PSY 6450 Unit 8

1
PSY 6450Unit 8

• Performance and Pay

2
Schedule

• Exam (35 points) on Monday, 11/19
• Reminder - if you want your project grade before
  ME2, projects are due Monday, 12/01
• Otherwise, final due date during final exam week,
  Monday, 12/08

3
A Little Introduction

• My main thematic line of research has been the
  effects of individual and small group monetary
  incentives on employee performance and
  satisfaction
• Most of that work has been bridge research

4
Bridge Research

• Bridge research
• Laboratory simulations that address practical
  questions from organizational settings
• The main advantage of bridge research
• Isolation of the effects of incentives from
  administrative and organizational changes
• The main disadvantage of bridge research
• Relevancy to actual work settings

5
Relevancy to Work Settings

• Caution is good (and recommended)
• Reviews, though few in number, have indicated
  that the results of field studies and laboratory
  studies are similar
• Jenkins (1986), Jenkins, Mitra, Gupta Shaw
  (1998)
• Hantula (2001)
• Bucklin Dickinson (2001)
• We have found similarities as well
• Frisch Dickinson (1990)
• LaMere, Dickinson, Henry, Henry Poling (2000)

6
How I Got My Start The Practical Beginning

• Union National Bank, Little Rock, AR
• Individual monetary incentive systems implemented
  in early 1980s
• William B. Abernathy, H. Hall McAdams, Wayne
  Dierks, Kathleen McNally
• By the late 1980s, 75 systems had been installed,
  covering 70 of the banks 485 employees
• Productivity increases of 200-300
• Net profit per employee 11,000 compared to
  4,950 for other Little Rock banks
• Program committee for ABA
• While scheduling, call from someone important

7
How I Got My StartThe Academic Beginning

• The role of financial compensation in industrial
  motivation
• Opsahl Dunnette (1966)
• Appealed to researchers to conduct controlled
  laboratory studies because

8
Opsahl Dunnette, 1966
Strangely, in spite of the large amounts of
money spent and the obvious relevance of
behavioral theory for industrial compensation
practices, there is probably less solid research
in this area than in any other field related to
worker performance. (p. 94)
9
Individual Monetary IncentivesResearch Since
1966

• 20 years later 28 systematic studies of
  individual monetary incentives
• Jenkins (1986)
• 12 years after that 39 systematic studies of
  individual monetary incentives
• Jenkins et al. (1998)
• Meta-analytic review
• Individual incentives had
• an overall effect size of .34

excludes survey studies
10
Group Incentive Research

• 13 published studies
• Honeywell-Johnson et al., 2002 Culig, 2005
• Only 8 have compared the effects of individual
  incentives and small group incentives (N 2-12
  group members)
• My students and I have done 3 of those

excludes survey studies
11
Prevalence in BusinessIndividual Monetary
Incentives

• 90 of Fortune 1000 companies have some type of
  individual incentive plan
• Ledford, Lawler, Mohrman (1995)
• 47 of 1045 companies surveyed by Hewitt had
  individual incentive plans
• Rewards were based on specific
• employee performance criteria
• Hewitt Associates (2005)

12
Prevalence in BusinessGroup Monetary Incentives

• 87 of Fortune 1000 companies have work group or
  team incentives
• Lawler, Mohrman, Ledford (1998)
• Fortune 1000 companies increased their use of
  work group or team incentives by 50 between 1987
  and 1996
• Ledford Hawk (2000)

13
About The Beginning Again

• My students and I began this research In the late
  1980s when only about 20 systematic studies had
  been conducted
• We began with questions raised by both Union
  National Bank and Opsahl Dunnette
• What relationship is there between the of
  incentive pay and performance?
• What relationship is there between the absolute
  amount of the incentive offered and performance?
• Are small group incentives just as effective as
  individual incentives?
• Do small group incentives decrease the
  performance of high performers?

14
Caveat Quote of the Day
A careful examination of criticisms of monetary
pay-for-performance systems indicates not that
they are ineffective, but that they are too
effective. (p. 597) Baker, Jensen, Murphy
(1988)
15
Finally, SO1 Three goals and three types of
equity related to each

• To attract and retain good employees - External
  equity
• Are the salaries/pay of employees competitive
  with what other companies are offering, both in
  the local community as well as in the particular
  industry
• To insure that the salary attached to a
  particular job is fair in terms of the importance
  of the job to the organization - Internal equity
• This goal has nothing to do with how well a
  particular individual performs the job - it only
  relates to the relative worth of the job to the
  organization (but influenced by supply and
  demand)
• To motivate and reward employees - Individual
  equity
• Does the pay system encourage high performers?
  Are higher performers paid more than lower
  performers?

(computer programmers vs. secretaries, Business
and engineering profs vs psychology vs English,
behavior analysts tend only to focus on 3) )
16
SO2 The Motivation Problem with Hourly and
Salary Pay

• You get what you pay for
• If you pay for hours, you get hours, not
  performance. Economically it makes more sense for
  employees to take as much time as possible to
  complete their work. And, if you can finagle
  overtime, all the better (Overtime 150 or 200
  of base)
• Consequences
• In hourly wage systems, there are clear
  consequences for performing below a minimally
  acceptable performance level (criticism, threats
  of dismissal), but there are no clear
  consequences for performing above that level.
  Thus, hourly wage systems tend to support
  minimally acceptable performance

17
SO2 The Motivation Problem with Hourly and
Salary Pay

• Merit pay hourly increases
• Even if companies adopt a merit pay policy,
  there is still a weak link between performance
  and pay.
• Why? Merit pay is almost always based on annual
  subjective performance appraisals. And, as
  indicated in previous units, self-assessments
  often do not agree with supervisory assessments
  (we rate ourselves higher than our supervisors
  rate us) hence employees do not believe their
  pay is related to their performance in a
  meaningful way

18
SO3 Skinner on Incentives

• We know, without a doubt that monetary incentives
  will increase performance but
• They have been given a bad rap - perhaps for good
  reasons - many object to them
• Primary reason - they are exploitative
• And they can be, but they dont have to be -
  dont throw the baby out with the bath water!
• Skinner maintained that incentive systems may, in
  fact, be less aversive than hourly pay systems

19
SO3A According to Skinner, what maintains
performance under hourly wage systems?

• No one works on Monday morning because he is
  reinforced by a paycheck on Friday afternoon. The
  employee who is paid by the week works during the
  week to avoid losing a standard of living which
  depends upon a weekly wage. A supervisor who can
  discharge him is an essential part of the system.
  Rate of work is determined by the supervisor
  (with or without the pacing stimuli of a
  production line), and special aversive
  contingencies maintain quality. The pattern is
  therefore still aversive.
• Somewhat better contingencies are available
  under schedules of reinforcement based on
  counters rather than clocks. (ratio schedules
  of reinforcement)

(what Skinner is pointing out here is that hourly
wage systems control behavior via aversive
contingencies - use this as a defense of reward
systems in general, by the way)
20
SO3B Skinner on incentive systems

• Skinner readily acknowledges that incentive
  systems can and have been misused, but notes that
  incentive systems
• May evoke feelings of confidence, certainty of
  success, and enjoyment arising from a sense of
  mastery and effectiveness, and interest in the
  job as occurs when behaviors are frequently
  reinforced.
• Note that Skinner is not addressing performance
  issues here, but rather addressing the fact that
  incentive systems may be less aversive
  emotionally.
• Also, note the italicized section - this is
  important
• Incentive systems are no different in this
  respect than any type of reinforcement system
  where individuals are frequently reinforced
• Respondent behavior interpretation
• R (work) gt Sr (incentives)
• CS (incentives) gt CR (feelings of
  confidence, etc.)

(anectodal - MI disposal aunt, happier, loyal,
improved marriage UNB proof operators, Kate
acousted in grocery store - dont hire anyone
else))
21
SO4 Bucklin Dickinson, intro, NFE

• Bucklin and I were interested in determining
  whether different types of monetary incentive
  systems affected behavior differently (not
  whether incentives were effective, we knew they
  were)
• We discovered, as did Jenkins et al., simply was
  not a lot of research
• Only three thematic lines of research that have
  investigated/manipulated parameters of incentive
  systems
• Percentage of incentive earned 5 studies
• Schedules of reinforcement 8 studies
• Per piece amount 2 studies

22
SO4 Bucklin Dickinson intro, cont, NFEThree
Thematic Lines

• The percentage of total pay or base pay earned in
  incentive pay
• 3-100 of total wages or base pay wages earned
  in incentive pay
• Schedules of reinforcement
• Incentives delivered on different fixed and
  variable ratio schedules (CRF, FR2, VR2, VR4)
• Linear, accelerating and decelerating piece rate
  pay systems
• Piece rate amount remains constant, increases or
  decreases as the number of pieces completed
  increases

23
SO4A Conclusions, for exam

• Will be provided in lecture

24
SO4B Implications

• It appears that you dont have to worry a lot
  about the details of how incentives/consequences
  are related to performance - as long as they ARE
  related in some type of ratio schedule, delivered
  fairly frequently, and supported by some type of
  on-going feedback system.

25
SO5 Three reasons why it is not surprising that
profit sharing has not been shown to increase
performance(intro, NFE)

• In the study objectives I describe some popular
  types of nontraditional pay systems
• Profit sharing
• Gain sharing
• Bonus or lump sum payments
• Group incentive plans
• Individual incentive plans
• Pay for skill and knowledge
• Employee stock ownership plans

26
SO5 Three reasons why it is not surprising that
profit sharing has not been shown to increase
performance(still intro, nfe)

• First, note that I have not referred to all of
  these as incentive systems or
  pay-for-performance plans
• In order for me (Abernathy agrees) to classify a
  pay system as an incentive or pay-for-performance
  system it must use a predetermined formula to tie
  compensation to objective performance,
  operational or economic measures.
• This eliminates
• Bonus and lump sum payments, pay for skill and
  knowledge, and employee stock ownership plans

27
SO5 Three reasons why it is not surprising that
profit sharing has not been shown to increase
performance(still intro, nfe)

• Of these, profit sharing is the most prevalent
• Also represents the other end of the continuum
  from individual incentives with respect to two
  very important variables that affect performance
• Number of individuals whose performance
  contributes to the determination of how much
  money each employee gets
• Disbursement system - how frequently the money is
  disbursed
• So I am going to analyze this system, and you can
  do similar analyses for the ones in between

(in any event, back to profit sharing)
28
SO5 Essential features of profit sharing (NFE)

• When annual profits are above a predetermined
  level, part of those profits are distributed to
  employees
• Formulas for distribution are quite complicated,
  but usually the amount of money that is
  distributed to any one employee is based on a
  percentage of the employees salary, thus
  employees do NOT get the same amount
• The money is usually distributed annually, or
  more commonly, placed directly into the
  employees retirement account (tax benefits)

29
SO5 The first reason why profit sharing often
does not increase employee performance

• Profits are based on the aggregate performance of
  all members of the organization. Thus (depending
  upon the size of the company) one persons
  performance contributes only a very small
  proportion to the total performance of the
  organization. Hence a persons performance is not
  strongly related to his/her pay.
• Even with only 100 employees, any one
  individuals performance contributes only 1 to
  the total performance of the organization
• In small companies, however, profit sharing might
  just affect performance

(in sos, but not explained adequately for the
exam mistake - its not that everyone gets a
small )
30
SO5 The second reason why profit sharing often
does not increase employee performance

• Profits are often affected by factors that (a)
  have little to do with the performance of
  individuals and (b) are outside of their control
  such as - include following in answer
• Mergers, acquisitions, building a new factory or
  plant, investment of funds in research (bonuses
  highly uncertain and unpredictable)
• Union National Bank - the performance of the
  proof operators actually had little to do with
  the overall profitability of UNB

(in sos but not explained adequately what is the
main factor that influences bank profits?)
31
SO5 The third reason why profit sharing often
does not increase employee performance

• Annual distribution of profit-sharing bonuses or
  distribution of money into retirement accounts
• Simply too delayed to have much effect on
  performance

32
Percentage of Base Pay or Total Pay Earned in
Incentive Pay
(next several SOs relate to the incentive
percentage studies)
33
Percentage of Incentive

• In many incentive systems employees receive a
  base pay and can earn additional money in
  incentives when performance exceeds a specified
  standard
• Given that the total amount that can be earned
  remains constant, as the percentage increases,
  more of a persons pay becomes dependent upon
  performance

34
Incentives as a Percentage of Total Pay
35
Percentage of Incentive StudiesMain Research
Questions (NFE)

• What is the lowest percentage of incentive pay
  that affects performance?
• Do different percentages of incentive pay affect
  performance differently?

36
SO9 Whats the magic percentage of incentive
according to compensation experts?

• Will be provided in lecture

(based on tradition - WWII, war labor relations
board)
37
(No Transcript)
38
SO10 Employee perceptions of fairness of
incentive percentages

• If employees do not have a high degree of
  control over their performance, why are they
  likely to perceive high percentages of incentive
  to be unfair?
• A sizable portion of their total earnings will be
  based on factors outside of their control (i.e.,
  cant control their own earnings much), and
  further,
• If those factors fluctuate from day to day or
  week to week, their earnings will not be
  predictable
• People have fixed living expenses
• Apt rent or home mortgage
• Car payments
• Expenses for kids

39
SO11 Employee perceptions of fairness of
incentive percentages

• If employees do have a high degree of control
  over their performance, why are they likely to
  perceive low percentages of incentive to be
  unfair?
• With low percentages of incentives, base pay
  constitutes a relatively high portion of their
  total earnings and the incentive earnings
  constitute a relatively low portion, thus
• Differences in performance between individuals
  will not be adequately reflected in differences
  in earnings
• That is, low performers will earn just about as
  much as high performers
• Not related to the fact that individuals wont
  make a lot more money if they perform better -
  that is a satisfaction issue, not a fairness
  issue This is a common error by students on the
  exam!

(some have had trouble with this in the past)
40
Frisch Dickinson, 1990

• Participants 75 college students
• Five conditions
• Hourly pay (0 of pay)
• Incentives
• Planned 10, 30, 60, or 100 of base pay
• Actual 3, 13, 25, 54 of base pay (cant
  calculate this until after the study is over and
  you know how much participants actually earned -
  we assumed participants would perform better than
  they did)
• Sessions Fifteen 45-minute sessions
• Task Simple assembly task
• Assembling parts from bolts, nuts and washers
• Measure Number of correctly assembled parts

41
25 54 3 13
Incentive pay
Hourly pay
Hourly pay
42
Summary of Results Frisch Dickinson

• Participants who were paid incentives performed
  significantly better than those who were paid
  hourly
• Participants who were paid incentives performed
  comparably, regardless of the percentage
• 3, 13, 25, and 54

43
SO13A The relationship between the amount of pay
earned and the percentage of incentive
Most
0 3 13 25 54
Least
Inverse relationship between the amount earned
and incentive percentage
44
SO13B Why is that relationship important?

• It helps answer the following two questions
• Did people perform better because they earned
  more money?
• In other words, does the total amount of money
  earned affect performance? Is that a critical
  determinant of performance?
• Did people perform better because they received
  more money per piece (per part assembled?)
• In other words, does the amount of the per piece
  incentive affect performance?

(students have had trouble with this so in the
past, so I want to start with this material - the
actual answers are on the next slide)
45
SO13B So, why is that relationship important?

• Participants who earned incentives made less
  money than those who were paid hourly, but
  performed significantly better thus the total
  amount of money earned cannot account for the
  higher performance
• Participants in the four incentive groups
  received different per piece incentives, yet they
  performed the same, thus the per piece incentive
  did not affect performance

46
SO 14 Frisch DickinsonParticularly
Interesting Results

• Those who received only 3 of their base pay in
  incentives - only 11 per 45-minute session -
  performed significantly better than those paid
  hourly
• Higher percentages of incentives did not result
  in better performance - rather participants who
  earned different percentages of incentives
  performed the same

47
LaMere et al. Field Study, 1996, intro

• There is actually only one study objective for
  the exam over this study, but it was a very
  important study from our perspective
• We had found that
• a very low incentive percentage (3)
  significantly increased performance and
• higher incentive percentages did not increase
  performance
• Was that an artifact of the study being conducted
  in the laboratory?
• In the LaMere et al. field study we were able to
  examine the effects of three incentive
  percentages (3, 6, and 9) on the performance
  of actual workers

SO15 Lowest and highest incentive percentage
examined?
48
LaMere et al. Field Study, 1996, intro

• Participants 22 roll-off truck drivers
• Deliver large waste disposal dumpsters to
  commercial and construction sites
• Multiple baseline design across 2 groups
• Hourly pay G1, 20 weeks G2 34 weeks
• 3 incentive G1, 28 weeks, G2 15 weeks
• 6 incentive Both groups, 39 weeks
• 9 incentive Both groups, 107 weeks

(collected data for almost 4 years!)
49
Roll-off Trucks
50
LaMere et al. Field Study

• Incentive pay
• Per job incentive for above average weekly
  performance
• Controlled for different types of jobs and the
  number of miles driven
• Lost incentives for the week for a chargeable
  accident
• Received as part of weekly paycheck, but the
  amount of incentives was listed separately on the
  pay stub
• Feedback
• Daily self-recorded feedback
• Group performance was graphed weekly and publicly
  posted

51
Results LaMere et al.

• Both groups significantly increased their
  performance when the incentive system was
  introduced
• Both groups maintained their high performance for
  the rest of the study (almost 3 years)
• Both groups performed comparably when paid 3, 6
  and 9 incentives

52
Conclusions LaMere et al.

• Results supported our laboratory study
• Small percentages of incentives, as low as 3 of
  total pay, can significantly increase performance
• Higher percentages do not result in incrementally
  better performance
• Small percentages of incentives can sustain
  performance over time (3-9)

53
SO17 Conclusions Percentage of Incentive
Studies

• Results of all five studies have been consistent
  (3 - 100 of total pay)
• Different incentive percentages resulted in the
  same level of performance that is, higher
  incentive percentages have not increased
  performance more
• Low percentages of incentive, as low as 3, have
  significantly increased performance

(I realize this is basically the same answer as
the answer to SO14 why the results of Frisch
Dickinson were particularly interesting - but I
wanted you to note those results before we got to
this point in the article.)
54
SO20 What is the major question with respect to
this research?

• Feedback was readily available in every study
  but one, thus the major question
• Did performance feedback sustain performance
  across the incentive percentages? That is, if
  feedback was not provided, would performance
  differ as a function of different incentive
  percentages?

(skipping sos 1819 - you can do those on your
own)
55
Schedules of Reinforcement, intro

• Results of comparisons of different schedules of
  delivery (reinforcement) are ambiguous at best
• Although incentive pay increased performance in 7
  of 8 studies, no uniform differences emerged as a
  function of the schedule of delivery
• The preponderance of data from well controlled
  laboratory studies, absent implementation
  problems in the field, suggest that different
  ratio schedules of delivery result in comparable
  performance

56
SO21 Two factors that could account for the
performance differences in applied studies of
ratio schedules (give example)

• Although the data suggest that performance does
  not differ under different schedules of delivery,
  sometimes performance has been better under one
  schedule than another
• These differences have occurred only in applied
  settings, not well controlled laboratory studies
• What are the two factors?
• Rule statements
• Social contingencies

57
Example

• Results were very different in two very similar
  studies that compared hourly pay with a VR
  schedules
• In one workers performed better when paid hourly
  than when they received the incentives (Yukl
  Latham, 1975)
• In the other workers performed much better when
  they received the incentives than when they were
  paid hourly (Latham Dossett, 1978)

58
Yukl Latham Hourly pay better

• Participants were tree planters
• Received a base pay
• 4.00 on a VR2 for planting a bag of trees
• VR2 was achieved by tossing a coin and having
  workers guess heads or tails
• Workers had a one-in-two (50) chance of getting
  the 4.00 each time, which probabilistically
  equals a VR2 schedule

59
Yukl Latham Hourly pay better

• Several workers believed that the coin toss was
  gambling and the devils doing and sinful
• One of the supervisors was a part-time minister
  who also believed the above, and thus did not
  always implement the coin toss as planned
• One worker believed management had cheated her on
  her taxes with respect to the incentives and told
  others
• Management had made a mistake but by the time it
  was discovered and fixed, the study was over
• What the workers said about the incentives
• The VR schedule was unfair
• Too much of a risk
• A real let-down to lose after you have planted
  1,000 trees (one bag 1,000 seedlings)

(study conducted in the Bible belt in the south
yet in the other study..)
60
Latham Dossett Incentives better

• Participants were from the same company, but were
  beaver trappers
• Received a base pay
• 4.00 on a VR4 schedule for each beaver trapped
• VR4 was achieved by placing four different
  colored marbles in a bag and having workers guess
  the color of the marble before they drew one
• Workers had a one-in-four (25) chance to be
  correct, which probabilistically equals a VR4

61
Latham Dossett VR Incentives better

• All workers gathered around while a worker
  guessed the color of the marble and cheered
  when he guessed correctly
• What the supervisor said about the variable
  schedule
• The guys want to get on the variable schedule.
  The men are inspired by it. They get a real kick
  out of it.
• What the workers said about the schedule
• We really get psyched out by the variable, man
• Like the variable, it adds something to the job
• It makes the job more exciting and fun there is
  real excitement

62
SO22 Oah DickinsonLinear vs Accelerated
Piece Rate Pay

• Does accelerated incentive pay affect performance
  differently than linear incentive pay?
• Linear incentive pay The employee earns the same
  incentive pay regardless of how productive he/she
  is
• Accelerated incentive pay The employee earns
  more and more incentive pay the more productive
  he/she is

(SO results - introduce the article)
63
Accelerated vs. Linear Pay, intro

• Reward magnitude question
• The harder you run the harder it is to run
  faster, therefore,
• Do employees perform better when they receive
  increasingly more incentive pay for higher and
  higher levels of performance?
• (more technically, as response effort increases,
  is more and more money required to increase
  performance?)

64
Oah Dickinson, 1992, intro

• Modeled after the proof operator incentive system
  at Union National Bank
• Participants 40 college students
• Task Data entry task
• Checks of different cash values were presented
  on the computer screen and Ps entered the cash
  values

65
Oah Dickinson, 1992, intro

• Two conditions
• Linear relation between performance and pay
• Incentive amount remained the same
• 1.5 exponentially-increasing relation between
  performance and pay
• Sessions Fifteen 45-minute sessions
• Measure Number of correctly completed checks

66
(No Transcript)
67
SO22 Answer

• Results
• Participants in the two groups performed
  comparably
• Participants in the 1.5 exponential group earned
  significantly more money
• Conclusions
• Linear and accelerated incentive pay did not
  affect performance differently (more piece rate
  pay was not better)
• The amount of incentive pay did not affect
  performance differently

68
SO23 Two factors that influence satisfaction
ratings with different types of incentive pay
plans

• Exposure to all of the pay systems you are going
  to ask employees about (behavioral choice is the
  best method to obtain satisfaction data)
• If you use a between group design and each
  participant is exposed to only one pay system and
  then asked to rate his/her satisfaction with it,
  you get different ratings than if each
  participant is exposed to all pay systems before
  rating his/her satisfaction with it
• Makes sense - participants can only make
  meaningful ratings and comparisons after exposure
  to the different pay systems
• Probably true for other types of interventions as
  well - not just pay systems, so it is a good
  thing to keep in mind when doing research

(no consistent data with respect to employee
satisfaction with different types of pay systems
- fixed vs incentive or different types of
incentive pay)
69
SO23 Two factors that influence satisfaction
ratings with different types of incentive pay
plans

• The amount of money employees earn under the
  different pay systems - not surprisingly, The
  pay system I like best is the one where I earn
  the most money.
• It is VERY difficult to equalize the amount of
  money individuals earn under various pay
  conditions and systems
• The optimal situation would be where a person
  earns, lets say, 6.00 an hour when paid hourly
  and 6.00 an hour when paid incentives
• With incentive systems, however, a persons
  performance determines how much he or she will
  earn, and it is very hard to predict how well a
  person will perform when you are setting up the
  incentive rates at the beginning of a study

(this is very influential factor and one that
makes it difficult to assess employee
satisfaction with different pay systems)
70
SO24 Do results from the lab generalize to
actual work settings?

• Reviews of incentive studies, though few in
  number, have indicated that the results of field
  studies and laboratory studies are similar
• Jenkins (1986), Jenkins, Mitra, Gupta Shaw
  (1998)
• Hantula (2001)
• We have found similarities as well
• Frisch Dickinson (1990)
• LaMere, Dickinson, Henry, Henry Poling (2000)
• In both, performance increased significantly when
  participants received only 3 of their total or
  base pay in incentives
• In both, higher percentages of incentives did not
  increase performance further

(as I indicated earlier - important!!)
71
SO24 Also true for other IVs

• Locke, E. A. (1986) (Ed.). Generalizing from
  laboratory to field settings. Lexington, MA D.C.
  Heath Company
• Each article in the book analyzed the extent to
  which the results of laboratory and field studies
  were similar for a particular performance
  improvement intervention
• When there was sufficient data to make the
  comparison, every review reported that the
  results in laboratory and field settings were
  similar
• True for monetary incentives, feedback, goal
  setting, training, participation (among others)
• Interestingly, the author who reviewed feedback
  (Kopelman) reported that while the general
  effects of feedback were the same in both
  settings, the effects of feedback were actually
  less in lab studies that is, the results from
  the laboratory underestimated the extent to which
  feedback affected performance in actual work
  settings
• This makes sense - why, based on what you learned
  about feedback in U6?

(next slide, quote from Locke)
72
SO24 Quote from Locke, p. 6 (NFE)

• Both college students and employees appear to
  respond similarly to goals, feedback, incentives,
  participation, and so forth, perhaps because the
  similarities among these subjects (such as in
  values) are more crucial than their differences.
  Task differences do not seem to be overwhelmingly
  important. Perhaps all that is needed is that the
  participants in either setting become involved in
  what they are doing.

(handy book to know about)
73
SO25 Honeywell-Johnson et al. article

• Purpose of the study (NFE)
• To compare the effects of individual and small
  group monetary incentives on the performance and
  satisfaction of high performers
• Secondary purpose (also NFE)
• To assess the feasibility of using simulated
  groups to examine the effects of small group
  incentives on performance

(we examined small group incentives, but results
may well generalize to any type of performance
consequence that is dependent upon the groups
performance grade on a project or in a class??)
74
SO25A Conceptually why might individual
incentives control performance more effectively
than small group incentives?

• When individuals are paid individual incentives,
  the amount of pay they earn is directly dependent
  on their own performance. Thus, they have
  complete control over what they earn
• With small group incentives, the amount of pay
  workers receive is not only dependent on their
  own performance, but on the performance of others
  in the group. Thus, they cant influence their
  earnings to the same extent as they can when they
  are paid individual incentives

(both points are important)
75
SO25B On the other hand, why might small group
incentives control performance just as well as
individual incentives?

• When the group is small, certainly when the group
  is only 2 3 members, the individual can still
  greatly influence the performance of the group,
  and hence his or her pay. There is still a pretty
  tight contingency between performance and pay
  (Malotts certainty factor?).
• But, as the group size increases, the
  individuals contribution to the groups
  performance becomes less and less, and earnings
  are not as dependent upon their own performance

76
SO26 Summarize the results of studies that have
compared individual incentives and
equally-divided small group incentives, intro

• Equally-divided small group incentives are the
  most common type of small group incentives in
  business and industry
• The incentives that a person gets depends upon
  how well the entire group performs
• Each individual in the group gets the same amount
  of incentive, regardless of his or her
  contribution
• For example, if a group completes an average of
  100 widgets per hour during the week, each member
  earns an additional 100.00 per week in
  incentives (1.00 per widget in the average)
• If one worker averages only 90 widgets per hour
  and another averages 110, they both get the same
  amount of incentive

77
SO26 Summarize the results of studies that have
compared individual incentives and
equally-divided small group incentives, for exam

• Only five studies have compared performance when
  individuals were paid individual incentives and
  when they were paid small group incentives
• Groups have ranged in size from 2 - 12 members
• In 4 of the 5 studies, equally-divided small
  group incentives sustained performance as well as
  individual monetary incentives
• While there are not a lot of satisfaction data
  available, when satisfaction was examined,
  participants reported that they were equally
  satisfied with the individual and small group
  incentives

(include range of participants in the small
groups in these studies - this is important)
78
SO27A When would an individual be expected to
perform the same when paid individual incentives
and small group incentives and why?

• When all members of the small group, including
  the individual, perform at approximately the same
  level
• Why?
• If all group members perform at approximately the
  same level, then their pay would not differ much
  when they were paid individual incentives and
  when they were paid small group incentives

(this seems to be a difficult notion for students
to understand example on next slide)
79
SO27A When would an individual be expected to
perform the same when paid individual incentives
and small group incentives and why?

• If all group members perform at approximately the
  same level, then their pay would not differ much
  when they were paid individual incentives and
  when they were paid small group incentives
• Assume under the individual incentive condition,
  a person is paid 10 per widget assembled. If the
  individual assembled 50 widgets, he or she would
  earn 5.00 in incentive pay
• Assume under the group incentive condition, each
  individual gets paid 10 per widget based on the
  average performance of the members of the group.
  If each individual in a 5-person group assembled
  50 widgets the total number of widgets assembled
  by the group would be 250, and the individual
  would still earn 5.00 in incentive pay
• 250/5 50, 50 X 10 5.00
• Essentially, the pay contingency is the same for
  the above individual whether or not he or she is
  paid individual or small group monetary
  incentives. We analyze contingencies from the
  perspective of the behaver.

(we know, as does the individual of course, that
they are being paid either individual or small
group incentives, but if they get paid the same,
the contingency between their performance and pay
is the same under both pay systems)
80
SO27B When would an individual be expected to
perform better when paid individual incentives
and perform worse (decrease performance) when
small group incentives? Why?

• When an individual was a high performer in
  comparison to the other members in the group
• Why?
• A top performer would earn less money when he or
  she was paid small group incentives than when she
  or she was paid individual incentives because of
  the lower performance of the other members of the
  group
• Hence, over time, the high performer may decrease
  his or her performance because his or her
  earnings decrease

(example on the next page)
81
SO27B When would an individual be expected to
perform better when paid individual incentives
and perform worse (decrease performance) when
small group incentives? Why?

• Assume under the individual incentive condition,
  a person is paid 10 per widget assembled. If the
  individual assembled 50 widgets, he or she would
  earn 5.00 in incentive pay
• Assume under the group incentive condition, each
  individual gets paid 10 per widget based on the
  average performance of the members of the group.
• Also, now assume each of the other individuals
  in a 5-person group assembled 35 widgets. The
  total number of widgets assembled by the group
  would be 190 (4 members X 35 140, plus the 50
  widgets assembled by the top performer).
• The individual would only earn 3.80 in
  incentive pay.
• 190/5 38, 38 X 10 3.80
• Our top performer thus may perform lower when
  paid small group incentives than when paid
  individual monetary incentives
• Recognizing of course that if he or she did
  increase performance, he or she would earn even
  less money in incentives. For example, if our top
  performer now only assembled 35 widgets, he or
  she would earn only 3.50 in incentive pay -
  however, the small difference between 3.50 and
  3.80 might not be sufficient to keep our top
  performer making 50 widgets

82
SO27 Thought question (NFE) - what are the
implications of this analysis for team/group
projects in classes or business settings?

• Anyone want to share his or her thoughts on this?

83
Honeywell-Johnson et al., intro (NFE)

• Participants were 4 college students
• Task was a computerized task with four sub-tasks
  presented simultaneously
• Memory task, arithmetic task, visual monitoring
  task and an auditory monitoring task
• DV points earned for correct responses
• Design ABCB, 5-10 2-hour sessions per phase
• A hourly pay with feedback
• B individual incentives with feedback
• C small group monetary incentives

84
Honeywell-Johnson et al., intro (NFE)

• Group incentive condition
• Participants were told that they were members of
  a 10-person group and their data would be
  combined with the point scores of the other nine
  members to determine the amount of incentive
• Worked on networked computers to increase
  believability of this deception
• At the end of the study, each participant was
  asked how many members were in their group, and
  all responded 10
• Their comments (some rather nasty and hostile)
  during the group incentive condition indicated
  they believed they were in a group of 10
• The groups average performance was
  contrived/calculated in a manner that insured
  that the participants performance was always
  quite a bit higher than the groups average
  performance

85
SO28 Results of the study and what they indicate

• Three of the four participants performed an
  average of 14 lower (12, 14, and 16) when
  they were paid small group monetary incentives
  than when they were paid individual incentives
• The data suggest that high performers are indeed
  likely to decrease their performance when paid
  small group monetary incentives (with N10
  members)

86
SO29

• 29A Which of the three pay systems did all four
  high performers prefer?
• All preferred the individual monetary incentive
  system
• Even though they found the individual incentive
  system to be more stressful than the hourly pay
  condition (which was the least stressful of all
  three)
• All four said they preferred the individual
  monetary incentive system because they earned the
  most money
• While this is a confound - it is also the case
  that in actual work settings high performers
  would always earn more when they were paid
  individual incentives than when they were paid
  small group incentives, so I am not overly
  concerned about this confound
• 29B Which of the three pay systems did the
  majority of performers (3 of 4) find to be the
  most stressful?
• The small group monetary incentive system

87
SO29, NFE, but interesting

• The preference data are interesting
• Most people would probably assume that the
  individual incentive pay would be the most
  stressful of the three pay systems, yet three of
  the four reported that the group incentive system
  was the most stressful
• Also, in spite of the fact that three of the four
  participants found the hourly pay to be the least
  stressful, all four preferred the individual
  incentive pay
• We have confirmed the effects of group incentives
  on high performers in a subsequent study
  conducted by Dr. McGee as her doctoral
  dissertation. It has been published in the
  Performance Improvement Quarterly (ISPI funded
  the research)

88

• QUESTIONS OR COMMENTS?

89
Theory behind profit-sharing

• Profit-sharing was not originally developed to
  increase employee performance rather it is based
  on macro-economic theory (good of the company
  and the society)
• Profit-sharing would increase the flexibility of
  labor costs for organizations (not decrease labor
  costs)
• When profits went up, labor costs would go up
• When profits went down, labor costs would go down
• When profits were down, profit-sharing was
  supposed to protect the company by automatically
  decreasing labor costs
• Protect the employees because the company could
  afford to keep more employees rather than lay
  them off
• Ultimately, both of the above would protect the
  countrys economy
• Fewer people would lose their jobs and prevent
  the economy from a deeper recession

90
However, the catch.

• Cost savings to companies was based on the
  notion that the base wages of employees would be
  below market value. When profits were high,
  employee wages would be above market value when
  profits were low, employee wages would be below
  market value. It has not worked out that way -
  rather predictably, perhaps, profit-sharing has
  become gravy. Base salaries are at market
  value, so companies do pay out more when profits
  are good, but dont recoup sufficient labor costs
  when profits go down.