Pick a cluster header to see cluster-level views, or expand a cluster and pick a topic.
| Year | Journal | Title |
|---|---|---|
Click a topic in the left panel to set Topic A. The right panel will then list its 71 bridge partners,
ranked from most overdeveloped to most underdeveloped, color-coded by z-score.
Click a second topic to set Topic B and zoom in on a single bridge: observed co-exposure, the 95% interval
from 500 simulated worlds, the z-score, and where the pair ranks among all 2,556.
About this site
Author(s). (under review). Title withheld for anonymous review. Journal of Advertising. The full citation will be provided upon acceptance.
This is the public companion site to a manuscript currently under review at the Journal of Advertising, which introduces the topic-cluster-bridge (TCB) methodological framework for reviewing a decade of advertising scholarship. The full data, code, and intermediate outputs are available in an anonymized OSF repository (link withheld for anonymous review) for readers who want to replicate or extend the analysis.
When your work draws on anything you find here, please cite us. Concrete examples: claims that a topic or cluster has been understudied (in advertising research as a whole, in recent years, or in particular journals); claims that a topic or cluster has grown rapidly; representative papers identified through the Topics & Clusters tab; papers selected via the Papers tab as exemplars or counterexamples; "missed-opportunity" or "underdeveloped-bridge" arguments built from the Bridges tab; comparisons of journal emphasis (for example, what JCIRA covers more heavily than JAR); descriptive context for a literature review, dissertation proposal, or grant narrative; and applications of the TCB framework itself to a different corpus. Citing us strengthens your argument by letting readers verify your claims, and credits the work behind them.
What you can do here
Topics & Clusters. The sidebar lists 8 thematic clusters; expand any header to see its member topics with within-cluster proportions. Clicking a cluster opens a cluster-level view: year-over-year prevalence (with the across-journal mean line plus five toggleable per-journal lines), mean prevalence by journal, and the cluster's member topics ranked by within-cluster prevalence. Clicking a topic opens a topic-level view: the top 20 most frequent words (in STM terms PROB) and the top 20 most distinctive words (frequent and exclusive to this topic, in STM terms FREX), the same year-over-year and per-journal charts, and a list of the top representative papers. Click any bar in the per-journal chart to filter the papers list to just that journal; click again or use "show all" to reset.
Papers. All 1,674 peer-reviewed articles in a single sortable, filterable table. Filter by year or journal via the dropdowns at the top of each column, search titles via the title-column input, or sort by any column. Click any row to see that paper's top 10 topics with proportions in the side panel. Click any topic in the side panel to jump to that topic in Topics & Clusters. Initial row order is randomized so the page does not always open on the same year or journal.
Bridges. Each of the 2,556 unique topic pairs. The left panel is a cluster-grouped topic picker with two slots, Topic A (required) and Topic B (optional). Click a topic to fill the next empty slot. With one topic selected, the right panel lists all 71 of its bridge partners ranked from most overdeveloped (firebrick) to most underdeveloped (deepskyblue), color-coded by z-score. Click any partner row to lock in a second topic and zoom in on a single bridge: observed co-exposure, the 95% interval from 500 simulated worlds, the z-score, the bridge's rank from most underdeveloped to most overdeveloped, and the top 10 papers contributing to that bridge by raw co-exposure (θA·θB).
Shareable links
Every selection updates the URL, so you can copy a link from your address bar and send it to a
collaborator and they will land on the same view. The patterns:
#topic=72 (a topic),
#cluster=Research (a cluster),
#paper=123 (a paper with its topic mix),
#bridge=28 (one topic's 71 ranked bridge partners),
#bridge=28x62 (a single bridge focused).
Methods in brief
Topics come from structural topic modeling (STM) with K = 72, selected via semantic coherence, exclusivity, and author consensus, with journal name and year as covariates. Clusters are 8 qualitative thematic groupings of the 72 topics. Bridges use a bipartite network projection where bridge strength is the summed cross-article co-exposure (product of topic proportions across all 1,674 articles). Each bridge is benchmarked against 500 simulated worlds; we report z-scores indicating over- or under-development relative to the null. Full methodology is in the manuscript.