Communication PhD Explorer

About this explorer

A free tool for people applying to US doctoral programs in Communication, Media Studies and Journalism. It is modeled on the Political Science Program Explorer and rebuilt for Communication with program-level results, explainable matching, a shortlist and a bilingual interface. Built by Junyu Jiang.

What is in the data

72 universities and 2370 tenure-line faculty (assistant, associate and full professors) from the units that grant a research PhD in communication, media or journalism. Lecturers, teaching and clinical faculty, professors of practice, emeriti, visiting and affiliated faculty are not included, because the tool is about who can supervise a dissertation. Data was read from official department directories on 2026-10-22.

How the programs were chosen

Candidates came from the NCA Doctoral Program Guide and from the US universities in the ShanghaiRanking 2025 Communication subject list, which was used only to find candidates. Each candidate was then checked against its official pages for a research doctorate; universities that only offer master’s degrees were dropped. The site shows no rankings: they say little about whether a department fits your research.

What is published about each person

A one-sentence research summary written for this site, the research areas the department lists, the research labels below, and links to the official department profile and, where the department links them, a personal site or Google Scholar. No photos are stored or shown, and profile text from department pages is not republished: it is read once, at build time, to compute the matching vectors, and only those vectors are published.

Research labels

Every faculty member carries labels in five dimensions, assigned by a language model from the public profile text (department profile, personal site when linked, listed research areas). A ★ marks a primary emphasis. Labels are interpretive aids: they describe documented emphases, not the whole of anyone’s work, and they can be wrong. Each record states its sources and a confidence level.

Research topics — The communication questions the research is about.
Political communication, public opinion, campaigns, deliberationJournalism studies, news production, news audiences, press freedomMisinformation, disinformation, fact-checking, trust in media and institutionsDigital platforms, algorithms, AI, data and society, platform governanceMedia effects, media psychology, attention, emotion, persuasion processesHealth communication, risk and crisis communication, health campaignsScience and environmental communication, climate, public understanding of scienceInterpersonal, relational, family, and small-group communicationOrganizational communication, work, leadership, institutionsStrategic communication, public relations, advertising, marketing communicationCritical and cultural studies, identity, race, gender, sexuality, powerRhetoric, public address, argumentation, discourseMedia industries, media economics, policy, law, regulationGlobal, international, and intercultural communication; migration and diasporaEntertainment media, games, fandom, popular culture, sports mediaCommunication history, theory, philosophy of communicationHuman-computer interaction, computer-mediated communication, social robots, VR/ARLanguage and social interaction, conversation analysis, nonverbal communicationChildren, youth, families, and media; media literacy; education
Research approaches — How the research is done.
Experiments and causal inference (lab, online, field, survey experiments)Survey research, panels, measurement and scale developmentComputational methods: text-as-data, NLP, machine learning, large-scale digital trace dataNetwork analysis and diffusionContent and framing analysis (manual or automated)Qualitative methods: interviews, ethnography, focus groups, case studiesCritical, interpretive, rhetorical, and textual analysisHistorical and archival researchPsychophysiological, neuroscientific, eye-tracking, and biometric measuresMixed methods and multi-method designsDesign research, system building, and tool developmentTheory building and conceptual work
Data and evidence — What evidence they analyze.
Social media and platform data (posts, comments, trace data, APIs)News, broadcast, film, advertising, and other media contentSurvey and experimental dataInterviews, observations, and fieldworkArchives, historical documents, speeches, and textsAdministrative, audience-measurement, and large-scale behavioral dataPhysiological, eye-tracking, and device log data
Settings and actors — The arenas, actors, or populations studied.
Platforms, algorithms, and online communitiesNewsrooms, media organizations, and creatorsElections, parties, governments, and policy arenasHealth systems, patients, and public health agenciesOrganizations, workplaces, and nonprofitsFamilies, couples, friendships, and peer groupsAudiences, publics, and citizensChildren, students, and educational settingsActivists, social movements, and civil societyMarginalized, minoritized, and diasporic communities
Regional focus — Where the research is situated.
United StatesChina (incl. Hong Kong, Macau)East Asia other than China (Japan, Korea, Taiwan)South and Southeast AsiaMiddle East and North AfricaSub-Saharan AfricaLatin America and the CaribbeanEurope (incl. Russia and the post-Soviet region)Cross-national, comparative, or global

How matching works

  1. Semantic score. Each faculty member is represented by the average of two 384-dimensional bge-small-en-v1.5 vectors: one for the one-sentence research summary and keywords, one for the profile text. Faculty vectors are precomputed from the public profile and stored as 8-bit integers (the profile text itself is not published); your text is embedded in your browser with Transformers.js (a one-time download of about 34 MB, cached by the browser). Raw cosine similarities from this model are compressed (almost every pair scores between 0.5 and 0.8), so they are calibrated against 32 realistic probe descriptions: a faculty member at the median of a typical query’s distribution scores 0, one in the top 0.5% scores 85, and the curve flattens above that.
  2. Keyword score. A BM25-style score over the same profile text, plus exact matches of each person’s keyword phrases, normalised to 0–100.
  3. Detected priorities. Regions, methods and topics named in your text become chips. By default a chip adds a small bonus (3 points, up to 9) to faculty who carry that label; click it once to make it a requirement, and again to switch it off.
  4. Fit = 75% semantic + 25% keyword + bonuses (weights adjustable). Program fit = 50% of the best match + 30% of the second + 20% of the third, so departments with several plausible advisors rank above departments with one star.

Nothing you type leaves your device. There are no accounts and no analytics on this site.

Limits

Corrections

Every faculty page has a “Report an error” link that opens a GitHub issue; corrections are folded into the next build. Faculty who would like their information changed or removed can open an issue or email junyujiang0508@gmail.com, and the record will be updated or taken down. The code and the data pipeline are public at github.com/Junyu-Jiang0508/communication_phd.

Universities included