- Created by: sophiehartley
- Created on: 13-06-18 13:53
Motivating factors and bias: Charlton et al interviewd 13 fingerprint analysts and found several motivations for their job including job satisfaction and solving high profile cases. Examiners are emotionally driven and motivated to achieve results.
- Expectation bias - expectation of what will be found is actually found e.g. downgrades conflicting evidence
- Confirmation bias - look for confirming evidence rather than conflicting
- Anchoring effects - relying too heavily on an initial piece of info
- Contextual bias - other information aide rom that being considered influences decisions.
- Role effects - identify as part of the prosecution or defence teams, can cause subconscious bias.
Dror et al studied contextual bias. Used 27 uni volunteers. Manipulated emotional state by providing background info. 96 prints, half had ambiguous, other half had clear. Two states: low context (bike theft/buglarly), high context (murder/personal attack). Results show that in the clear condition matches were unaffected but in ambigious condition 58% matched in high, 49% matched in low. suggests decisions were swayed by emotional context information.
Key Research: Hall & Player
- Aim: to investigate whether context effects fingerprint identification by experts.
- Experiment: independent measures design, IV = low context (forgergy) high context (murder). DV = if they read report, if they found a match or no match, if they would provide evidence at court.
- Sample: volunteers byt all fingerprint experts working for Met. Police fingerprint bureau (70) experience ranging from 3m-30y. 58 active practitioners and 12 managerial roles.
- Materials: volunteer's fingerprint scanned onto the computer, super-imposed onto £50 note - made it a poor print so it was representative. Allowed to use normal equipment to identify. Controls = same fingerprint, same set for comparisons.
- Procedure: 35 got forgery, 35 got murder. No time limit. All anonymous. Treat it like a normal work day. Feedback sheet asking whether it affected analysis or not. Asked to state match, no mathc, not enough detail, insufficient detail.
- Results: 57/70 read the report. 52% who read high context felt they were affected, compared to only 6% of low context. Whilst they said it affected them, it made no difference on their decision. Only 17% in HC and 20% in LC would present it in court.
- Conclusion: context does effect them, but not their final decision. Details of crime may be surplus to requirements, as they can remain objective and detached.
Strategies to reduce bias
Independent analysis of the latent mark and comparison mark = combatting expectation bias
- Latent mark (from crime scene) should be analysed independently before analysing the comparison mark. This ensures the analysis is unbiased because they won't look for particular features.
- Based on evidence from Dror, when they found fingerprint examiners differed in reliability because they saw fewer key elements if studied comparison first. Dror suggests that the latent print should be analysed in isolation.
- Kassin et al support this, saying examiners should work in a linear way, working firstly from crime scene evidence to comparison to target.
Working in isolation from other evidence and conclusions = combatting contextual bias
- The examiner should be unaware of any of the crime scene evidence to prevent that creating any context bias which may led to misidentifying a print. Furthermore, the verifier should also be blind to the crime details and to the conclusion of the examiner, which will also prevent confirmation bias.
- Evidence from Dror et al would support this suggestion, as they found that more matches were found with more emotional crime detail, therefore if no crime detail is given, it will prevent this potential context bias from occurring.