Start academic social media research by confirming an approved data-access route for the proposed project. Then define the sample, record the collection method, plan data handling and obtain the required institutional review. Public visibility, an accessible search tool and an ethics approval do not by themselves establish platform permission. This guide covers study design after those access questions are resolved.
Confirm the platform’s research access route
Reddit’s developer access overview identifies Reddit For Researchers as its authorized academic research route. An ordinary Data API client or third-party tool is not a substitute for that program’s approval.
TikTok Research Tools has eligibility criteria and project approval. Match the researcher, institution and purpose to those requirements before designing the collection.
Check the source-specific requirements for every network. A general public search product, an export feature or an MCP integration does not grant research-program eligibility or override a source’s restrictions.
Start from the research question
The question helps choose the network, unit of analysis and sample. The examples below describe candidate data types, conditional on an access route and project approval that allow the study. They are not instructions to collect those records through any available tool.
| Example question | Data that fits | |
|---|---|---|
| How people explain a problem | How do renters describe damp and mould problems? | Approved Reddit research records or permitted YouTube comment data |
| How a topic is framed in video | How do short videos present heat pumps? | Video records and relevant text available through the project’s approved sources |
| How a public announcement was received | How did people react to a new transport policy? | Public response records from approved access routes on the selected networks |
| How communities organise information | How do hobby communities write their rules? | Community descriptions and rules available through an approved research route |
Be honest about what the data can say. Public posts tell you how people who post publicly on that network discuss something. They do not tell you what a population thinks.
Design a sample you can defend
- Name the collection exactly. Record the source dataset, endpoint or search and exact input fields. A phrase such as “posts about heat pumps” is not a sufficient method description.
- Fix the time scope. Distinguish publication date, collection window and any archive-update delay. Use only filters documented for the selected source.
- Choose the order deliberately. Record relevance, popularity, time ordering or another selection method. A sort name does not establish a random sample.
- Collect across networks only where needed. Analyse within each source first because units, coverage and counters can differ.
- Describe selection effects. Record inclusion rules, exclusions, pagination and the stopping rule. Do not imply that a ranked result list covers the entire network.
Document so others can repeat it
A documented method makes the collection easier to evaluate and repeat, though changed or removed sources can prevent identical later results. For every permitted collection, record:
- Network and search. For example, YouTube Search videos.
- Exact query and settings. The words, operators, filters and sort.
- Date and time collected. Search results change hour to hour.
- Number of results kept. And any you excluded, with the reason.
- Where the data is stored. File names of your exports.
Where the project’s approved access arrangement and use rules allow Superlurk, saved results include their search inputs and collection time and can be exported as CSV or JSON. Product retention does not determine what your project may store or share. A social research log provides a template that can also be used with an official research dataset.
Get the data into your analysis tools
Exports work with the usual tools: CSV opens in a spreadsheet or in R and Python; JSON keeps nested fields for programmatic analysis. For qualitative work, import the text columns into your coding software and code by theme. Our export guide covers the details, and comment analysis shows a coding approach for comment threads.
Budget from the actual collection plan and any required access arrangement. If using permitted Superlurk searches, review the quote for each search, lookup and additional page. The free trial has 10 credits; a study below 100 credits does not therefore fit the trial. No total study price is implied by this guide.
Ethics: the questions your reviewers will ask
The Association of Internet Researchers’ Internet Research: Ethical Guidelines 3.0 is a historical reference approved in 2019 and published with a 2020 citation. Use it alongside the guidance and project review your institution currently requires. Questions to resolve include:
- Public is not the same as consent. People post publicly for an audience, not usually for research. Consider expectations of privacy in each community.
- Quotes can identify people. A verbatim quote can often be found with a search engine. Decide through the project review whether to quote or paraphrase. Label any composites explicitly; do not present constructed text as a participant’s verbatim statement.
- Minimise what you keep. Collect what your question needs; avoid storing profile details you will not analyse.
- Sensitive topics need special care. Superlurk's terms forbid collecting data about children or about sensitive characteristics (health, religion, sexuality, political views) of identifiable people. Later anonymization does not make a prohibited collection acceptable. Choose a source and approval arrangement built for the proposed study.
- Platform terms and model training. Superlurk's terms also forbid redistributing results in bulk or using them to train models without the platforms' permission. Plan data sharing and any computational work accordingly.
- People can ask to be removed. Anyone in Superlurk results can request removal through the deletion request page.
A worked example: how short videos explain heat pumps
A student proposes studying how short videos explain heat pumps to homeowners. This is an illustration of study design, conditional on approved source access and institutional review; it is not a completed or approved study.
- Access: confirm whether the student and project qualify for the selected research datasets, and document the institutional review and platform approvals before collection.
- Sample: specify the exact approved query fields for heat-pump explanations and costs, the language, publication window, collection dates and inclusion rule.
- Context: inspect a predefined set of relevant comments only where the approved source provides them. Record how videos and comments were selected.
- Documentation: store the permitted records, source identifiers, query inputs and collection times under the project’s retention and access controls.
- Analysis: code only the available, reviewed text for cost, noise and savings claims, and keep questions separate from assertions. Do not infer spoken content from metadata alone.
- Dissemination: apply the approved quotation, identifiability and sharing plan. Distinguish source observations from the student’s interpretation.
Determine cost only after selecting the approved access route and reviewing its actual collection requirements. State the limits of the design: the chosen source datasets, query fields, one collection window, language scope and the sampling rule. Do not claim a representative account of all homeowners.
Frequently asked questions
Can I use social media posts in academic research?
It depends on the approved access route, project purpose, platform requirements and institutional review. Public visibility alone does not establish permission for academic collection. Reddit and TikTok document specific research access programs.
How do I collect social media data for a thesis?
First confirm that the source and access arrangement permit the proposed study. Then define the sample, exact inputs, collection window, inclusion rules and handling plan. Record enough detail to explain the collection, while acknowledging that changing sources may prevent identical later results.
Is social media search data a random sample?
No. Search results are selected and ranked by the source or search service. Describe the sample and collection method rather than treating it as representative of everyone on a network.
Should I quote posts word for word?
Decide this through the project’s review and dissemination plan. Searchable quotations can make an author identifiable. Clearly distinguish verbatim quotations, paraphrases and any explicitly labelled composites.
Can I research sensitive topics such as health or politics through Superlurk?
Superlurk prohibits collecting data about sensitive characteristics of identifiable people. Do not treat later anonymization or a general institutional approval as permission to make a prohibited collection.
Sources
How this guide was written. Updated on 11 November 2026 against current Reddit and TikTok access documentation, Superlurk’s acceptable use policy and the historical AoIR Ethical Guidelines 3.0 reference. The example is a study-design illustration, not an approved project, price quote or collection result. An institution’s ethics review and the relevant platform’s data access approval are separate requirements.