Find me someone like him

1
Find me someone like him

• The PhD experience so far

2
Structure

• Overview
• Review to date
• Discussion of current preoccupation
• Conclusion
• Questions

3
Recruitment by analogy

• Generate a framework to facilitate the
  identification of potential golden employees
  from a cohort of job applicants using a computer
  simulation.

4
Questions

• Are potentially Golden employees being
  eliminated by pragmatic shortlisting methods?
• Could a computer-based system process all
  applicants and indicate potential via analogy to
  the golden employee, considering behavioural,
  personality and other factors?
• How would such a selection system change the
  makeup of the short listed candidates?
• Answering these questions has formed the crux of
  my research

5
Issues

• Degraded grades
• Contemporary assessment
• What to measure?
• How to measure it
• How to compare measurements

6
Intellectual Framework
Effectiveness of Tertiary Qualifications
tertiary.ppt
Dimension Elicitation (rep grids)
Personality
Simulation design and scripting
Similarity MDS.ppt
Contemporary assessment tools and metrics
Culture
2004
2006
7
May 2006 An appraisal

• Does the proposal still stack up?
• Reality check against specific case

8
History

• PhD deals with graduate recruitment - Data very
  difficult to obtain
• Commercially sensitive in nature
• Consulting to several organisations building
  relationships
• Sworn to confidentiality

9
Selection 900 applications 80 appointed All had
(possibly specific) bachelors degree as
minimum All had A GPA as minimum All satisfied
professional certification criteria All satisfied
testing standards All were qualified via two
interviews
1997 Graduates (80)
10
By 2005 50 had left the firm 1 had made it to
Partner
1997 Graduates (80)
11
The One Earnings for the firm
Charge out rate _at_ 550 per hour Consistently been
80 chargeable
2005
Charge out rate 80 per hour Employees must be
70 chargeable
1997
12
The One His Earnings
Partnership allocation equates to 375 550K
2005
Graduates started on 35 38K
1997
13
The One His Potential
Major commercial centre Partnership allocation
equates to 4 7million pa
2010
He is 29 years old and may be seconded to one of
the major international branches within 5 years
2005
14
So..

• There is massive incentive for grads to perform
  at top level.
• Very few actually do
• This is, without doubt, a golden employee
• Why so few when only the cream are processed
  and appointed?

15
Process

• Standard tests are administered interviews
  (more novel methods have been proposed)
• Interviews carry most weight
• This process is typical, but evidence suggests it
  fails to appoint most appropriate candidates
• Retention rate is low
• Few are promoted to highest levels, many Partners
  are appointed from outside the firm

16
Costs drive shortlisting process

• 2005
• Grad recruitment budget 320,000
• 1750 applicants
• 110 appointed
• Shortlisting is a pragmatic response designed to
  remove most of the applications based on a
  single, often arbitrary and somewhat blinkered
  criteria.

17
Hmmm..

• What if, for detecting such golden employees,
  the HRM literature has it wrong.
• What if interviewing is the correct method
• Why is it correct?
• Why is it failing?

18
Whats happening is Rapid cognition

• Tennis coach example
• What are they looking for?
• What is a common behaviour of golden employees
  and highly successful entrepreneurs, outside the
  realm of operational attributes?
• Relationship with the firm
• Parent/child, Spousal
• Commitment to the firm beyond merely that of an
  employee

19
Relationships

• Is this ability to closely relate to an
  artificial entity the factor which is being
  looked for during the rapid cognition process of
  an interview?
• Is it possible to create a computer-based model
  such that a comparison with the golden-one can
  be made?

20
Hmmm

• More complex with the passage of time
• Research Framework needs to
• Weight personality and behavioural dimensions in
  terms of performance significance
• Utilise a valid framework for comparison in order
  to generate a meaningful analogy
• Model the rapid cognition patterns which dominate
  interviews
• How can a relationship be modelled?

21
Despair sets in

• Complex set of interacting variables which change
  over time
• Is there any research which has attempted to
  model the interactions and exchanges of human
  beings within an intimate over time?

22
Gottman

• Research led to John Gottman and colleagues.
• Gottman uses a Non-Linear Dynamics (Chaos)
  approach to model marriages could this be
  expanded to model golden/firm relationship?

23
Parallels are striking

• Marriage
• Honesty
• Loyalty
• Affection
• Compatibility
• Goal coincidence
• Reliability
• Trustworthiness
• Dependability
• Companion

• Employees
• Aptitude
• Potential
• Personality
• Loyalty
• Motivation
• Ambition
• Timeliness
• Initiative
• Communication

24
Gottmans model based on general systems theory

• dQ1/dt f1(Q1,Q2,Q3,..)
• dQ2/dt f2(Q1,Q2,Q3,..)
• Where Qs are variables, each of which indexes a
  particular unit in the system
• Equations derive the rate of change of the units
  over time

25
Mathematical Biology

• Modelling wide range of deterministic and
  stochastic systems with a set of non-linear
  differential equations has become a very
  productive area Mick Roberts

26
The method
Identify the phenomenon to model
Determine the stability of Each steady state
Identify the key elements that govern the
phenomenon
Describe the behaviour of the model near each
steady state as each parameter varies
Build a dynamic mathematical model that
encapsulates the Interaction of these elements
What does the model say qualitatively? How does
it compare with observation?
Bad Good
Find steady states
Experiment
27
The elements

• Steady states
• States where the system will remain unless
  disturbed
• Unstable or stable?
• Phase-Space
• n-dimensional graphical representation of all
  possible variable values

28
Force field portrait

• Plotting all values in Phase-Space could be
  immensely time-consuming, so
• Example
• du/dt f(u,v)
• dv/dt g(u,v)

29
Null Clines, Vector force field, Phase-Space
g0
f0
F0,gv
FF0
F0,g0
u
30
The equations of marriage

• Wt1 b rwWt IHW(H1)
• Ht1 a rHHt IWH(Wt1)
• Where
• H W attitudes of Husband Wife after tth
  interchange.
• a b base level attitude baggage
• rH rW coefficients of inertia
• IWH IHW Influence functions

31
Wifes influence function null cline
S1
S2
Husbands influence function null cline
S3
32
S1
S2
S3
33
Gottmans method

• Incorporates catastrophic events, repair and
  damping functions
• Codes 15 minutes of couples interchange using
  SPAFF coding scheme (700 couples)
• Weights codes
• Codes observational data manually

34
Success

• Gottmans model has 94 accuracy in predicting
  newlyweds subsequent divorces

35
Questions to be answered

• Can Gottmans model be expanded to explain a
  relationship with an inanimate object?
• Can a simulator engender such a relationship in a
  short period of time if so how?

36
If interviewing is correct then why is it not
working?

• Multiple personality issues mean interviewer gets
  false impression of candidates personality
• Maybe the ability to form a relationship with an
  artificial entity is best detected by another
  artificial entity - simulator

37
Conclusion

• Operational and cognitive aptitude, base
  personality and general communication skills may
  not be enough to identify a golden one
• Perhaps the missing element is a form of
  relationship ability found to a greater extent in
  golden ones
• The research proceeds with an investigation into
  such a model and simulated environment data
  generator

38
Questions ?