Notes on Mediation Model and Confirmation Bias in Lineups

Mediation Model: Used to analyze correlations.
People in single-blind administrations are more likely to choose the suspect.
This is due to the pressure exerted by the administrator on the witness.
The relationship between knowing the suspect and choosing the suspect disappears when accounting for the pressure applied.

A study was conducted in Ohio to investigate confirmation bias in investigations.
Administrators were participants who believed they were in a regular eyewitness study.
Mock witnesses (confederates) either identified the suspect or a filler with varying confidence levels.
The study examined how investigators would use this information when they knew who the suspect was.

When administrators were blind to the suspect's identity:
- Approximately 84-86% of IDs were recorded, regardless of whether the witness identified a filler or the suspect.
When administrators were not blind:
- If the witness identified a filler (wrong choice) with low confidence, administrators often did not record the answer.
- If the witness identified the suspect (right choice) with low confidence, administrators were more likely to record the answer.

Non-blind administrators have the freedom to interpret the witness's identification.
If a witness is unsure and makes a mistake, the administrator might:
- Discourage the witness or not record the identification.
If a witness tentatively identifies the suspect:
- The administrator might affirm the choice and record it.

Administrators may unconsciously:
- Accept evidence that supports their hypothesis (who the suspect is).
- Reject evidence that contradicts their hypothesis.
This can subtly influence the outcome of the lineup.

Confirmation bias can affect how evidence is interpreted and recorded.
Non-blind administrators may unintentionally influence witnesses or manipulate the process to confirm their expectations.