BIG FIVE continues

1
BIG FIVE continues

2
Openness to Experience

• Alternately labeled culture, intellectance,
  openness. Which best fits depends in part on
  whether factors are derived lexically
  (intelligence) or via questionnaire (openness)
• Correlated with intelligence, but it is an active
  intelligence, liking to think

3
Openness continued

• Correlated also with education (causal direction
  unclear), number of career changes, aesthetic
  interests and sensitivity, absorption (being
  fully engaged in what intellectual tasks only?
  one is doing), broad values, but the aesthetic
  and feeling components may lead to greater
  susceptibility to depression.

4
Conscientiousness

• Self-disciplined, dutiful, organized,
  responsible, reliable, hard working. Of the big
  5, most related to success across jobs and
  situations.
• Others (E, A) predict success only in particular
  situations.
  
• Cs prediction is stronger when occupational
  autonomy is high than when it is low.

5
Agreeableness

• It includes altruism, affection, humaneness,
  sincerity.
  
• Most related to pro-social tendencies.
• Most related to good parenting in mothers (High
  A, low N make best mothers.

6
Normative change

• Previous studies have shown that students become
  less authoritarian, ethnocentric, dogmatic,
  aggressive, during college, more complex in
  their outlooks, more original, unconventional,
  intelligent, psychological minded, cognitive
  commitment, absorption, open to experience,
  traits related to conscientiousness (orderliness,
  diligence, dependable, sense of control), no
  consistent changes extroversion-related traits
  (sociability, etc). Similarly for neuroticism,
  which either does not change or decreases.
  Previous work suggests that Openness,
  conscientiousness, agreeableness will increase,
  neuroticism will decrease, extroversion will not
  change. Cal Berkeley study using a Reliable
  Change Index supports the predicted change
  patterns for college students.

7
Blocks attack on the Five Factor

• Key points
• -- To be global, fully adequate, it must be
  better than competing set of constructs.
  
• --It is a model, rather than a theory. No
  functioning psychological system within the
  individual is postulated..

8
Limitations of Factor Analysis

• Factors derived because they have lots of
  commonality with other measures says nothing
  directly about their real world relevance.
  
• -- Variables that dont correlate with the mass
  of variables are often dismissed as residual,
  even though they are reliable and may be very
  important in real world contexts. Powerful
  factors can have trivial consequences when
  looking at particular concerns.
• --The mix of variables input into the factor
  analysis will determine which variables are
  large, explaining lots of common variance, and
  which ones are small.

9
Block cont.

• --Such prestructuring is common. Leads to
  replicable factors, but still says little about
  importance of the factors.
  
• --There are still no agreed-on criteria for
  determining the number of factors to be
  extracted, or for rotating factors to obtain the
  best psychological meaning of the factors.
  (e.g., is orthogonal rotation psychologically
  inappropriate and forced. How does one know
  that a factor is not a bloated complex of
  relatively independent constructs?)
• --In any data set, there are an infinite set of
  factor loadings that equally explain the data.
  Unguided factor analysis cannot choose the most
  appropriate, the one most psychologically
  meaningful.

10
Block cont.

• --Because FA starts with correlation matrices,
  all problems affecting correlations (reliability
  of variables, linearity of relations among them,
  method variance inflations, merging samples
  males and females where correlations may be
  quite different and meaningfully so among the
  subsamples impulsivity and introspectiveness
  correlate positively for male grad students,
  negatively for AF pilots, unidirectional
  relationships e.g. wittiness requires
  intelligence, but intelligence does not mandate
  wittiness.
• Correlations e.g., with agreeableness dont
  well reflect the contextuality that may be
  necessary for the relationship to occur.
  

11
Block concluded

• --People are too prone to dismiss factors that
  dont replicate, although they may be valid for
  particular samples.
  
• --Comparing multiple models of the same data must
  be done by arbitrarily identified criteria via
  subjective criteria. (Cudeck Henley, 1991)
  
• --Anthropologically speaking, we appear to have a
  bias for a relatively small number of factors,
  yielding a picture of a world which is at the
  level of complexity we intuitively prefer.
• --Therefore, the faith that the five-factor
  proponents have in this method as the sufficient
  for deciding personality structure is unwarranted.

12
Then, Costa McCrae

• 1. Sought to determine the degree to which the
  five factor model is recoverable from other
  measures such as Blocks California Q-Set (CQS)
  or Jacksons PRF. But they have not gone the
  other direction. They claim greater congruence
  than warranted. The CQS yielded 32 eigenvalues
  greater than 1 Block has identified 20 reliable
  factors. Even if just 8 factors,
  introspectiveness, narcissism, forcefulness
  emerge as three factors in the CQS not in the big
  5.
• Similar results with PRF.
• 2. Extend Big 5 to thinking about psychiatric
  disorders. But that is an impoverished sampling
  of symptoms to represent disorders, which have
  identified many components.
• 3. The revised NEO-PI-R added six dimensions for
  A and C. But when normed on 1000 subjects, N C
  correlated -.53 and E and O .40, uncorrected for
  attenuation. Block replicated the large N - C
  correlation.

13

• Costa and McCrae appear to admit that the model
  is theory-driven rather than determined by
  empirical inevitability. They also admit that if
  you keep all six facets on each dimension, you
  lose orthogonality.
• The big five has never been empirically tested
  against alternative dimensional offerings to see
  which best carves nature at its joints.
  
• So, given where we are, how good is the Big 5 for
  the scientific tasks ahead? Will it help or
  hinder the science of personality?

14
RECENT I/O APPLICATIONS OF BIG FIVE

• Bernardin, H. J., Cooke, D. K., Villanova, P.
  (2000). Conscientiousness and agreeableness as
  predictors of rating leniency. Journal of
  Applied Psychology
• Results A correlated .33 (p rating level C correlated -.33 with average
  rating level. No significant interaction. N, E,
  and O were essentially unrelated to rating
  leniency. The result remained essentially the
  same when professors ratings were used as a
  control.

15
Hurtz, G. M. Donovan, J. J. (2000).
Personality and job performance The big five
revisited. Journal of Applied Psychology

• Method
• Used only measures designed originally to
  measure the Big 5.
  
• Found 26 studies, yielding between 35 and 45
  measures for each of the five dimensions.
  
• Moderators
• Worker occupation Sales, customer service,
  manager, skilled and semiskilled.
  
• Training proficiency (training performance and
  end-of-training tests) vs. job proficiency.
  
• Performance clustered into 3 dvs task
  performance (technical performance, use of
  equipment, etc.), job dedication (effort,
  persistence, reliability), interpersonal
  facilitation (cooperation, team player, etc.)
• Computed composite score correlations (rather
  than average correlations) for lower-level
  measures to estimate higher factor correlations.
  If inter-test correlations not provided, used
  those from the HPI manual.
• Similarly for overall job performance, if
  correlations not provided, entered .55 to compute
  composite score.
  
• Corrections for artifact distributions,
  sampling error, measurement artifacts for both
  predictor and criterion variables

16

• Results
• Overall, C Stability (neuroticism) had
  significant true-score correlations, .22,.14.
  Agreeableness, at .13, was not at 90 confidence
  above zero.
• By category, C was more predictive for Sales and
  customer service for managers and skilled
  workers, it wasnt significantly above .00 at 90
  confidence interval. E and ES were significantly
  predictive for Sales, Customer service, managers,
  but not for skilled or semiskilled workers.
  Openness was significant only for customer
  service.
• C and ES were also significant for Job
  performance but not for training performance.
  
• ES was predictive for all three components, above
  90 confidence, but correlations were small.
  (table 4). C was largest across the board, but
  only above 90 confidence interpersonal
  facilitation. A was related only to the
  interpersonal facilitation dimension above 90
  confidence.
• Discussion
• Conscientiousness may add a small proportion of
  variance in predicting job performance overall,
  others for predicting specific dimensions. But
  that would depend on the degree to which they
  have been picked up and overlap with other
  selection techniques (interviews, etc.)
• Some results based on small correlations.

17
Judge, T. A. Bono, J. E. (2000). Five-factor
model of personality and transformational
leadership. Journal of Applied Psychology

• Transformational leadership (TL) (Bass, 1985)--
  inspires followers with a vision beyond their own
  self interest, rather than relying on exchanges
  (i.e., transactional leadership). Four
  dimensions
• idealized influence -- charisma
• inspirational motivation -- articulating the
  vision
  
• intellectual stimulation -- stimulating follower
  creativity by challenging their view of status quo

18

• Hypotheses
• -- Neuroticism negatively related (self
  confident, self esteem missing, but needed for
  TL)
  
• -- Extroversion positively related (related to
  sociability, charisma, social dominance, needed
  for TL)
  
• -- Openness positively related (creativity,
  originality, divergent thinking, need for change,
  willingness to transform)
  
• -- Agreeableness positively related (empathy,
  altruism, generosity, concern for others needed
  for TL).
  
• -- Conscientiousness -- no hypothesis offered.
• -- Transformational leadership will predict
  subordinate satisfaction, job satisfaction, work
  motivation, and ratings of leader effectiveness.
  
• -- TL will predict leader outcomes controlling
  for transactional leader behavior.

19

• Method
• Participants were 316 members and 223 (48 of
  whom participated by returning materials)
  graduates of community leadership programs.
  
• In first session, completed personality survey
  (NEO-PR-I). Gave surveys to supervisors and
  subordinates (80 in each case).
  
• Transformational Leadership was measured by MLQ,
  Form 5X (Basss measure). Covers the four
  dimensions
  
• Idealized influence-attributed, Displays a
  sense of power and confidence
  
• Idealized influence-behavior, Talks to us about
  his/her most important values and beliefs
  
• Inspirational motivation, Articulates a
  compelling vision for the future,
  
• Intellectual stimulation, Reexamines critical
  assumptions to question whether they are
  appropriate
  
• Individualized consideration, Spends time
  teaching and coaching me.
  
• Transactional leader behaviors also measured by
  MLQ 5X
  
• (See items, p. 756).
• Also measures of subordinate satisfaction with
  leader, job satisfaction, organizational
  commitment, work motivation, each from at least
  two subordinates.
• Leader effectiveness measured by five items from
  immediate supervisor.

20

• Transformational leader behaviors all loaded on
  one factor.
  
• E, O, A all correlated with TL, but O had
  non-significant beta when others were entered, so
  O may not contribute.
  
• A broader definition of Neuroticism, drawn from
  Judge Et al.s core self-evaluation theory,
  functioned no better than neuroticism as measured
  by the big 5.
• TL predicted all the anticipated outcomes except
  job satisfaction (Table 6).
  
• Eight control variables (gender, age, job and
  organizational tenure, public vs. private
  industry, organization size, number reporting to
  the leader) did not affect the TL - leader
  effectiveness relationship (controlling increased
  slightly from .35 to .37).
• Discussion
• The reliable multiple-R between the Big 5 and
  transformational leadership is .40. The
  strongest single correlation was .32 (with
  agreeableness). But these are stronger than
  found for job performance (.22, .30)
• The strongest correlation with agreeableness was
  surprising, but it is related to group
  performance.
  
• Surprised that Openness was not predictive once
  agreeableness and extroversion were entered, that
  Neuroticism didnt predict (even when defined
  more broadly).
• But the results support the construct of
  transformational leadership and its effects
  strongly, generalizing across different types and
  levels of organizations.