Can we create a different kind of social network? Learning from Mastodon


This purpose of the proposal is to use a mixed methods approach to analyze the current use of is to further investigate anecdotes regarding the sense of community, intimacy and depth of relationship on the site using its publicly available data and discussions with users.


Research questions

  1. What are differences with respect to how Mastodon and Twitter are being used?
  2. How are Mastodon-specific tools including tags, content warnings, local and federated timelines, post-level privacy controls being used by the community?
  3. Are there any other notable characteristics of Mastodon and how it is administered?



In October 2016, a small number of open educators and educational technologists began to explore and experiment with an open-source Twitter-like social network space called Mastodon. Mastodon is “a free, open-source social network. A decentralized alternative to commercial platforms, it avoids the risks of a single company monopolizing your communication… Anyone can run their own Mastodon instance and participate in the social network seamlessly.” (


Between November 2016 and April 2017, its flagship instance has grown from under 10,000 to almost 50,000 registered users. There are reportedly close to currently 150,000 on all federated instances (


While its most noticeable difference from Twitter is its 500 character limit, Bali identified several other benefits including being open-source, in its infancy, content warning labels and its intimacy ( Lynds and Richard also wrote an early article identifying the potential for this type of site (


Further conversations with the developers noted the careful consideration of the needs of marginalized groups that drove both software design decisions and the site administrator’s approach to moderation and community-building.

“In the early days of the project some of our most meaningful criticism came from less privileged folks and Gargron et al. did a good job of adding features to address content ownership and privacy concerns. Idealogues who are mostly worried about exact ostatus compatibility don’t see the value or don’t even see these things as positive changes. They are more interested in the engineering details of the project than the human impact.” – personal communication with @Trev (since deleted from Mastodon)


As the has continued to grow, anecdotal evidence points to the possibility that from those roots, a new type of community is forming.

“So that’s also a beautiful mastodon thing: friendship retrieves you, even among vagabonds like us.”  – Kate Bowles (


“One of the most valuable things about Mastodon is how @Gargron is building an entirely new platform with comparatively few resources and maintaining that platform’s openness. I think there’s a lot to model in what is being accomplished through Mastodon; how do we build new things in open ways that can be sustained by communities and not legacy administration? – Chuck Pearson


Proposed methodology

The research will begin with a literature review of the types of analyses that were conducted using early Twitter data to identify a series of meaningful quantitative and qualitative approaches to describe the mastodon network and compare it to Twitter.


Data from the Mastodon API will be pulled, cleaned and reviewed. It is expected that the types of information that will be analyzed will include the post length, length and complexity of discussion threads, number of posts and rate of posts, active length of threads and size and complexity of networks. This analysis will first be completed for the entire instance and then possibly for one or more segments of users with a focus on identifying the specific uses of academic and scholarly participants.


Once the Mastodon dataset is better understood, the specifications required to pull an equivalent subset of Twitter data. Efforts will be made to identify a dataset that is as similar as possible depending on the types of information gathered by the APIs. The descriptive results from the Mastodon and Twitter datasets will be compared.


It is hoped that most of the preliminary work can be completed using a combination of open-source tools that combine text and code like Jupyter Notebooks and Python. This approach is preferred as it allows others with varying levels of technical expertise to follow the steps in the analysis providing them with open access to both the data and the analysis. I used this approach last year to analyze an open data set with a reasonable level of success.


Further computer-assisted segmenting and well as network and content analysis will be completed using both the Mastodon and Twitter datasets in alignment with the methods identified in the literature review with a preference for the use of tools that support open access to both the data and the analysis. Depending on the findings, further qualitative content analysis may be of value.


I will then seek to share these findings and discuss them with a subset of Mastodon users, likely users involved in scholarly and academic work. These discussions will likely take place a synchronously in the Mastodon platform and synchronously using other tools.


Synchronous sessions may be completed privately or in groups. It will also likely be possible to interview a number of academic and scholars who joined Mastodon in the first wave, but have not remained in the space to determine how their experiences align with those who have chosen to stay.


The focus of these sessions will be to discuss with the participants the results of the data analysis to gather their thoughts on those findings. They will then also discuss in more detail how they are using/ used the various Mastodon-specific features and gather their perceptions on the value of those features.


Combining the results of these quantitative and qualitative approaches should provide robust insight to the structural and communal similarities and differences of communities, particularly academic and scholarly communities, uses of these tools.


Initial comments from reviewers:

We have reviewed your application and we are interested in having a followup conversation with you. In particular we’d like to chat about research design. We are interested in exploring what sorts of comparisons you think might be made between Twitter and Mastodon. For instance, who would we be comparing and what might we be looking to establish? It seems that the platforms serve very different groups. Just a few questions for you to ponder for this meeting 🙂