Honesty and integrity tests

1
Honesty and integrity tests

• Practical and psychometric effectiveness
• Dr Iain Coyne
• Institute of Work, Health Organisations
• University of Nottingham, UK
• Paper presented at HR Section of NIP
• Bilthoven (2006)

2
Outline

• History of integrity testing
• Types of integrity tests
• Quality issues
• scope, accuracy, fairness, acceptability
• Validity Meta-analyses
• Negatives
• Conclusions

3
History

• Been around since 1950s
• Saw an increase in use in US in 1980s
• Sensitisation as a result of scandals
• Employee Polygraph Protection Act
• Their use
• 5000 US organisations use them annually (Sackett
  Harris, 1984)
• 10-15 of all US employers (OBannon et al, 1989)
• 2.5 of UK personnel managers used integrity
  tests to select for honesty (Coyne, 2001)

4
Types
Overt
Covert
Attitudes Admissions
Personality
Reid Report PSI
PRB Hogan Reliability Giotto
5
Scope

• Initially view them as specific as they only
  measure integrity or related constructs

Specific
Broad
Overt
Covert
Theft
CWB
6
Accuracy

• Internal consistency
• 0.85 for Reid Report and PSI
• 0.73-0.97 for PRB
• 0.63 for Hogan Reliability Scale
• Test retest
• 0.91 on PSI over 1 week
• 0.90 for Hogan over 4 weeks
• Ones et al (1993) meta-analysis
• Mean alpha 0.81
• Mean test-retest 0.85

7
Fairness

• Group differences - Ones Viswesvaran (1998)
• Females 0.16 SDs higher than males
• Blacks 0.04 SDs lower than Whites
• 40 0.08 SDs higher than lt40
• Coyne (2001) 458 working adults
• Females 0.08 SDs lower than males
• Blacks 0.47 SDs higher than Whites
• 40 0.27 SDs higher than lt40

8
Fairness

• False positives
• Paradox of screening out honesty
• 50 of people may be incorrectly classified
  (Murphy, 1993)

9
Acceptability
Single method
Multi-method
Appropriate to use Do not like it Only a small
sample refuse to take it Overt more job relevant
and less offensive/invasive than covert
Ranked around the middle when compared to other
methods. Not liked but not disliked
10
Validity

• Criterion
• External criteria
• 0.23 0.62 with various measures of dishonest
  behaviour (theft, absenteeism etc)
• Time series
• Theft and termination rates dropped after
  introduction of IT
• Contrasted groups
• dishonest groups perform worse than
  non-dishonest

11
Meta-analysis Ones Viswesvaran (2001)
12
Schmidt Hunter (1998)
13
Schmidt Hunter (1998)

• GMA IT or GMA Structured interview
• Highest increase in utility
• Both can be used with entry-level or experienced
  applicants
• Both combinations are less expensive than other
  combinations
• Both measure in part conscientiousness which is
  shown to relate to job performance

14
Negatives
Supervisor ratings
Labelling
Cross-culture?
Faking
Screen out managers
Personality is better
15
Personality better?
16
Summary
17
Conclusions

• Overall research has shown favourable evidence
  for psychometric effectiveness of ITs
• Need to consider labelling, faking, false
  positive and acceptability
• Personality scales and structured interviews also
  show positive results
• Need to consider training
• What about bad barrels?