Work Relevant Affect Constructs

1
Work Relevant Affect Constructs

• Clearing Away Confusions

Howard Weiss Purdue University ESRC Conference on
Work Attitude Measurement October, 2007
2
Premise 1 Affect constructs are and have
been important topics in the study of
behavior (morale, satisfaction, emotion,
stress, commitment, etc.)
Premise 2 In spite of importance, confusion
rather than clarity has reigned in this area
My Goal Today To provide some organization and
clarity for this domain
3
So..
Well talk about affect and what it means,
about job satisfaction and commitment, about
emotions and moods, etc.
And try to sort out all these related constructs
Plus - measurement implications
4
Topics

• Attitudes and emotions in the context of job
  satisfaction
• Commitment
• Time frames for well being constructs
• Different criterion relationships for constructs
• Measurement

5
My Overall Message Is

• We have believed were capturing much more with
  job satisfaction than we actually have
• We have to expand our assessment and analysis of
  work experience to encompass all sorts of states
  and experiences
• We have to ask what do we want to know and why
• Do we have a practical bent, or
• Do we want to know what people think and feel and
  experience at work
• Either way were not there

6
A Brief History of Research on Affect and
WorkWeiss Brief (2001)

• Research began in the 1930s, rich with true
  affect constructs
• Hersey (1932) Daily mood reports
• Fisher Hanna (1931) Emotional maladjustment
• Hoppock (1935) Diaries (that no one returned)
• And then there was Kornhauser and Sharpe (1932)
• Questionnaire
• Facet measurement
• Single organization
• Satisfaction-performance analysis

7
A Brief History of Organizational Researchon
AffectWeiss Brief (2001)

• By the end of the decade a very narrow paradigm
  developed for studying satisfaction in U.S.
• Broad concept of affect became the narrow concept
  of job satisfaction
• The single organization survey became the method
  of choice
• Raw description replaced theory building
• Very applied focus held sway
• Kornhauser and Sharpe emerged victorious

8
Over Time Researchers Mixed Satisfaction,
Emotion/Affect and Attitude Into One Big
Conceptual Stew
Key Definitions of Job Satisfaction

• Locke (1976)
• A pleasurable or positive emotional state
  resulting from an appraisal of ones job
• Cranny, Smith and Stone (1992)
• An affective (that is emotional) reaction to
  ones job

9
But.

• Miner (1992)
• It seems desirable to treat job satisfaction
  as generally equivalent to job attitudes
• Brief (1998)
• Job satisfaction is an attitude toward ones
  job

10
So which is it?Emotion or Attitude

• For a long time and for most researchers the
  answer would be both
• Attitude is emotion
• Smith, Kendal and Hulin (1969)
• Vroom (1964)
• Why?
• Little thorough conceptual attention at the start
• The tripartite model of attitudes of course
  justified the mix

11
Two Models of Attitude Structure
Classic Tripartite Model
Attitude
Affect
Beliefs
Behaviors
Attitude Evaluative Judgment
Newer Model
Behaviors
12
What is affect?

• Definitions abound but most researchers talk
  about these distinguishing characteristics
  (whether mood or emotions)
• Subjective experience
• Physiological changes
• Behavioral tendencies
• Time bound
• Does satisfaction (attitude have any of these
  characteristics?)

13
Lets look at some dataOne Managers Emotional
States
14
Another Managers Emotional States
15
Together
16
One Graduate Student
Parents called with a birthday surprise
Was not the regular tension in lab meeting
Stats exam returned
Lunch with a friend
?
Argument with a coworker
?
Breakdown in front of advisor
nosebleed
17
A Second Graduate Student
In charge of training session that got off to a
good start
Received proofs of paper in press
Good talk with advisor
Got equipment working
Best friend got job interview
Worked on writing research proposal
Officemate is leaving
Problems with equipment
18
We have to remember

• That with satisfaction what we are measuring an
  attitude which is an evaluative judgment about
  ones job
• That affective experiences both influence those
  judgments but are conceptually distinct
• The same is true about beliefs about ones job

19
Where is the empirical evidence?

• Factor analyses separate affect from evaluations
  and beliefs (when affect is measured correctly) -
  ex. Crites et al. (1994)
• Affect and beliefs separately and independently
  predict attitudes/satisfaction - ex. Weiss, et
  al. (1999)

20
Lets look at some dataOne Managers Emotional
States
21
Where is the empirical evidence?

• Factor analyses separate affect from evaluations
  and beliefs (when affect is measured correctly) -
  ex. Crites et al. (1994)
• Affect and beliefs separately and independently
  predict attitudes/satisfaction - ex. Weiss, et
  al. (1999)
• Affect and beliefs predict behavior differently

22
Is this just semantics?
You can judge for yourselves
But I believe the confusion between affect and
attitudes

• Is just plain sloppy conceptualization
• Has obscured differences among important
• but related constructs
• Has contributed to the neglect
• of the study of true emotion at work
• Has contributed to the inability to find
  consistent
• satisfaction-performance relationships
• Has produced real measurement problems

23
Lets look at a related set of constructs

• Meyer and Allen describe three types of
  commitment
• Affective commitment
• Normative commitment
• Continuance commitment
• But, what Meyer and Allen call affective
  commitment has very little affect in it. Mostly
  its about value congruence and identity.
  Important but not affect

24
Finding Affect in Affective Commitment

• Objectives
• Validate partition of affective commitment
  dimension
• Begin development of new measure
• Method
• Large single organization survey

25
Finding Affect in Affective Commitment

• Some results
• Five factor model is best fit to data
• Affective
• Value congruence/identity
• Normative
• Alternatives
• Investments
• Affective commitment and value congruence were
  highly correlated but had independent,
  significant relationships with ITL
• Daily hassles predicted the affective commitment
  when value congruence was controlled for but did
  not predict value congruence when affective
  commitment was controlled for

26
Final Items for Index

• Affective Commitment (a .87)
• Generally, on a day-to-day basis, I am happy with
  my work at -----.
• I enjoy working at -----.
• Generally, on a day-to-day basis, I am proud to
  an employee of ----.
• Commitment through Value Congruence (a .85)
• I feel like being an employee of ---- can help me
  achieve what I want in life.
• I really feel as if ------ values are my own.
• Working at ------ is consistent with my personal
  goals.
• Commitment through Investments (a .76)
• If I left ------, I would feel like I'm starting
  all over again.
• It would be difficult for me to leave ------- and
  give up the benefits available in -----.
• I continue to work at ---- because leaving would
  require considerable sacrifice.
• Commitment through Alternatives (a .85)
• I would have difficulty finding a good job if I
  left -----.
• One of the problems with leaving ------- would be
  the lack of available alternatives.
• Normative Commitment (a .75)
• I would feel guilty if I left ------.
• I would not leave -------- now because I have a
  sense of obligation to the people here.
• If I left -------, I would feel like I had let
  people down.

27
Finding Affect in Affective Commitment

• In another study with teachers using our scale
• Affective Commitment correlated with burnout but
  Value Congruence did not
• Overall, I think we can distinguish between a
  commitment based on daily experiences/happiness,
  etc. and a commitment based on identity,
  perceived fit or value congruence.
• But this is very preliminary, particularly the
  measurement

28
A Preliminary Taxonomy
To fully assess the nature of life at work we
have to assess all three families
What people think and feel
29
Lets Briefly Talk About The Time Frames of
Constructs

• Constructs with time frame indeterminate (time
  frame not part of construct definition)
• Satisfaction
• Well being
• Constructs with longer episodes
• Loneliness
• Depression
• Stress (?)
• Constructs with shorter episodes
• Emotion
• Mood
• Stress (?)

30
A Preliminary Taxonomy
To fully assess the nature of life at work we
have to assess all three families
What people think and feel
31
Q. Why is this important?A. Different
Performance Relationships

• The satisfaction-performance relationship has
  been the Holy Grail of work research
• But now we see its not one cup but (at least)
  three cups
• One cup holds the effects of evaluations on work
  behaviors
• One cup holds the effects of beliefs
• One cup holds the effects affective states

32
Another way to look at it
Judgment Driven Behaviors
Features
Beliefs
Affective Experiences
Satisfaction
Events
Affect Driven Behaviors
Weiss and Cropanzano, 1996
33
What are measurement implications? Lets
brainstorm a bit

• At the very least, we need separate measures of
  affective experiences, beliefs, evaluations.
• Each will have something to say about work
  experience
• Each will predict something important about work
  behaviors
• I think we have to carefully understand the
  relevant states of being among workers, both
  momentary and elongated

34
What are measurement implications? Lets
brainstorm a bit

• Surveys are very good for some things, very
  limited for others
• Theyre good at the evaluation part, the belief
  part.
• They my be OK at the state of being part
  (telling whether you feel bored or depressed,
  etc.)
• Theyre very limited at capturing experiences and
  events (frequency of this or that, etc.)
• Over time people stop thinking episodically and
  start thinking semantically

35
In conclusion

• The measurement of affect and attitude among
  workers has been characterized by imprecise
  conceptual development
• Precision will breed better measurement and
  better utility

Thank You