Skip to main content

​
Overview

Alias employs advanced text similarity algorithms to detect duplicate and near-duplicate responses. This helps identify participants who may be submitting copy-pasted or templated responses to multiple questions or across different surveys. There are two types of duplicate checks performed by Alias:
  1. Self-duplicate
  2. Cross-duplicate

​
Self-duplicate

Self-duplicate detection identifies cases where a participant submits similar or identical responses to multiple questions within the same survey. Example:
  • Question 1: “Describe your typical day.”
  • Response 1: “Yoga, healthy eating, focused work, regular breaks, evening reading.”
  • Question 2: “What does your daily routine entail?”
  • Response 2: “Healthy eating, focused work, regular breaks, and evening reading.”
In this case, the responses to Questions 1 and 2 are nearly identical, triggering a “Self-duplicate response” flag. Self-duplicate checks have no minimum length requirement, so even short repeated responses can be flagged.

​
Cross-duplicate

Cross-duplicate detection identifies cases where multiple participants submit similar or identical responses to the same question across different submissions. Example:
  • Question: “What drives you professionally?”
  • Participant 1 Response: “Challenge, learning, innovation, making an impact.”
  • Participant 2 Response: “Challenge + learning + innovation + having an impact.”
Here, the responses from Participants 1 and 2 are nearly identical, triggering a “Cross-duplicate response” flag. To avoid flagging common short responses (e.g., “Yes”, “No”, “Maybe”), cross-duplicate checks only apply to responses with 20 characters or more.

​
Response Grouping

In addition to flagging duplicate responses, Alias clusters similar responses into groups. The response_groups object in the API response assigns an integer to each question, representing the cluster its response belongs to. For example: 
"response_groups": {
  "Q1": 1,
  "Q2": 2,
  "Q3": 1,
  "Q4": 3
}
This indicates that responses to Q1 and Q3 are similar and belong to group 1, while Q2 and Q4 belong to separate groups. Response grouping provides an easy way to identify and review clusters of duplicate or near-duplicate content across your survey data.

​
Interpreting Duplicate Detection Results

When a duplicate is detected, the corresponding question’s response is flagged in the checks object of the API response. For example: 
"checks": {
  "Q1": ["Cross-duplicate response"],
  "Q3": ["Self-duplicate response"]
}
Use these flags to identify and investigate potential cases of copy-pasted, templated, or bot-generated responses.

​
Next Steps

  • Learn how behavioral tracking can help detect fraudulent activity.
  • Understand how effort scoring provides a more nuanced assessment of response quality.
  • Explore the API reference for details on the checks and response_groups objects.