Exploring Mental Health Communications among Instagram Coaches

Ehsan ul Haq, Lik-Hang Lee, Gareth Tyson, Reza Hadi Mogavi, Tristan Braud, Pan Hui

Introduction

With mental health issues affecting one in five Americans, the stigma surrounding traditional clinical settings often drives individuals toward Online Social Networks (OSNs) for guidance. Instagram, in particular, has seen the rise of a new class of "mental health influencers"—a diverse group ranging from licensed therapists and psychologists to life coaches and authors. Despite their significant reach, little was known about how these professionals communicate or what specific linguistic strategies drive audience engagement. This study analyzes nearly 98,000 Instagram posts from 127 verified accounts to quantify the relationship between professional identity, linguistic style, and user interaction.

Methodology

The researchers employed a robust computational pipeline to analyze a dataset of 97,918 posts collected via the CrowdTangle API between 2019 and 2021. The methodology consisted of four primary pillars:

  1. Account Classification: Using FastText word embeddings and K-Means clustering, the researchers categorized influencers into five distinct groups based on their self-declared biographies (e.g., "Authors," "Coaches," or "Therapists").
  2. Linguistic Analysis: The text of post captions was processed using Linguistic Inquiry and Word Count (LIWC) 2015 to identify semantic categories such as affective processes, cognitive processes, and biological drives.
  3. Support Quantification: To distinguish between types of help, the team used pre-trained Random Forest and XGBoost classifiers to label posts as providing Information Support (IS)—facts and advice—or Emotional Support (ES)—empathy and encouragement.
  4. Engagement Modeling: Because "likes" follow a non-Gaussian distribution, the study utilized Negative Binomial Regression to predict engagement. This model normalized for follower counts to ensure that results reflected the quality of communication rather than just the size of the account.

Professional Identity and Support

The study found that professional labels are strong predictors of the type of support offered. For instance, accounts identifying primarily as "Coaches" tend to provide higher levels of both emotional and informational support compared to those identifying as "Authors." This suggests that the self-adopted label influences the creator's perceived role and their subsequent communication strategy on the platform.

The Language of Engagement

One of the study's most striking findings is the "Negative Emotion Paradox." While conventional wisdom might suggest that social media audiences prefer positivity, the data revealed that the explicit use of words categorized as "Negative Emotions" (e.g., anxiety, trauma, anger) actually increases engagement. Specifically, adding one word related to negative emotion increases engagement by 0.01x. The researchers suggest that being direct about struggles makes content feel more authentic and relevant to followers seeking help.

Furthermore, the study highlighted a preference for Information Support over Emotional Support. While empathy is valued, audiences on Instagram appear to prioritize concrete advice and factual knowledge. For certain account types, such as those combining therapist and coach identities, moving from low to medium Information Support resulted in a 0.2x increase in engagement.

Platform Friction and Audience Behavior

The research also identified a significant penalty for "platform friction." Influencers who included calls-to-action leading users away from Instagram—such as links to external websites or other social platforms—saw a notable decrease in engagement. This suggests that both the Instagram algorithm and user behavior favor content that keeps the experience within the app's ecosystem.

Conclusion and Future Directions

This research provides a data-driven look at the burgeoning field of digital mental health advocacy. It reveals that engagement is driven by a complex interplay of professional identity and linguistic specificity, where directness and informational value often outweigh general emotional encouragement.

However, the study notes several limitations. The reliance on "likes" as a proxy for engagement does not capture the qualitative depth of a post's impact on a user's well-being. Additionally, the analysis focused exclusively on text, omitting the visual and video elements that are central to the Instagram experience. Future work aims to incorporate image analysis and sentiment analysis of user comments to provide a more holistic understanding of how mental health communication functions in the digital age.

Summary updated

Abstract

There has been a significant expansion in the use of online social networks (OSNs) to support people experiencing mental health issues. This paper studies the role of Instagram influencers who specialize in coaching people with mental health issues. Using a dataset of 97k posts, we characterize such users' linguistic and behavioural features. We explore how these observations impact audience engagement (as measured by likes). We show that the support provided by these accounts varies based on their self-declared professional identities. For instance, Instagram accounts that declare themselves as Authors offer less support than accounts that label themselves as a Coach. We show that increasing information support in general communication positively affects user engagement. However, the effect of vocabulary on engagement is not consistent across the Instagram account types. Our findings shed light on this understudied topic and guide how mental health practitioners can improve outreach.
BibTeX
@inproceedings{haq2022exploring,
  author = {Ehsan ul Haq and Lik-Hang Lee and Gareth Tyson and Reza Hadi Mogavi and Tristan Braud and Pan Hui},
  title = {{Exploring Mental Health Communications among Instagram Coaches}},
  booktitle = {ASONAM},
  year = {2022},
  url = {https://openreview.net/forum?id=PcjtY3zh85E},
}
Figure