Skip to main content

How to Use Data Sampling in Rayyan

  • Updated

Data Sampling lets you create named subsets of your references — called samples — to divide screening work among your team, reduce selection bias through randomization, and improve Rayyan’s Relevance Ranking predictions. Once created, samples appear as a filterable facet in the Filters panel so each team member can work on their assigned subset.

Note Data Sampling is available on all plans. Free and Essential plans are limited to 1 sample per review. Advanced and higher plans offer unlimited sample creation.

How to create a sample

  1. Go to the Review data page or the Screening stage and click Samples in the toolbar
  2. Under Choose sample Data Source, select All References or Filtered (to sample from a pre-filtered subset)
  3. Choose a sampling method (see table below)
  4. Enter a Sample Name and a Percentage or Number of Articles — these two values are linked and update automatically when you change one
  5. Click + Add more samples to create multiple samples in the same session (e.g., one per team member)
  6. Click Create
Note Total percentages across all samples in a single session must be 100% or less. You can create samples covering less than 100% — remaining references simply won’t be assigned to any sample.
Important Create all your samples in a single session to avoid overlap. If you create a sample, then return later to create another, the new sample may include references already assigned to the first. Creating all samples at once ensures clean, non-overlapping division of work.

Sampling methods

Method How it works Best for
Randomized References selected randomly Reducing selection bias — recommended for most workflows
System ID Selected in order of internal Rayyan ID Reproducible, sequential division
Title Ordered alphabetically by title Alphabetical division
Publication Date Ordered by publication date Dividing work by time period
Rating Ordered by Relevance Ranking score Prioritizing high-relevance references for specific reviewers

How to use a sample during screening

Go to the Screening stage, open the Filters panel, find the Sampled Data facet, click Enable filter, and select your assigned sample name. The reference list updates to show only those references. Samples do not restrict access — any team member can view any sample. The review owner is responsible for communicating which sample each member should work on.

Best practice Use Randomized sampling when dividing work among screeners — it ensures each person receives a representative cross-section of the full dataset. When different team members screen different randomized samples, Rayyan’s Relevance Ranking learns from a broader set of decisions, improving predictions for the remaining unscreened references.

How to manage samples

Samples cannot be edited after creation — you can only rename them. To rename a sample, locate it in the Sampled Data facet, double-click the name, enter the new name, and press Enter to apply.

To delete one or more samples, select them in the Sampled Data facet, click the facet settings icon (···), and click Delete selected. Deleting a sample removes the grouping only — no reference data is deleted.

Continue to: How to Use Ratings and Relevance Ranking in Rayyan

Related articles

Comments

0 comments

Article is closed for comments.