Telegram is a primarily commercial space, promoting and selling workout programs, weight loss tools, or pills and powders for burning fat or gaining muscles.
Overall, the topic of weight loss is delicate both in terms of its content and its ecology, with potentially harmful content appearing in the near vicinity of supportive and motivational posts.
Media literacy aims to empower citizens to critically engage with their (increasingly personalised) information and media landscape. Media literacy efforts have gained urgency in the rise and rapid spread of misinformation, disinformation, fake news and other problematic information on social media and beyond. However, media literacy efforts, including corrective information, fact checking, and pre- and debunking have been topics of discussion due to mixed effects. Scholars have warned for possible backlashes, such as polarisation and increasing mistrust in journalism and institutions (Bounegru et al. 2018; Chan et al. 2021; Hameleers & van der Meer 2020; Phillips & Milner, 2021; Vraga et al. 2022).
In response, Phillips and Milner (2021) have put forward the concept of ‘ecological literacy’, which calls for a focus on media ecologies rather than individual harmful content and sources. Their concept of ecological literacy, a networked media literacy approach, is helpful here. In this winter school project, we aim to operationalise this notion, that moves from the single post to a networked view, by tracing the spread of (problematic) messages in the realm of sensitive topics such as dietary advice, current climate activism and blockchain investment opportunities.
We aim to do so not by starting from the assessment of information as being true, false, or misleading, but rather by making the act of mapping and tracing content central to our inquiry, scrutinising how specific information travels through different platforms and how it may transform through the act of circulating. Such tracing may lead to views on how the message is being reinforced, complicated, annotated, endorsed, criticised or debunked throughout its dissemination.
Starting point for the analysis is Instagram, Telegram & TikTok
, - three platforms have been under scrutiny or criticised in popular media for pushing problematic content to a younger audience (Milmo & Hern, 2022, Veranen, 2022, Onibada 2021) - where messages on one or more of the following topics will be traced across platforms regarding health & diet tips. Recent articles point out teenagers are confronted with highly problematic online (English-language) content on dieting advice, with terms like ‘weight loss’ sometimes bordering on pro-ana voices, or with the hidden agenda of selling surgery or vitamin supplements. How does this kind of problematic information present itself, and how does that differ across platforms? (Also see Milmo & Hern 2022, Beeld & Geluid 2022)
2. Initial Data Sets
Top 1000 posts on Tiktok, Instagram & Telegram when querying for Weight Loss
Datasets can be requested with the researchers.
3. Research Questions
RQ: What does the information ecology of weight loss look like on the platforms of TikTok
, Instagram and Telegram?
Sub-RQ: Who are the influencers in this environment, what sounds can be heard, which objects and sources are presented, where do we find pollution and toxicity, what content is looming on the edges?
4. Methodology and initial datasets
Research protocol. High res pdf can be downloaded here
We started with the query ‘Weight loss’ OR ‘Weightloss’, and fed this to the three respective platforms (Instagram, TikTok
& Telegram). First, we set up clean research accounts for Instagram & TikTok
with a fictitious user, born January first 2005 - to reflect a young adult, just out of the minor age. Data collection happened using 4CAT with the Zeeschuimer plugin installed.
In the next step, we used the query, and harvested the top 1000 posts that were shown as search results on the web version of Instagram and TikTok
using the Zeeschuimer browser plugin. Post data collected was exported to the 4CAT server for further analysis. For Telegram, we queried for most prominent public groups using ‘weight loss’ in their title, to scrape a total of 1000 messages from the top 10 groups.
For data analysis purposes we roughly used three different techniques:
- Visual analysis using the ImageSorter. Image sorter allows to plot the visual part of the posts on a larger grid, organised by visual similarities (like hue or lightning). This is a way to organise content by similarities to get a more general understanding of visual vernacular used to communicate about weight loss.
- Close reading the top 50 posts of Instagram & TikTok. We used close reading strategies as a way to ‘spend time with our data’, to better understand what we were looking at - and try to give some early answers to the research questions. Insights gathered here are also early indicators for the landscape to design and visuals to use.
- Textual analysis. For the textual analysis we drew both on the hashtags used and the subsequent co-hashtag analysis to try and understand the larger conversation happening around ‘weight loss’, this allowed us to scale the insights gathered through close reading, to also grasp how certain topics redirect people to more other neighbouring topics to allow people wander to more fringe and sometimes problematic subjects. Lastly, we extracted the urls in the body or account bio (Instagram/TikTok and Telegram respectively) to try and understand what other platforms and content was being referred to. This gives a better understanding of the way in which a particular discussion is connected and referred to - it helped lay the land for our map making efforts trying to display the ecology.
- Other analyses include a hashtag / song correlation analysis for TikTok, to look into the functioning of songs as an organising mechanism as well as the manual coding of the URLs.
. In the last step we used the insights from the research to co-design the map. This process fits the tradition of allegorical maps, where landscapes and architectural elements represent concepts and narratives. The map below uses visual elements from ‘The land of make-believe’, a poster by Jaro Hess from the 1930s that represents different fairy tales in one larger landscape. There, we find known characters like the Pied Piper, Jack and the Beanstalk, and Jack and Jill. In this visualisation, we similarly combine different narratives that occur in the online spaces on weight loss. On this map, castles represent platforms, trees illustrate co-hashtags, and pathways may lead to other platforms and hashtag spaces. Behind the mountains, potentially harmful zones appear as looming monsters.
The visualisation below represents the landscape you might encounter when turning to social media for advice on weight loss. The map can be considered a creative device for exploring existing narratives and composing new stories. In line with the concept of ecological literacy, we invite viewers to reflect on this landscape. What stories would you tell, moving through the landscape? What surroundings look familiar, where do you still know your way and direction - but also, at what point might you get disoriented, get lost in the thick woods, or slip down dangerous slopes?
Media ecology of ‘weight loss’. Find the full pdf here (warning: big file; may take some time to open!).
The findings below unpack the different journeys present on the map, it does so by explaining what we found to support the landscape we outlined above.
TikTok is a motivational space where people support each other in their weight loss journey - sometimes for commercial benefits. Users invite others on a weight loss journey while showing their own ‘before’ and ‘after’ pictures, often supported by music, such as “No Role Modelz" by J. Cole. From the top engaged with posts 58 percent are before/after video following a similar weight loss vernacular. Posts spotlit dramatic transformations, often omitting the work that happens between the before and after.
tries to keep users safe, by directing users to a warning page about eating disorders
, where TikTok
encourages users to look for professional help. When searching for terms weightlosspills or weightlossjourney (found co-hashtagged in 11% of videos) TikTok
takes the users to the warning landing page.
However, when searching for weightloss a user can view many videos containing the hashtags #weightlossjourney and #weightlosspills. A user then can click on the potentially harmful hashtags, and view many videos without bumping into the warning page. A user can thus easily bypass warning red flags put up by TikTok
for potential harmful content. Similarly, the co-hashtags #obesitymedication and #500Calories were not caught by TikTok
’s professional help filter and contain potentially harmful content. Half of the accounts that posted the top engaged videos direct to commercial platforms where they try to sell products, coaching programs, or ask for donations. Post with #weightloss often include with hashtags related to calorie counting and calorie deficit. Videos in which users are sometimes encouraged to eat as little calories as possible.
From TikTok to other platforms
TikTok does not allow users to link in a video post or subtext in the post. However there is the possibility to lure viewers away from the platform via links in individual accounts. We counted how many users in the most engaged category with posts linked to other social media platforms, including within Linktrees, which is commonly used to circumvent TikTok
’s limitation of only one link in the author’s biography. Most referred to platform is Instagram (40 percent). 12 percent outlink to Twitch, which was not linked to on Instagram or Telegram. Personal email addresses are also often used to get in touch outside of TikTok
(32 percent). Amazon and other ecommerce sites with affiliate links or wishlists were frequently linked (30 percent), in contrast to the explicit product marketing rife on Telegram.
The visual vernacular (Gibbs, 2015) of #weightloss on TikTok
is strong. There are several key elements that are imitated throughout the genre (Zulli & Zulli, 2022). Users usually depict themselves in their home trying different poses to illustrate their weight loss. Before/after videos are very common, often captioned with their weight loss amount and time frame, with many time frames being exceptionally short (i.e. 100lbs in 6 months), collapsing potentially implying drastic and harmful techniques (example 1
, example 2
). Transformation videos typically make use of music only, no voice or other sound effects. Certain communities associated themselves with specific songs.
- No Role Modelz - j. cole (12 occurrences)
- The One that Got Away - Katy Perry (12 occurrences)
- Love You So - The King Khan & BBQ Show (10 occurrences)
For Instagram from the initial dataset containing just over 1,000 posts (exactly 1054), we decided to do a close reading of the top 50 posts, ordered by engagement (the sum of likes and comments).
On Instagram, users also share their weight loss journey. However, the narrative seems more focused on displaying individual progress, with less visibility of peer support, although there are hashtags that link to communities such as #wlscommunity and others. The tone of voice on Instagram is motivational, but this motivation is infused with claims such as 'stronger body, stronger mind'. At first glance, this does not seem harmful, yet posts promoting products and diets with sometimes questionable promises of rapid weight loss are prominent, and these messages are found alongside tips and tricks on a healthy lifestyle. For example, the 'flat tummy water' (red water, lime juice, cucumber, ginger and mint leaves) claims to 'eliminate toxins'.
Dataset overview of images
Colours in some sections describe healthy opposed to unhealthy meals
In analysing the images of the entire dataset, it was interesting to note that junk or unhealthy food appears in colours such as orange or red, while healthy and low-calorie foods appear in vivid colours such as yellow and green - colours which in colour psychology are often associated with happiness, joy and energy (Moon, Chang Bae et al., 2015).
In all the posts where people show their progress and weight loss, which they call “before-and-after”, the images convey that one side is better than the other. In fact, in many of these before-and-after posts, the before posts show a sad expression, and the after (after the journey) shows the same person with a smile.
Commercial agendas on Instagram
We have seen that 59% of the accounts in the top 50 results of the most engagement promote steps, such as "link in bio", that prompt the purchase of products or services aiding weight loss or work out enhancement. The type of users of the site are entrepreneurs who most often have affiliation agreements (and/or are ambassadors) with companies that sell products. They also seem to be building a community around health, but we found that the call to action under Instagram photos often promotes clicking on the link in the bio, which leads to a personal website, YouTube
or Linktree. The latter application leads to multiple links to other social media platforms, including less popular ones such as Onlyfans. Some of Linktree's links also contain links to direct payment, whereby other people can donate money directly.
The network above is a co-hashtag analysis which details how hashtags co-occuring in the same post. This gives a sense of the different sub conversations you might encounter when querying for weight loss on Instagram. Colours represent statistical communities. The network mostly shows that there are different ongoing conversations, which roughly breakdown as follows:
- The most prominent and connected one are the posts that concern workout, fitness and bodybuilding
- The second, slightly less connected discussion is of weight loss, and its instruments used for weight loss (specific, tips, diets, ways of counting and calculating) and lastly
- Smaller sub conversations conclude food related posts, and conversation around slimming.
The Telegram space is a primarily commercial space, promoting and selling workout programs, weight loss tools, or pills and powders for burning fat or gaining muscles. Some invite you to pay with crypto currencies. The posts shown on the investigated channels often depict fitness gurus at the gym, showing their abs.
More toxic Telegram channels on weight loss and fitness engage in body shaming. Men posing with guns, ‘hunting’ the obese. Another narrative on Telegram posed that a fit and healthy body might not need a covid vaccination. There were also cases found where channels were promoting potentially harmful drugs.
The top hashtags on Telegram are about motivation, fitness and food. The context in which these hashtags are placed are mostly healthy food and weight loss journeys. Hashtags are rare on Telegram, most are copy pasted from other platforms. When looking at There were a few links directing to Facebook and Pinterest, and a lot more links pointing to Youtube. From the three researched platforms, only Telegram outlinked to Pinterest or Facebook.
Overall, the topic of weight loss is delicate both in terms of its content and its ecology, with potentially harmful content appearing in the near vicinity of supportive and motivational posts. It needs noting that a subsequent journey for any user might be a walk over the ridge of a cliff, as well as a well demarcated footpath.
Therefore it is important to underline that translating from insights we got when encountering the content we encountered - to the collective map making exercise coils use more attention. Specific insights and questions that came up when making the map together include:
- We currently worked with a dataset that could be considered a self-ethnographic journey, the researchers started with a similar starting point: the query weight loss. How to better include and represent the perspective of the users themselves?
- Platforms are paving the way - thereby also including checks-and-balances and mitigations strategies. First exploration suggests that users are only given a warning sign redirecting to self-protection when directly searching potentially harmful words. When ‘organically’ redirected from existing posts (eg. through a hashtag), no warning is given. This would need more research.
- In the currently drafted map, the platforms are regions part of a bigger territory.. When encountering both implicit and explicit connectedness of the landscape, how to represent this on a map without losing the complexity of the ecology / or using the ‘map-reader’ over the complexity?
- How to best represent the commercial agenda which we often found present, but only looming one or two clicks away?
How to understand one's position in a larger media ecology? Where to find the ‘you are here’ and how to subsequently orientate yourself? This research has been an early attempt to operationalize the idea of ‘ecological literacy' (Phillips and Miner 2021), by focussing on the media ecology rather than individual content and sources. Taking the act of mapping and tracing central to our inquiry we tried to better understand the landscape one is moving through when looking for advice on weight loss on three different platforms, TikTok
, Telegram and Instagram. The map we produced is the output of exactly such a collective mapping exercise, trying to capture in a landscape the different roads, connections and relations between different territories and walks we encountered in our research. In further research we would like to explore the double function of the map - not only in the way it allows to collectively produce a specific media ecology imaginary - but also in the sense that is provides a tool for narrating potential journeys one could make travelling the mountains of Instagram, and hiking the little uphill streams of TikTok
Beeld & Geluid (2022, December 15). Tiktok door de ogen van Jongeren deel 1. Beeld en Geluid, blog post, retrieved from https://beeldengeluid.nl/kennis/blog/tiktok-door-de-ogen-van-jongeren-deel-1
Bounegru, L., Gray, J., Venturini, T., & Mauri, M. (2018). A Field Guide to" Fake News" and Other Information Disorders: A Collection of Recipes for Those Who Love to Cook with Digital Methods, Public Data Lab, Amsterdam (2018).
Chan, M. S., Jones, C. R., Hall Jamieson, K., & Albarracín, D. (2017). Debunking: A Meta-Analysis of the Psychological Efficacy of Messages Countering Misinformation. Psychological Science, 28(11), 1531–1546. https://doi.org/10.1177/0956797617714579
Gibbs, M., Meese, J., Arnold, M., Nansen, B., & Carter, M. (2015). #Funeral and Instagram: Death, social media, and platform vernacular. Information, Communication & Society, 18(3), 255–268. https://doi.org/10.1080/1369118X.2014.987152
Hameleers, M., Powell, T. E., van der Meer, G. L. A., & Bos, L. (2020). A picture paints a thousand lies? The effects and mechanisms of multimodal disinformation and rebuttals disseminated via social media. Political Communication, 37, 281-301. https://doi.org/10.1080/10584609.2019.1674979
Milmo, D & A. Hern (2022, December 15). Tiktok self-harm study results: 'Every parent's Nightmare'. The Guardian. Retrieved from https://www.theguardian.com/technology/2022/dec/15/tiktok-self-harm-study-results-every-parents-nightmare
Moon, Chang Bae et al. Mood Lighting System Reflecting Music Mood. Color research and application 40.2 (2015): 201–212. Web.
Onibada, A. (2021, July 1). Influencers in Norway will legally have to disclose their photoshopped images. BuzzFeed News. Retrieved January 26, 2023, from https://www.buzzfeednews.com/article/adeonibada/influencers-norway-law-filter-photoshop
Phillips, W., & Milner, R. M. (2021). You are here: A field guide for navigating polarized speech, conspiracy theories, and our polluted media landscape. MIT Press.
Vermanen, J. (n.d.). Naaktbeelden van Honderden slachtoffers per maand in telegram-groepen gedeeld. KRO. Retrieved January 26, 2023, from https://kro-ncrv.nl/naaktbeelden-van-honderden-slachtoffers-per-maand-in-telegram-groepen-gedeeld
Vraga, E., Tully, M., & Bode, L. (2022). Assessing the relative merits of news literacy and corrections in responding to misinformation on Twitter. New Media & Society, 24(10), 2354-2371.
Zulli, D., & Zulli, D. J. (2022). Extending the Internet meme: Conceptualizing technological mimesis and imitation publics on the TikTok platform. New Media & Society, 24(8), 1872–1890. https://doi.org/10.1177/1461444820983603