Choose an order for the question you have
Open Research, choose a topic or search term, and select Sort by. Year · newest first is the default; Year · oldest first helps trace a line of work from its earlier entries. Citations · most first can help find established discussion in the collected corpus. Title · A–Z is useful when you know part of a title and want a stable alphabetical scan. Sorting applies to all matching, grouped results before pagination, and CSV export follows that same order.
The Publication year filter narrows results to the earliest known public year of each linked work. A preprint first made public in 2023 and linked to a 2024 journal publication remains a 2023 work here. This is a discovery convention, not a claim that the journal published it in 2023. We retain a journal-only year when a source lacks a precise date; a known year does not imply a known day.
What an indexed citation count means here
Each paper result shows a citation count when it is available, the provider, and its retrieval date. OpenAlex is preferred when a valid measurement is available for a linked work; otherwise the site uses Crossref. These providers observe different citation coverage. Their counts are not interchangeable and the combined order is an aid to discovery, not a calibrated ranking. OpenAlex documents cited_by_count as a work-level field; Crossref supplies deposited citation metadata through its REST API.
Unknown is different from zero
Citations not indexed means this snapshot has no usable measurement for the work. It does not mean nobody has cited it. A known zero remains a numeric measurement and sorts ahead of unknown counts. We never add citation counts from a preprint and its journal version. One linked work uses one provider measurement, preserving the measurement’s actual retrieval date.
Citation sorting displays how many matching results have a known count. Read that coverage line before interpreting the first page. A title without a DOI can still be valuable while lacking citation enrichment. The collector’s bounded OpenAlex budget rotates through unique DOIs and refreshes the oldest enrichment snapshots first. It does not promise that every measurement was refreshed during the latest collection. Check each row’s date.
A repeatable two-pass reading workflow
For an unfamiliar topic, start with all dates and a topic filter. Scan citations to find candidate background reading, then switch to newest first to identify recent additions. Save the URL for each view. Open a candidate’s original source and record the exact claim you intend to use, its publication version, and the experiment or theorem that supports it. Our recommendation is to keep those notes alongside the bibliographic information rather than treating the count as evidence for the claim.
For a dated research meeting, apply a publication year filter and export the results. The CSV includes publication year, indexed citations, provider, and retrieval timestamp so collaborators can understand the view. Use the browser’s Back button to return to previous sort and filter states. If a result is absent, inspect topic rules, source coverage, date precision, and identifier linking before concluding that the paper or research activity does not exist.
Keep the limits with the result
Our interpretation of these controls is deliberately task-specific: newest can mean timely, and cited can mean widely discussed, but neither says that a finding is correct, reproducible, relevant to your application, or affordable to implement. A strong literature review explains why a selected paper answers the question, not merely where it appears in an ordered list. When source coverage differs or counts are stale, retain that qualification in your exported notes.