TailorMade: Selecting Employees for Organization Fit

1
Tailor-MadeSelecting Employees for Organization
Fit

• Fritz Drasgow
• CIOP
• November 15, 2002

2
Person-Organization Fit Matters!

• Jennifer Chatman (1991) showed that
  person-organization fit is important. It
  predicted
• Job satisfaction 1 year later
• Intent to stay with organization 1 year later
• Actual turnover 2.5 years later

3
Rethinking Selection

• Traditionally, organizations have selected
  employees based on applicants KSAs and job
  requirements.
• This has led to a focus on
• Cognitive ability
• Specific knowledge and skills relevant for a
  particular job

4
Rethinking Selection

• Butthe traditional approach pays little
  attention to predicting satisfaction and tenure
• Andwith increased use of project teams, lattice
  career paths, and organizational competency
  models, selection for skills needed for a
  specific job may be less valuable

5
Selection in the 21st Century

• Focus on general characteristics of
  applicantscognitive ability continues to be
  important and conscientiousness is now recognized
  as important
• Also explicitly consider the match between the
  values of an individual and the culture of the
  organization

6
Selection in the 21st Century

• In addition to predicting task performance, new
  goals include predicting
• Job satisfaction (Judge et al. 2001
  meta-analysis r between job satisfaction and
  job performance is .30 .52 for complex jobs!)
• Tenure
• Contextual performance
• Effectiveness in teams
• Effectiveness in diverse jobs

7
Person-Organization Fit

• Selecting for P-O fit facilitates achieving these
  new goals
• But this is not a panaceathere are a variety of
  difficulties and challenges in P-O fit
  selectionand we will talk about many of them
  today

8
Definitions

• Person-Job (P-J) fit the match between a
  persons knowledge, skills, and abilities and the
  requirements of a specific job (demands-ability
  fit).
• Person-Organization (P-O) fit the congruence of
  an individuals personality, beliefs, and values
  with the culture, norms, and values of an
  organization

9
Definitions

• Person-Vocation (P-V) fit the match between an
  individuals vocational interests and the
  activities of an occupationHolland RIASEC codes
• Person-Group (P-G) fit the compatibility
  between individuals and their work groups
  (Kristof, 1996, p. 7)

10
Definitions

• Needs-Supplies fit the match between an
  employees needs what he/she receives from the
  organization (pay, benefits, interesting work,
  etc.)

11
A Conceptual Framework The People Make the
Place

• Ben Schneiders Attraction-Selection-Attrition
  (ASA) model
• An organizations founder defines its goals and
  establishes policies and practices to achieve the
  goals.
• Together, the goals, policies, and practices
  create the structures, processes, and culture
  that define an organization.

12
ASA Model

• Attraction An individual is attracted to an
  organization when his/her preferences,
  personality, and values are perceived as
  congruent with the structure, processes, culture,
  and values of an organization
• Selection An organization recruits and hires
  individuals who can contribute to its
  effectiveness organizations choose people with
  varying competencies but similar values and
  personalities

13
ASA Model

• Attrition People leave organizations when they
  dont fit in
• Sohow important are the people?
• The attributes of people, not the nature of the
  external environment, or organizational
  technology, or organizational structures, are the
  fundamental determinants of organizational
  behavior Ben Schneider (1987, p. 437)

14
ASA Model

• The combined effects of attraction, selection,
  and attrition is to produce an organization where
  the people are relatively homogeneous
• Kurt Lewin said behavior is a function of a
  person and an environment, B f(P,E)
• Schneider the environment is a function of
  people behaving, E f(P,B)
• The people make the place!

15
P-O Fit

• It is thus the values and personalities of the
  people in an organization that should be examined
  when characterizing the O part of P-O fit.
• Chatman (1991), for example, found substantial
  differences in values across seemingly similar
  organizations (West Coast offices of eight large
  U.S. public accounting firms)

16
More on Types of Fit

• See Amy Kristofs Integrative Review (1996)
• Complementary fit A person adds something that
  is missing in the organization
• Supplementary fit A person has characteristics
  that are similar to those already present in the
  organization
• E.g., select for complementary fit for knowledge
  and skills but supplementary fit for values and
  personality

17
Types of Fit

• Perceived fit A measure of fit determined from
  the responses of a single person
• via direct ratings How well do you fit in
  here? or
• via a calculation Ask respondent both How
  important is it to you to confront conflict
  directly? and To what extent is conflict
  confronted directly in this org.?

18
More on Types of Fit

• Objective fit A measure of fit computed from
  the responses of an individual AND responses of
  one or more people from the organization
• e.g., the ratings of To what extent is conflict
  confronted directly in this org.? are made by
  other people from the organization.

19
Subjective Fit

• Caution its not clear that perceived fit is
  anything other than job satisfaction
• I like my coworkers
• I respect my supervisor
• I fit in here

20
Discriminant Validity of Fit Perceptions

• Cable DeRue, JAP, 2002, looked at the
  discriminant validity of subjective fits
• P-O fitjudgments of congruence between an
  employees personal values and an organizations
  culture
• P-J fit judgments of congruence between an
  employees skills and the demands of a job
• Needs-supplies fit judgments of the congruence
  between employees needs and they rewards they
  receive (pay, benefits, training)

21
Cable DeRue

• Predicted and found that Subj. P-O fit would be
  related to organizational outcome variables
• Organizational identification (r .48)
• Perceived organizational support (r .53)
• Peer-rated citizenship behaviors (r .22)
• Turnover (r .17)
• Subj. P-J and N-S fit were less related to these
  outcomes

22
Cable DeRue

• Predicted and found that Subj. N-S fit would be
  related to job-focused attitudes
• Job satisfaction (r .61)
• Career satisfaction (r .38)
• Occupational commitment (r .43)
• Subj. P-O and P-J fit were less related to these
  outcomes

23
Cable DeRue

• Predicted But Did Not Find Subj. P-J fit to
  be most related to
• Occupational commitment (r .24, beta .01)
• Peer-rated job performance (r .00)
• Pay raise in past year (r .13)
• Subj. P-O and N-S fit were more related to
  these outcomes

24
Flies in the Ointment Computing Fit Statistics

• OReilly, Chatman, Caldwell (1991, p. 490)
  argue that a fit statistic should focus on the
  salience and configuration of variables within a
  person rather than the relative standing of
  persons across each variable.
• idiographic measurement
• but also allow comparisons across situations

25
Fit Statistics

• OReilly et al. (1991) also argued that people
  and situations should be described in
  commensurate and relevant dimensions (p. 490).
• I.e., avoid describing people with one set of
  characteristics and situations with a totally
  different set of characteristics.

26
Organizational Culture Profile

• OReilly, Chatman, Caldwell (1991) developed
  the OCPit assesses 7 dimensions of
  organizational culture
• Innovation
• Stability
• Orientation toward people (fair, supportive)
• Orientation toward outcomes (results oriented,
  achievement oriented)

27
Organizational Culture Profile

• Easygoing vs. aggressive (calm, reflective, low
  level of conflict)
• Attention to detail (precise, analytical)
• Team orientation
• See Chatman Jehn, 1994

28
Organizational Culture Profile (OCP)

• Consists of 54 value statements
• Use a Q-sort (Block, 1978) methodology for rating
• Item-category pattern is 2-4-6-9-12-9-6-4-2

29
OCP Individual Preferences

• The 54 items are ranked by each individual
• Directions How Important is it for this
  characteristic to be a part of the organization
  you work for?
• Most desirable Most undesirable
• One-year test-retest reliability .73

30
OCP Organizational Values

• Chatman (1991) had an average of 16 accountants
  per firm make ratings
• Directions
• Important values may be expressed in the form of
  norms or shared expectations about whats
  important, how to behave or what attitudes are
  appropriate.
• Please sort the 54 values into a row of nine
  categories,

31
OCP Organizational Values

• placing at one end of the row those cards that
  you consider to be the most characteristic
  aspects of the culture of your organization, and
  at the other end those cards that you believe to
  be the least characteristic.

32
OCP P-O Fit Score

• OCP preference profile the ranks of the 54
  values
• Compute the average across 16 respondents for the
  O profile
• P-O fit was computed as the correlation of a
  persons ranks with the average profile of
  his/her firm

33
OCP P-O Fit Score

• OCP profiles are ipsativeevery persons
  profile sums to 270, the mean rank is 270/54 5.
• Ipsative variables have quirky statistical
  characteristics, e.g., if we computed the 54 x 54
  correlation matrix of values, it would be
  singular (one eigenvalue 0) regardless of your
  sample size.

34
OCP P-O Fit Score

• Another problem with ipsative measures is that
  their constant mean across people.
• E.g.,for one person, many of the OCP values may
  be very important for another person, the values
  may be largely irrelevant
• but both would have means of 5

35
Profile Similarity Indices

• Jeff Edwards (1993 2002) has slammed a variety
  of PSIs, including

36
Edwardss Argument

• PSIs
• Are conceptually ambiguousby combining
  heterogeneous elements into one score, it is
  impossible to know the construct underlying the
  PSI
• Discard possibly important information assume
  positive and negative differences have equal
  effectsand absolute level of profile elements is
  irrelevant

37
Edwardss Argument

• Possibly the most telling argument is that PSIs
  assume a highly restrictive relationship.
• E.g., suppose you use D1 X1 Y1

38
Edwardss Argument

• Then
• So that X and Y have equal, but opposite
  regression coefficientsthis seems very unlikely
  to occur

39
Edwardss Argument

• Instead, Edwards recommends polynomial
  regression,
• Here X and Y are allowed to have any regression
  coefficient
• Note you can test the increase in R2 obtained
  by going from the difference score to the more
  general equation

40
Comments on Polynomial Regression

• Obviously, it would be impossible to enter the
  OCPs 54 values into one regression equation
• So, compute scale scores for the OCPs seven
  factors

41
Comments on Polynomial Regression

• Maybe the bottom line is variance explained
• When predicting job satisfaction, Bretz and Judge
  (1994) found
• D1 explained 12 incremental variance beyond
  controls (adj. R2 .18)
• Linear and interaction terms explained 20 more
  variance (adj. R2 .38)
• Would you rather explain 18 or 38?

42
Recommendations

• Identify the scope of organizational culture you
  want to assessworkgroup, plant, division, entire
  organization
• Sample 16 people from this unit of analysis (?
  Chatman, 1991, had good results with 16)

43
Recommendations

• Form multi-item composites to assess underlying
  values (e.g., Van Vianen, 2000)
• Use polynomial regression (at least linear and
  interaction terms X1, Y1, X1Y1, X2, Y2, X2Y2,
  etc.)

44
New Directions

• Chris Robert, Arzu Wasti, and some UIUC folks are
  working on an OC measure that parallels the
  dimensions of national culture
• Individualismpeople see themselves as separate
  and unique from groups of people in their society
• Collectivismones identity is defined by ones
  membership in groups

45
New Directions

• and Power Distance, which is the degree to which
  status inequalities are accepted as necessary and
  proper
• Verticallegitimate power is given to those at
  the top, and people below follow orders
• Horizontalpeople are equal

46
Harry Triandis

• Butrather than two separate dimensions, Triandis
  sees four cultural syndromes
• Vertical collectivist
• Horizontal collectivist
• Vertical individualist
• Horizontal individualist

47
A New Approach

• Robert et al. suggest assessing the culture of an
  organization in a way that is parallel to
  TriandisI.e., 4 organizational syndromes
• Then the match between the way it is around
  here and a persons preferences provides a new
  view of P-O fit
• Research is in progress

48
Using P-O Fit in Practice

• In selectionobjective P-O fit is related to job
  satisfaction, commitment, and turnover
• Organizations already try to select for P-O
  fitbut several studies have found that
  recruiters can assess P-Recruiter fit, but not
  P-O fit
• Need an assessment tool and procedure

49
Using P-O Fit in Practice

• First, for self-selection, like a realistic job
  preview
• E.g., put a P-O fit instrument on the
  organizations web site, let possible job
  applicants complete it, give them feedback about
  fit
• Dineen, Ash, Noe, JAP, 2002, found objective
  P-O fit was related to applicants attraction to
  the organization good obj. fit ? pos. attraction

50
Using P-O Fit in Practice

• Why not tell all applicants they fit? One actual
  web site seems to do this
• Dineen et al.
• You wont achieve the goal of a leaner, better
  fitting applicant pool
• Giving false feedback to lure applicants may
  backfire when they find they dont really fit

51
Using P-O Fit in Practice

• Second, as an explicit selection tool
• Give applicants an instrument like the OCP
• Compute objective P-O fit
• Use objective fit as one consideration when
  making hiring decisions
• Companies consider fit when hiringthis might add
  empirical rigor to the process

52
In Sum,

• Objective P-O fit really does matter
• At the very least, organizations should help
  job-seekers self-select by providing a means for
  assessing P-O fit on their web sites
• Will job-seekers self-select out? Need research
  on this

53
In Sum,

• There is good reason to use an assessment tool to
  evaluate P-O fit as part of the selection process
• How to assess the O part? Chatman used 16
  people/organization, but from highly homogeneous
  businesses.
• Should we sample 16 people randomly from a
  functional area? plant? division? or nonrandomly
  sample the CEO and a few other leaders?

54
In Sum,

• Again, P-O fit matters, but exactly how
  organizations should use this fact to improve
  selection is something we still have to figure
  out