Apps

🧪 This platform is in early beta. Features may change and you might encounter bugs. We appreciate your patience!

Susceptibility Bias

Also Known As: Confounding by Indication Channeling Bias Allocation Bias
Statistical Error ID: susceptibility_bias

Definition

Susceptibility bias occurs when the groups being compared in a study have different baseline risks for the outcome of interest. This often happens in observational studies where treatment decisions are influenced by disease severity or patient characteristics. The resulting outcome differences may reflect the pre-existing risk profile rather than the treatment effect.

Examples

An observational study finds that patients receiving a new cancer drug have higher mortality than those on the standard treatment. However, the new drug was disproportionately prescribed to patients with more advanced disease, making it appear less effective than it actually is.

A study finds that patients who receive annual full-body MRI scans have higher rates of cancer diagnosis than those who don't, leading a journalist to claim MRI scans cause cancer. In reality, people who seek out expensive preventive scans tend to have stronger family histories of cancer or prior health concerns, making them inherently more likely to have cancer detected.

A health insurer's data shows that members enrolled in a disease management program for diabetes have more hospitalizations than non-enrolled diabetics. The insurer concludes the program is ineffective, ignoring that enrollment was targeted at the highest-risk, most poorly controlled patients who were already on a trajectory toward hospitalization.

Verification Steps
Verification Steps
Binary yes/no questions that an AI must answer to detect a reasoning pattern in a text.
Each of the 452 aspects has verification steps — simple yes/no questions designed to systematically detect whether a pattern appears in a text. For ad hominem: "Does the argument attack a person rather than their claim?" For false dichotomy: "Are only two options presented when more exist?" This ensures consistent, reproducible analysis.

Binary (yes/no) questions an LLM must answer to identify this aspect:

  1. 1

    Do the treatment and control groups differ in their baseline risk for the outcome?

    Type: binary
  2. 2

    Was treatment assignment influenced by patient prognosis or disease severity?

    Type: binary
  3. 3

    Could physicians have selected treatments based on perceived patient vulnerability?

    Type: binary
  4. 4

    Does the analysis fail to adjust for baseline risk differences between groups?

    Type: binary
Deep Dive
The expandable detail section on each aspect page with examples, psychology, and counter-strategies.
The Deep Dive section provides in-depth information about each aspect: a real-world example showing the pattern in action, an explanation of why it works psychologically, practical advice on how to counter it, alternative names, and links to related aspects.

Hierarchical Context