More Related Content
Similar to Shl dragos iliescu biz (2013 11-13)
Similar to Shl dragos iliescu biz (2013 11-13) (20)
More from RevistaBiz (20)
Shl dragos iliescu biz (2013 11-13)
- 2. Making HR measurement strategic
• It does not require claims, but actions
• Requires a melange of Financial and Technical reasoning and
expertise
◦ Financial: because any strategic metric finally should be resolved
to a cost-benefit analysis
◦ Technical: because measurement is impossible without
understanding the construct measured
2
© SHL 2012
November 19, 2013
- 3. Established HR measurement models on ...
• e.g. Cascio & Boudreau (2008)
• We can now measure very well Cost & Gain for:
◦ Absenteeism
◦ Employee separations
◦ Employee Health & Welfare
◦ Employee Attitudes, e.g. Engagement
◦ Talent Investment
◦ Development programs
◦ Selection & Staffing
3
© SHL 2012
November 19, 2013
- 4. An example: Efficiency of personnel selection
• What methods can and will be efficient?
• When are tests useful?
• Can I assess how many misdecisions (mishires) I will do with a
specific selection array?
• Can I improve on this? When is it worth it?
4
© SHL 2012
November 19, 2013
- 5. Terminology
• Base Rate (success rate)
◦ The proportion of current job incumbents recruited without the use
of tests who perform well in the position
• Selection Ratio
◦ Proportion of applicants who are selected
• Test Validity
◦ Relationship between test scores and job performance
5
© SHL 2012
November 19, 2013
- 6. Terminology clarified in an example
• 100 applicants for 20 customer service positions
◦ (Selection ratio is 0.20)
• 60 out of these 100 applicants (60%) are likely to perform well
◦ (Base rate of 0.60)
• If applicants were selected at random we would expect:
◦ 12 are likely to be successful in the job (60% of 20 selected)
◦ 8 are likely to be unsuccessful (40% of 20 selected)
6
© SHL 2012
November 19, 2013
- 7. The Taylor Russell table
Estimated ratio of applicants hired who can be expected to be successful on the job.
This table is based on a current random selection success rate of 60%
LOW
HIGH
SELECTION RATIO
.40
.60
.05
.20
.5
.6
.60
.68
.75
.82
.88
.93
.96
.60
.65
.71
.76
.81
.86
.90
.60
.64
.67
.71
.75
.79
.83
.7
.8
.9
1.00
.99
1.00
1.00
1.00
.94
.98
1.00
1.00
.87
.92
.97
1.00
VALIDITY
.0
.1
.2
.3
.4
7
.80
.95
.60
.63
.65
.68
.70
.73
.76
.60
.61
.66
.64
.66
.67
.69
.60
.60
.61
.61
.62
.62
.63
.80
.83
.88
1.00
.71
.72
.74
.75
.63
.63
.63
.63
© SHL 2012
November 19, 2013
- 8. Quality of selection measures (Robertson and Smith, 2001)
+1
PERFECT PREDICTION
.63
Ability and Structured Interview
.60
Ability and Work sample
.54
Work Sample Tests
.51
Structured Interviews
.51
Ability Tests
.41
Job knowledge Tests
.40
Personality Tests
.35
Biodata
.26
References
.18
Years Job Experience
.10
Years Education
.02
Graphology
0
-0.01
RANDOM PREDICTION
Age
8
© SHL 2012
November 19, 2013
- 9. How can I improve?
•
Validity
HIGH
•
Current success rate
LOW
•
Selection ratio
LOW
9
© SHL 2012
November 19, 2013
- 10. When can I not improve (or it’s not worth it)
• There are very small numbers of applicants
• Current success rates are high
• Individual differences in job performance are small
• Job success is not critical, failures can be tolerated
• Anybody can prove successful
10
© SHL 2012
November 19, 2013