Arthur PDF

Academy of Management Journal - In Press
GAINSHARING AND ORGANIZATIONAL LEARNING: AN ANALYSIS

OF EMPLOYEE SUGGESTIONS OVER TIME
JEFFREY B. ARTHUR
R.B. Pamplin College of Business
Department of Management
Virginia Tech University
Blacksburg, VA 24061
Tel: 203 254-4000
e-mail: [email protected]
and
LYNDA AIMAN-SMITH
College of Management
North Carolina State University
P.O. Box 7229
Raleigh, NC 27695-7229
Tel: 919 515-8699
e-mail: [email protected]

Gainsharing and Organizational Learning:
An Analysis of Employee Suggestions over Time
Abstract
This paper addresses the issue of how gainsharing programs work by proposing a model of
gainsharing as an organizational learning system. We analyzed employee suggestions submitted
over the first four years of a gainsharing plan at one plant and found that although traditional
explanations of how gainsharing works predicted changes in volume of total suggestions,
changes in the content of those suggestions were consistent with an organizational learning
process. Suggestions indicating first-order learning were initially high but declined over time as
a percentage of total suggestions, while suggestions congruent with second-order learning
became a larger portion of the total suggestions over time.

After more than fifty years of research and experience, gainsharing has been described as
a practice in search of profits and in need of theory (Hammer, 1988: 339). The lack of more
theory-based empirical evidence on gainsharing effectiveness may be traced to its origins as a
successful management practice rather than an academic study or theory. Fascination with the
promise of gainsharing has existed in the U.S. business community since at least the 1930s. It
was then that Joseph Scanlon, an ex-steelworker, union official, and MIT lecturer advocated a
plan whereby companies combined a plant-wide group incentive system with a comprehensive
employee involvement program to encourage labor-management cooperation and the generation
of employee cost-saving ideas. Early support for the plan came in the form of popular articles on
successful Scanlon Plans at Adamson Co. (Chamberlain, 1946) and Lapoint Machine Tool Co.
(Davenport, 1950). Recent surveys indicate gainsharing is again experiencing increased
popularity, with estimates that over a third of large companies now rely on some form of
gainsharing program (Lawler & Cohen, 1992; Alexander Consulting Group, 1992).
Despite the increasing popularity of these plans, evidence for their effectiveness has
remained mixed. Although there are a number of qualitative studies and some more recent
experimental quantitative studies which have documented gainsharing successes (e.g., Hatcher &
Ross, 1991; Schuster, 1984; Wagner, Rubin, & Callahan, 1988), it is clear that a large number of
programs fail to achieve desired results. Some evidence indicates that more than half of the
gainsharing plans instituted do not survive beyond five years (Goodman, 1980; Scanlon, 1948;
White, 1979), and many appear to begin to lose effectiveness after the first two or three years
(Gray, 1971; Kaufman, 1992; McKersie, 1963).
These mixed results from gainsharing plans have led to calls by researchers to develop a
better understanding of how gainsharing programs work (Bullock & Lawler, 1984; Hammer,

1988; Lawler, 1988). The absence of a strong theory-based understanding of how these
programs work is particularly problematic because gainsharing represents a complex
organizational intervention that requires firms to make a large number of choices about their
implementation and measurement. Decisions must be made regarding the degree and form of
employee involvement, the composition of the incentive formula, as well as the percentage of
employee compensation to put to at-risk. Decisions must also be made about how and when to
adjust the gainsharing formula and payouts based on continuous monitoring of external changes
and the performance of the gainsharing plan (e.g., Ross & Ross, 1984).
When making these decisions, we contend that it is a mistake to view gainsharing as
simply a group incentive and suggestion program and hence to measure its success in terms of
short-term financial performance or the number of suggestions submitted. We argue that such
views and measures overlook the potential of gainsharing as an organizational learning system
with the ability to generate first- and second-order learning over time. By analyzing the content
of employee suggestions over time, this article addresses calls in the gainsharing literature for a
better theory-based understanding of how gainsharing works.
PERSPECTIVES ON GAINSHARING EFFECTIVENESS
A number of theoretical perspectives have been proposed to explain gainsharing
effectiveness (e.g., Welborne & Gomez-Mejia, 1995). Most of these theoretical treatments focus
on one of the two primary characteristics of gainsharing: employee participation and contingent
or performance-based pay. Early gainsharing research tended to focus on the role of participative
management and employee involvement in explaining the success of gainsharing plans. Both
Katz & Kahn (1966) and McGregor (1960) devoted sections of their influential books to
describing how the Scanlon Plan has been able to achieve success by embodying the concepts of

system-wide participation and Theory Y management. An underlying assumption in this
participation literature on gainsharing was that employees possessed an untapped reservoir of
effort and knowledge for improving organizational processes and effectiveness, and that the
Scanlon Plans participation and communication mechanisms as well as an equitable reward
structure release this reservoir in the interest of the firm (Frost, Wakeley, & Ruh, 1974; Hammer,
1988; McGregor, 1960). Research by Hanlon and others (Hanlon and Taylor, 1991; Hanlon,
Meyer &Taylor, 1994) supported the proposition that positive, work-related work group
communication increased with the introduction of gainsharing. Other empirical evidence also
supports the position that employee participation may drive performance outcomes in some
gainsharing plans (Gowen & Jennings, 1990; Rosenberg & Rosenstein, 1980). Conclusions from
the participative management studies of gainsharing must be tempered, however, by the fact that
an independent effect of the gainsharing bonus generally has not been statistically separated in
these analyses. In addition, these studies do not explain positive productivity results in
gainsharing plans such as Improshare which de-emphasize employee participation (Kaufman,
1992).
Several theoretical perspectives suggest that incentive payments play a critical role in
explaining how gainsharing works, within the broader context of contingent-based rewards
(Welborne & Gomez-Mejia, 1995; Gerhart & Milkovich, 1992). Expectancy theory (Vroom,
1964), for example, implies that employee effort under gainsharing is determined by the degree
to which there is a clear perceived connection between effort and the desired bonus payoff
(Goodman & Moore, 1976). Operant conditioning theory suggests that the size and timing of the
gainsharing bonus payout are also a critical factors in explaining employee reactions to these

plans (e.g., Mawhinney & Gowen, 1990). In general, research on incentive pay aspects of
gainsharing effectiveness has been relatively sparse.
Finally, several recent studies have focused on the importance of perceived fairness in
gainsharing plans in explaining effectiveness. Cooper, Dyck, & Frohlich (1992) argued that
perceived fairness of the distribution rule is needed in gainsharing to resolve the social dilemma
in gainsharing programs in which free-riders can get the same rewards as contributors. Using
two lab experiments, they found that productivity under a group incentive plan increased when
there was a perceived fair distribution rule (which included both a floor income and entitlement
components) that was participatively developed by employees. Welborne, Balkin & GomezMejia (1995) also found that perceived distributional and procedural fairness was related to
increased mutual monitoring by employees, an agency theory construct that might also explain
the effectiveness of gainsharing programs. These findings reinforce the observations by
Schuster (1984; 1985) and others (e.g., Collins, 1995; Gray, 1971) that the success of
gainsharing depends critically on employee perceptions of fairness regarding the development of
the bonus formula and the distribution of rewards. In the absence of perceived fairness,
employees withdraw effort and ideas that drive performance results.
Organizational Learning Perspective
Organizational learning is a fundamental concept in organizational theory (March &
Simon, 1958; Cyert & March, 1963) that has experienced a resurgence of interest among
researchers and practitioners in recent years and has found a prominent place in the
manufacturing and strategy literatures (e.g., Hayes, Wheelwright, & Clark, 1988). Although
much of the professional literature in this area has emphasized the value of becoming a learning
organization(e.g., Garwin, 1993; Senge, 1990), academic researchers have focused on issues

such as defining the construct of organizational (as opposed to individual) learning, identifying
the barriers to learning and adaptation by organizations, and determining when different types of
learning may be functional or dysfunctional for firm effectiveness and survival (Duncan & Weis,
1979; Easterby-Smith, 1997; Fiol & Lyles, 1985; Huber, 1991; Levitt & March, 1981; Miner &
Mezias, 1996).
Although academic scholars differ in their beliefs about the viability and efficacy of
organizational learning, there appears to be fairly broad agreement concerning the organizational
learning process, and the types of learning which can result from this process (Lant & Mezias,
1992). In perhaps the most widely used model, Argyis & Schn (1978; 1996: 16) described
organizational learning as an organizational change process that begins with organizational
members experiencing a perceived gap between what is expected (or aspiration level) and what
exists (c.f., Cyert & March, 1963; March & Simon, 1958). This perceived performance gap
stimulates a search process by organizational members, taking one of two forms. The first form,
labeled first-order (Hedberg, Nystrom, & Starbuck, 1976) or single-loop (Argyris & Schn,
1978) learning, consists of "a routine, incremental, conservative process that serves to maintain
stable relations and sustain existing rules" (Lant & Mezias, 1992: 48-49). The outcome of this
first-order learning process is expected to be incremental change or adaptation to further exploit
existing technologies, routines, and processes in ways which dont alter underlying assumptions
or values (Argyis & Schn, 1996; Cyert & March, 1963; March, 1991).
Alternatively, this inquiry can result in a second type of learning called double-loop
(Argyris & Schn, 1978; Ashby, 1960), higher- level (Hedberg, 1981; Fiol & Lyles, 1985); or
second-order learning (Hedberg, Nystrom, & Starbuck, 1976; Lant & Mezias, 1992). In contrast
to first-order learning, second-order learning has been described behaviorally as the search for

and exploration of alternative routines, rules, technologies, goals, and purposes" (Lant & Mezias,
1992: 49). From a more cognitive perspective, Argyris and Schn (1996: 24) define this type of
learning as those sorts of organizational inquiry which resolve incompatible organizational
norms by setting new priorities and weighting of norms, or by restructuring norms themselves
together with associated strategies and assumptions. In essence, second-order learning allows
organizations to break out of existing patterns of thoughts or behaviors by exploring qualitatively
different ways of thinking and doing things.
A perceived performance gap is a necessary, but not sufficient condition for first-and
second-order organizational learning to occur. Researchers have identified two additional
necessary conditions. First, organizational members must have the motivation, ability, and
opportunity to inquire into resolving this perceived gap on behalf of the organization [as opposed
to other alternatives such as withdrawal of effort or exit from the organization] (Argyris &
Schn, 1996). In addition, first- and second-order learning by individual organizational
members must be translated or externalized from the tacit knowledge of individuals into a form
that can be utilized by the organization (Duncan & Weiss, 1979; Huber, 1991).
Empirical Organizational Learning Research
Despite the seemingly strong agreement concerning the basic organizational learning
process and outcomes, operationalizing and measuring organizational learning in empirical
research has proven to be excruciatingly hard to do (Miner and Mezias, 1996: 95; Tsang,
1997). A basic conceptual and empirical problem in this research is how to identify an incidence
of first-or second-order organizational learning in field settings. The literature on organizational
learning appears to be somewhat divided on this point (Kim, 1993; Leroy & Ramanantsoa, 1997;
Lundberg, 1995; Miner & Mezias, 1996). Studies utilizing a behavioral perspective (c.f. March

& Simon, 1958; Cyert & March, 1963; Levitt & Marsh, 1988) tend to focus on antecedents and
changes in organizational structures, technologies, routines, and systems as manifestations of
organizational learning resulting from target-oriented, trial and error search processes
(Lundberg, 1995: 15). This perspective is embodied in simulation-based studies that focus
attention on environmental conditions associated with periods of radical versus incremental
changes in organizational structures, processes, and strategies (e.g. Lant & Mezias, 1992; Mezias
& Glynn, 1993).
In contrast, a more cognitive-oriented perspective focuses attention on changes in the
values and belief systems underlying individual and collective action. From this perspective, it is
the degree of change in the cognitive maps, or theories of action, of organizational members
that distinguishes between first- and second-order organizational learning (Argyris and Schn
1978; 1996). Identifying incidents of this type of learning requires looking more deeply into the
(often symbolic) meaning that is embedded in an organizational activity or event. Barleys
(1986) study of organization and role changes following the introduction of CT scanners in two
organizations provides a powerful example of this perspective.
Viewing organizational learning from a more cognitive perspective, Argyris & Schn
(1996) note that the distinction between first- and second-order learning may be a less binary
one than might first appear. Focusing on the degree to which problem solving inquiry leads to
solutions that challenge fundamental organization-wide norms or theories of action means that
it is possible to speak of organizational learning which is more or less double-loop. In place of
the binary distinction we have a more continuous concept of learning. (Argyris and Schn,
1978: 25; emphasis in the original).

Organizational Learning and Gainsharing
Although it is beyond the scope of this paper to attempt to resolve these conceptual and
empirical issues in the organizational learning literature, we argue that both a behavioral and
cognitive organizational learning perspective can be used to understand how gainsharing works.
From a behavioral perspective, a gainsharing program itself can be seen as a manifestation of
organizational learning. In other words, a gainsharing program is often established in
traditionally organized production firms as the result of a search process that has been motivated
by some performance crisis (i.e., gap between aspired and actual performance). In these cases,
the introduction of gainsharing could be classified as a second-order organizational learning
event in that it represents a significant change in the routine or system used to reward individual
contributions. Specifically, it shifts the basis of rewards from individual to a group-level
performance and it makes the distribution of these rewards contingent on improved
organizational performance. By embracing a more participative management philosophy and
structure, one could also identify this as an incidence of second order learning based on a shift in
values and theories of action associated with this change.1
In addition, it is possible to view gainsharing as an organizational learning system that
helps to generate first- and second-order learning at the individual and group level. Argyris &
Schn (1996) identified certain organizational structures, behaviors, and cognitive maps as
learning systems because they provide a framework for further problem-solving inquiry and
learning. Examples of these learning systems include channels of communication (forums for
discussion and debate, formal and informal patterns of interaction)...procedures and routines that
guide individual and interactive inquiry; and systems of incentives that influence the will to
inquire (Argyris & Schn, 1996: 28, emphasis added).

Viewing gainsharing as a learning system allows us to re-interpret the functions of
various structural characteristics of the gainsharing plan. For example, the gainsharing incentive
formula can be seen as initiating an employee search process by making monetary rewards
contingent on improving the existing level of organizational performance as measured by the
gainsharing incentive formula.
In addition, we can view the employee suggestion system as a critical mechanism for
transforming the content of individual-level search and knowledge to organizational knowledge.
Ducan and Weiss (1979: 85) define organizational knowledge as that knowledge which is
available to organizational decision makers and which is relevant to organizational activities.
To satisfy this definition, they note that organizational knowledge must be communicable (i.e.,
able to be articulated by individuals) and consensual (i.e., accepted by other organizational
members). In the Scanlon plan, employees communicate their ideas in the form of written
suggestions that they then make public by submitting them to a joint employee-management
department team and screening committee to be evaluated and implemented.
Argyris & Schn (1996: 25-28) posit that the extent of individual inquiry into problem
solving in organizational learning is affected by various behavioral norms and organizational
constraints. Similarly, proponents of Scanlon-type gainsharing plans have argued that the
introduction of more democratic norms and the use of labor-management review committees
serve to transform behavioral norms and to overcome organizational constraints by increasing
the employee identification with organizational-wide goals as opposed to individual or subgroup
(e.g., division, or union) goals (Katz & Kahn, 1966; Frost et al., 1974).
In the last stages of the organizational learning model, individual learning becomes
institutionalized and embedded in new actions (policies, programs, structures) and behaviors

(assumptions, routines, modes of interaction) that impact organizational performance. This
performance impact provides important feedback to individual members whose search process
may be altered by changes in the perceived performance gap. The gainsharing bonus thus
provides employees with explicit feedback on the organizational effects of changes in routines
and employee behaviors.
HYPOTHESES
As can be seen from the above discussion, employee suggestions generated by a
gainsharing program play a critical role in understanding how gainsharing works: both in the
traditional explanations of gainsharing effectiveness as well as from the perspective of an
organizational learning system. Although the ultimate goal of gainsharing is to enhance
organizational effectiveness, the creation of employee suggestions for performance improvement
can be seen as an important intermediate goal of these programs. Given the central role of
employee suggestions in explaining gainsharing effectiveness, the lack of empirical research on
changes in the volume of suggestion activity associated with gainsharing is somewhat surprising.
In this study, we used both traditional explanations of gainsharing effectiveness as well as an
organizational learning perspective to predict changes in the volume and content of employee
suggestions over time. The purpose of these analyses is to gain a more complete understanding
of how gainsharing works.
Changes in Volume of Total Suggestions
Based on the view of Scanlon Plan effectiveness as a function of increased employee
participation, we would expect the volume of gainsharing employee suggestions to increase
rapidly in the period immediately following the introduction of a Scanlon-type plan, then rise at
a decreasing rate, and eventually begin to decline. There are two main reasons to expect this

curvilinear pattern in the volume of total suggestions over time. First, the attention paid to
employees during the early part of the gainsharing program may lead to a Hawthorne effect in
which employees submit an inordinately high number of suggestions. As this attention is
redirected over time, suggestion volume would be expected to decline. This explanation is
consistent with Griffins (1988) finding that that there was a cycle for quality circle
effectiveness, with peak effectiveness occurring in the third year and then declining. BowieMcCoy, Wendt, & Chope (1993) also found a pattern of suggestions consistent with a
Hawthorne effect in the public accounting firm they studied, with the number of suggestions
peaking in the second year of the program and then declining. This same life-cycle observation
has been made by a number of researchers regarding the effectiveness of employee participation
and incentive programs (Cooke, 1989; Gray, 1971; McKersie, 1963).
A second reason to expect a curvilinear pattern for the total number of suggestions
submitted over time is that there exists a finite number of cost saving improvements which can
be made within a given production system. As Hammer (1988) notes, a key assumption of the
participation view of gainsharing is that employees have pent-up ideas which are released in
the form of suggestions once gainsharing is introduced. If this assumption is correct, it follows
that after these pent-up ideas have been submitted, we should begin to see a decline in the
number of suggestions over time.
Hypothesis 1: The volume of gainsharing suggestions submitted by employees will
increase rapidly after the implementation of gainsharing, then level off and begin to decline.
In addition, we expect the level of bonus payments to be positively related to the
variation in the number of suggestions submitted. Gainsharing payouts are expected to impact
the number of suggestions in at least two ways. First, they provide a reinforcement or reward for

previous suggestion making by employees. Expectancy theory (Vroom, 1964) would predict
that employee effort in making suggestions is dependent on both the degree to which that effort
translates into an actual suggestion and the degree to which making the suggestion pays off with
some desired outcome (Goodman & Moore, 1976). To the extent that money is motivating
employee suggestion-making behavior, employees will continue to engage in this behavior only
as long as the behavior is re-enforced by a bonus (Geare, 1976). This expected relationship is
supported by Grays (1971) conclusion that the level of gainsharing payout provided the best
explanation for changes in suggestion making over time at a steel processing company. In
addition, Hammer (1988) posited that bonus payments indirectly reinforce employees trust in
management and commitment to the gainsharing plan. In effect, the bonus payment represents a
promise kept by management to reward employee effort.
Hypothesis 2: The volume of employee suggestions over time will be positively related to
the amount of payout received from the gainsharing plan.
Note that both Hypotheses 1 and 2 are expected to be operating simultaneously. In other
words, initial suggestions are expected to improve firm efficiency dramatically and result in
relatively large initial gainsharing bonuses that then reinforce additional suggestion-making.
However, as additional improvements become more difficult to achieve, we would expect to see
both fewer suggestions and lower bonus payments. A multivariate statistical model is therefore
required to test for the independent effects of these variables.

Organizational Learning and Suggestions
In this paper we argue that seen from an organizational learning perspective, employee
gainsharing suggestions result from employee searches that lead to first- and second-order
organizational learning. If this argument is correct, then the content of employee suggestions
over time should follow a pattern consistent with the search pattern described in the organization
learning perspective. Based on this logic, we propose that the content of gainsharing suggestions
in the period immediately following the introduction of gainsharing will be characterized
primarily by first-order learning suggestions.
There are a number of reasons for this expectation. First, by definition, first-order learning
suggestions do not challenge the status quo in terms of underlying values of the organization and
the nature of the employee-management relationship. Thus the learning model would suggest
that these types of suggestions would dominate the early problem-solving searches by
employees. In other words, employees will be more likely to seek familiar solutions to problems
that do not disrupt basic values in the relationship (Argyris & Schn, 1978; Cyert & March,
1963). To the extent that these types of suggestions work in terms of improving plant
effectiveness, employees will continue to engage in the same search processes and enact the
same structures and behavioral norms (Levitt & March, 1988).
We posit, however, that there is a finite amount of labor cost saving that can be generated
by improvements within the existing production process and wage-effort bargain. As the firm
approaches the limits of the potential for first-order learning to generate additional cost savings,
we expect to see a change in the content of suggestions submitted. Specifically, first-order
learning suggestions are expected to decline as additional labor cost savings will need to come

from improvements that alter or challenge the existing practices and the implicit wage-effort
bargain.
Since we expect the absolute number of suggestions to decline over time, we also expect a
decline in the absolute number of second-order learning suggestions. However, we expect the
proportion of this type of suggestion to increase over time relative to first-order learning
suggestions. The proportion of second-order suggestions is expected to be relatively low in the
period immediately following the introduction of the gainsharing plan as trust is built up within
the system and employees learn to think about work in new ways. The growth in the relative
proportion of second-order learning suggestions over time can be seen as the result of increased
individual knowledge based on communication and trust in the system as well as the desire to
maintain the gainsharing pay-outs once the gains from first order-learning suggestions have
declined.
Hypothesis 3: The relative proportion of employee suggestions representing first-order
learning will decline over time.
By focusing attention on the changing content of employee suggestions over time we are
utilizing a cognitive-oriented organizational learning perspective discussed in the previous
section. Although the act of suggestion-making is a behavior, the content of the suggestion is an
indicator of the knowledge of the individual suggestion-maker which, through the act of writing
it down and submitting it to a gainsharing plan suggestion committee, becomes organizationallevel knowledge. Changes in organizational knowledge, as measured by changes in the content
of employee suggestions over time, can be seen as an important intermediate mechanism in an
organizational learning process (Pentland, 1992; Snyder and Cummings, 1998).

METHODS
Research Site
The research site for this study was a large manufacturing plant in an industrial
Midwestern metropolitan area. The plant manufactures heating and cooling systems for the auto
industry and employs approximately 1,500 workers. All of the direct hourly workers are
unionized, and all are members of work teams. A modified Scanlon-type gainsharing program
was negotiated in 1988 as part of a cooperative joint union-management effort at the plant to
respond to economic downturns and competitive changes in the auto parts industry during the
1970's and 1980's (e.g., Cutcher-Gershenfeld & McHugh, 1994). Interviews with labor and
management representatives indicated that industrial relations prior to 1988 followed a pattern
typical of the traditional industrial relations system described by Kochan, Katz, & McKersie
(1986). This system was characterized by a low-trust union-management relationship in which
management retained the right to make unilateral decisions regarding market strategy and
investments; employee participation was limited to union bargaining over wages, benefits, and
working conditions; and work organization was characterized by narrowly defined highly
structured jobs. One indication of this Tayloristic job design approach at the plant was the fact
that prior to 1988, the plant had 56 job classifications for production workers and 23
classifications for maintenance and skilled workers.
Both union and management saw gainsharing as a key to breaking from this pattern and
saving the plant from being shut down or sold, and both parties were involved extensively in the
development and implementation of the plan. For example, the plant put together a joint unionmanagement study team. This team benchmarked plans at other plants proposed a customized
gainsharing formula (or family of measures) that consisted of 10 performance measures. Each

of these performance measures was relevant to plant costs, had historical data with which to
form a baseline, and could be controlled by the workforce. Included in the formula were
measures of direct and indirect labor costs (including overtime) of bargaining unit members,
maintenance materials, perishable tools, scrap, rework, and supplies. The bonus pool was
determined by subtracting the actual expenses (labor and other costs) from a baseline of
allowable expenses (based on a two-year rolling average of earned labor dollars and historical
costs). This bonus pool was split equally between the company and all bargaining unit
employees. Employee payouts were distributed twice per year (June and January). Each work
area in the plant was kept apprised of their group's performance on the family of measurements
as well as the plants year-to-date performance through charts posted throughout the plant and
updated on a monthly basis.
In analyzing the impact of this plan, we chose to use the 48-month period from the
introduction of the gainsharing plan, in January 1989 through December 1992. We chose to
limit our study to this period because of a number of contextual changes occurred at the plant at
the end of 1992 that could potentially confound the analysis. First, a new collective bargaining
agreement that made several modifications to the original gainsharing agreement came into
effect at the end of 1992. In addition, at the end of 1992 a program that replaced the supervisor
with a bargaining unit team leader was fully implemented. Finally, the work at the plant
changed dramatically with the model changeover in the auto company they supplied in 1992/93.
Coinciding with this changeover was the start-up of a new facility at the plant which began
production of three new product lines in mid-1992, the call back of 104 laid-off employees and
creation of 32 new jobs in January, 1993, and large increases in employment in 1993 and 1994.
In contrast, the period from 1989 through 1992 represented a relatively stable period with no

significant interventions. The choice of this period thus allows us to focus on the hypothesized
gainsharing dynamics while minimizing the problems involved in trying to make inferences in
the context of multiple treatments. The length of time covered by this study (four-years with a
two-year average baseline for performance measures) compares favorably to other longitudinal
studies of gainsharing and other incentive pay plans (e.g., Bowie-McCoy et al., 1993; Gowen &
Jennings, 1990; Schuster, 1983).
Categorizing Employee Suggestions
Consistent with a Scanlon plan format, the plant introduced a formal procedure for
soliciting and processing employee suggestions as part of the gainsharing plan. An individual
employee submitted the suggestion on a standard form to a joint union-management review
team. The review team could accept, decline, or ask to investigate further. There was also a
provision for a review board, consisting of union and management leaders, which met quarterly
and made decisions on suggestions whose implementation costs exceeded $1,000, or on any
suggestion whose implementation could not be agreed upon by the review team.
Plant employees made a total of 495 suggestions during the four-year period of our study.
The average of approximately 10 suggestions per 100 employees per year is comparable to other
U.S. firms with suggestion programs. For example, a 1990 survey of 336 U.S. firms with
suggestion programs by the National Association of Suggestion Systems found an average of 11
suggestions per 100 employees per year (reported in Robinson & Schroeder, 1993). The number
of suggestions at this plant, however, is significantly below that reported in other gainsharing
cases found in the literature (e.g., Hatcher, Ross, and Collins, 1991). When questioned about
the relatively low number of suggestions at the plant, both union and management
representatives indicated that they did not believe that this reflected a low level of employee

support for the plan. Instead, they reported that many plant improvements were being made
outside of the formal suggestion system. Hatcher et al. (1991) found a similar phenomenon in
the gainsharing plans that they studied in which 38% of the employees not making formal
suggestions indicated that they decided to circumvent the gainsharing suggestion system.
Another reason for the relatively low number of formal suggestions at this plant was that no
additional individual rewards or formal recognition were given to suggestion-makers during the
period under study. Because, however, we chose to study a period in which we were able to
determine that no significant new products or processes were introduced in the plant, we can be
reasonably certain that our suggestion data do not exclude significant plant level changes
implemented during this period.
We obtained the written copies of 436 of the 495 suggestions reported in the plant
suggestion log. The missing suggestions had been lost or discarded. Using the plants
suggestion log we coded date submitted, the department of origin of the suggestion, and the
action taken (i.e., accept, decline, or investigate further). We found no statistically significant
differences between the suggestions that we obtained and the total number of suggestions
submitted at the plant. Approximately two-thirds of the suggestions were accepted--a figure
which remained constant throughout the four-year period under study.
To distinguish between different types of suggestions, each suggestion was content
analyzed by the authors. Definitions and examples of these different suggestion types are listed
in Table 1.
Table 1 about here
As shown in this table, we categorized a suggestion as first-order learning if the content of that
suggestion consisted of ideas for: 1) saving on the costs of materials for existing operations; or 2)

doing currently contracted work more cheaply "in-house. We reasoned that these two types of
suggestions were consistent with the concept of first-order learning in that they did not challenge
the status quo thinking in the plant or change the basic way in which work was performed. For
example, suggestions for savings on the costs of materials such as changing suppliers to obtain
lower cost materials or re-using materials that have previously been scrapped typically did not
attempt to alter the plants wage-effort bargain. Employees in this case were not offering to
work harder or even more effectively to increase output. Rather, they were suggesting changes
outside of the wage-effort bargain that involved primarily one-time savings on the costs of
materials. In effect, these suggestions reinforced (enacted) existing routines and behavioral
norms in the plant.
In contrast, we categorized suggestions as representing second-order learning if the content
of the suggestion consisted of ideas for increasing productivity (output per unit of labor) by
changing the way in which employees performed their tasks and / or by re-designing the product
or manufacturing process. We reasoned that these types of suggestions were qualitatively
different from the first-order learning suggestions in that there was a proposed alteration in the
wage-effort bargain between labor and management. This is not to say these suggestions
directly proposed production speed-ups or wage reductions. Under the collective bargaining
agreement, the gainsharing committees were prohibited from accepting any suggestions that
would alter the terms of the negotiated agreement. But the process and design change
suggestions did require that employees change their job tasks in ways that increased plant
efficiency. This can be seen in the process change examples in Table 1 in which employees
identified alterations in production steps and flow which, if implemented, would result in a
decrease in labor required to produce a given output.

It should be noted that many of these process and design change suggestions identified
relatively minor physical alterations to existing operations. Under a traditional behavioral
approach to organizational-level learning, one would not be likely to classify these as secondorder learning because they dont require radical changes in existing organizational structures
or routines. From a cognitive organizational learning perspective, however, the underlying
values represented by process and design change suggestions depart significantly from the
traditional labor-management relationship that existed at the plant. In a political environment in
which labor and management see each other as potential enemies, the idea that a union member
would identify and communicate to management ways in which labor content could be
minimized (however minor the actual change might be) can be seen as a radical in terms of the
thinking and values its represents.
In addition, many of the process change (second-order learning) suggestions, including the
examples shown in Table 1, indicate that employees were thinking outside of their current jobs
to consider ways in which improvements to the entire process (including those outside of their
immediate area) might be achieved. The assumptions and cognitive maps associated with this
type of suggestion differ significantly from traditional Tayloristic assumptions accepted by both
labor and management and embedded in detailed union contracts. These traditional values
proscribed narrowly defined tasks and roles for production employees who were not paid to
thinkespecially not beyond their immediate tasks.
At an aggregate level then, we argue that degree to which the content of suggestions
identify material savings and new work vs. process and design changes provides us with a good
approximation of level of change in employee mind frame and organizational values. This
change is from a more traditional adversarial labor-management relationship to a more

integrative relationship based on trust and mutual gain. This is precisely the kind of
transformation in thinking hypothesized by the early gainsharing proponents (e.g., Katz & Kahn,
1966; Frost et al., 1974). Using a cognitive-based organizational learning perspective, we have
identified these different types of suggestions as manifestations of first- and second-order
learning.
Measurement
A number of steps were taken to insure that the categories of suggestions we derived were
psychometrically reliable and valid. First, the two researchers provided independent ratings for
two random samples of 50 suggestions using the categories listed above. The coefficient kappa
measure of agreement between the researchers were statistically significant at .90 and .93
(p<.01). Discrepancies were discussed and resolved. We were unable to clearly decipher the
meaning of 18 of the suggestions and therefore did not include them in the analysis relating to
suggestion type. We found no significant differences between the excluded suggestions and the
total population of suggestions in terms of date submitted, the department of origin of the
suggestion, or the action taken on suggestion (i.e., accept, decline, or investigate further).
As an additional check of both the reliability and validity of our suggestion categories, we
asked a supervisor from the plant who was familiar with both plant operations and the suggestion
program to rate a stratified sample of 52 suggestions. These suggestions reflected the range and
frequency of categories listed in Table 1. The supervisor was presented with the text of the
actual suggestion and was asked to indicate whether it could be described as a material or
process (or both) suggestion as defined in Table 1 and whether the suggestion could also be rated
as New Work or Design Change. We considered there to be agreement between the
supervisors and our ratings if both sets of ratings would result in the suggestion being classified

as either first-order learning or second-order learning as described in Table 1. The coefficient
kappa (Cohen, 1960) for this comparison was statistically significant at .68 (p<.01). For the
purpose of hypothesis testing we summed the total number of suggestions as well as the number
of first- and second order suggestions for each month.
RESULTS
Comparing the average level of performance for the two years prior to gainsharing with
the average for the first four years of the plan, the plan appears to have been successful. Labor
costs (which represent over 70% of total costs) declined by 8.5 percent. Substantial declines
were also achieved in the use of overtime, perishable tools, rework, and supplies. The net
savings attributable to the gainsharing plan during this period totaled over $9 million. Each
eligible employee received a total of $4,442 in gainsharing payouts over the period, which
represents a 4.5% bonus for an employee earning the average rate in the plant. The number of
grievances per employee and the rate of absenteeism at the plant declined by over 50% and 20%
respectively.
Closer examination of these performance results, however, indicates that the bulk of these
gains occurred in the first two years of the plan. The net savings from the plan actually peaked
at the end of the second year, declined sharply at the end of the third year (with $4 million loss),
and finally recovered somewhat in the forth year. Corresponding payouts to employees followed
a similar pattern with over 80 percent of the total payouts to employees coming in the first two
years of the plan. Overall the plant recorded profits of $38.5 and $14.8 million in 1989 and
1990, but experienced losses of $7.9 and $4.9 million in 1991 and 1992. Only grievances and
absenteeism appeared to decline steadily over the four-year period.

The pattern of financial results from the plant is reflected in the total number of employee
suggestions submitted. Chart 1 presents changes in the average number of suggestions per
month for the four years under study.
Chart 1 about here
The data shown in this chart are consistent with the prediction in Hypothesis 1 regarding the
trend over time for total suggestions. The number of suggestions increased rapidly during the
first six months of the gainsharing program and after a short dip at the end of the second year
peaked at about 16 suggestions per month at the beginning of the third year. The average
number of monthly suggestions then declined steadily through the third and forth years,
averaging only about 5 suggestions per month in 1992. In addition, the chart indicates that the
number of suggestions appears to be positively associated with the amount of payout from the
gainsharing plan.
To test whether the observed trends and effects were statistically significant, we used
monthly data to analyze the following regression equation:
Suggestionst =a + b1(Time) - b2(Time2) + b3(Payoutt) + b4(Employmentt) + et.
In this equation, Suggestionst is the total number of employee suggestions each month. Time is a
trend variable indicating the number of months since the beginning of the plan. Payoutt equals
the amount of gainsharing bonus payout to employees for that six-month period, and
Employmentt is the number of employees on the payroll each month at the plant. The mean,
standard deviation, and correlations for all variables are presented in Table 2.
Table 2 about here
Because these are time-series data and thus potentially unsuitable for use with OLS regression, a
Box-Jenkins time-series modeling procedure (Box & Jenkins, 1976) was used to model the error

(or noise structure) in the regression equation (SAS/ ETS 6.0). Specifically, the autocorrelation
and partial autocorrelation residual plots were examined to determine the existence of nonstationarity, seasonality, autocorrelation, and moving average processes in the time series data
that would violate the assumptions of OLS regression analysis (McCleary & Hay, 1980; Ostom,
1990; SAS Institute, Inc., 1993). Examination of these plots indicated a stationary noise
structure and the absence of any significant seasonal, autocorrelation, or moving average
processes. In addition, the non-significant Durbin-Watson statistic (1.85) confirmed an absence
of autocorrelation and the appropriateness of using OLS regression with these data. As Ostrom
(1990: 28) states, if there is no indication of significant serial correlation, we can accept OLS
estimates without fearing a loss of efficiency or bias in the estimated variances.
The OLS regression results in Table 3 show that the Time and Time2 variables are both
statistically significant, but with opposite signs. This finding supports the hypothesized
curvilinear trend for total suggestions based on the Hawthorne effect and pent-up ideas
arguments. In addition, the significant positive relationship for the Payoutt variable, while
controlling for the independent effect of time, supports Hypotheses 2. According to these data, a
$100 increase (or decrease) semi-annual employee payout was associated with an increase (or
decrease) of six suggestions per month. Given that the average number of suggestions was just
over 9 per month, the effect of Payout size can be seen as fairly large in practical terms. The
overall model was significant (R2=.41; p<.001).
Table 3 about here
Results for first- and second-order learning suggestions
Hypothesis 3 stated there would be a decrease in the percentage of first-order suggestions
(and a corresponding increase in the percentage of second-order suggestions) over time. The

trend in the percentage of first-order suggestions is graphically displayed in Chart 2. This Chart
shows that as a proportion of the total, first-order suggestions decreased over time from over 60
percent of the total number of suggestions in the first months to less than 40 percent of the total
by the end of the period. It should be noted that because all suggestions used in the analysis
were categorized as either first- or second-order suggestions, the results for the proportion of
second-order suggestions are simply the inverse of the ones shown in Chart 2. To avoid
redundancy, only the results for one type of suggestion (first-order) are presented here.
Chart 2 about here
To test for the statistical significance of this observed downward linear trend for the
number of first-order suggestions submitted per month (and corresponding increasing linear
trend for second-order suggestions), we again used the trend variable, Time, defined as the
number of months since the beginning of the gainsharing plan (1-48 months) in the regression
model. We also included a variable measuring the total number of suggestions each month,
Total Suggestions, as a control variable.
As in the previous analysis, we began by modeling the noise structure of the regression
model (Box & Jenkins, 1976; SAS Institute, Inc., 1993). In this case, however, examination of
the autocorrelation and partial autocorrelation residual plots indicated the presence of significant
serial correlation and seasonality in the time series. Additional analysis revealed that a noise
model consisting of autoregressive lags at 2 and 9 months AR (2,9) was the most effective model
in controlling for serial correlation and seasonality in the data. Given this noise structure, we
used a method known as Yule-Walker, or the two-step full transform method (Harvey, 1981;
SAS Institute, Inc., 1993: 213-217) to estimate a regression model that included the two
autoregressive parameters identified above. This method employs generalized least squares

using O.L.S. residuals to estimate covariances across observations. These estimates are then
used to transform the data into a form that is suitable for use with O.L.S.2
The results of this new analysis in Table 4 show a significant negative coefficient for the
Time variable (p<.05) indicating that number of first-order suggestions are becoming a relatively
smaller portion of the total number of suggestions over time as hypothesized based on the
organizational learning model.
Table 4 about here
The size of the effect of Time appears small (only about one-third of a suggestion decrease per
year). Given the small number of suggestions each month, however, this represents about an 8
percent decline in the monthly average of first order suggestions in this organization each year.
The overall regression model is also statistically significant (R2=.81; p<.01) while the DurbinWatson statistic (1.90, n.s.) indicates the absence of significant autocorrelation in the
transformed regression model.
DISCUSSION
An important implication of these findings is a recognition of the various criteria that can
be used to judge the effectiveness of a gainsharing program (c.f., Leana & Florkowski, 1992). In
the case of the organization studied here, the number of employee suggestions first increased,
then decreased rapidly over time. Based on these data, one might conclude that the gainsharing
program had failed in that it had not been able to sustain a high level of employee involvement
over time. The correlation between the volume of suggestions and payout from the plan (a proxy
for increased productivity and cost-savings) is consistent with the position that gainsharing
produces a Hawthorne effect and that gains are unlikely to be sustainable over a longer term.
One explanation of this effect is that gainsharing provides a means and an incentive for

employees to express pent-up cost saving ideas on how to improve existing operations (Hammer,
1988). Once these ideas have been expressed and implemented, however, these programs tend to
lose effectiveness as performance, payouts, and suggestions all begin to decline (e.g., Goodman,
1980; Kaufman, 1992; McKersie, 1963).
An organizational learning perspective suggests a very different interpretation of the
results. Instead of viewing declining gainsharing payouts and suggestions as evidence of a failed
short-term program, this perspective posits that these are precisely the conditions that will lead to
the generation of second-order organizational learning and change in the organization. These
data also support this interpretation. By disaggregating the pattern of overall suggestion-making
into first and second-order learning suggestions, we found that the type of suggestions shifted
over time from a majority of first-order material savings and new work suggestions in the earlier
period to a majority of second-order learning production process changes by the end of the four
year period. The importance of managing both types of learning has been noted by
organizational learning scholars (e.g., March, 1991; Lant & Mezias, 1992). Thus, using an
organizational learning system perspective, we could judge this program a success.
The fact that the plant in this case experienced financial losses in the 3rd and 4th year of
the gainsharing plan appears to be incongruous with the observation that changes in the
proportion of second-order learning suggestions are associated with improved organizational
performance over time. One possible explanation for this incongruity is that a longer time frame
than was available for this study is needed to assess the impact of changes in patterns of firstand second-order learning on long-term organizational performance and survival. In the case of
this plant, it was able to regain profitability in the 5th and 6th year following the introduction of
gainsharing. Although an analysis of this period is beyond the scope of this study, the return to

profitability appeared to be a function of the plants ability to successfully implement the model
changeover in 1993 and to attract a significant amount of new production work to the plant.
Plant personnel attributed much of their success in both of these activities to their improved
work methods and to the cooperative labor-management climate they had developed. Patterns of
organizational learning associated with the introduction of gainsharing at the plant appear to be
at least indirectly related to both of these outcomes and suggest at least part of the explanation
for the plants longer-term success.
Any conclusions from this study must be tempered by the fact that data for this study
come from a single gainsharing plan in one organization. Although the basic structural features
of this plan are typical of other Scanlon-type gainsharing plans, the extent to which the results
found in this study generalize to other organizational contexts as well as other types of
gainsharing programs (i.e., Ruckers or Improshare) can not be assessed directly. As Lawler
(1988) notes, gainsharing can be seen as part of a larger, multi-phase organizational change
effort toward a more participative (high-involvement) system of management. We know very
little about how of variations in gainsharing formulas, participation structures and processes
impact organizational outcomes over time. It may be that the specific features of this
gainsharing plan made it more likely to observe a transition from first- to second-order learning
than other plans. Kims (1999) study of 211 organizations with gainsharing plans found that a
number of the characteristics of this plan were positively correlated with program survival,
including low reliance on consultants, union participation, employee approval of the plan, and
labor intensity.
On the other hand, it is possible that the average bonus level of 4.5% in this study was not
sufficient to motivate more active participation and suggestion making on the part of employees.

Although very little data exists on the impact of the size of the gainsharing pay-out, reports of
some successful plans indicate considerably higher pay-out as a percentage of base pay than
occurred in this plant (e.g., Gross and Duncan, 1998). There was, however, a considerable
variation in the amount of bonusranging from11.1% of base pay at the beginning of the second
year to 0% at the end of the third year. Both labor and management representatives indicated that
they recognized the importance of providing a reasonable bonusespecially at the beginning
of the plan when trust was being established. They also felt strongly, however, that any
adjustments to the payout formula be made jointly as part of the negotiated labor agreement and
not appear to done arbitrarily by management. Thus no adjustments to the bonus formula were
made during the period of this study. Finally, the average bonus of 4.5% in this plan appears
consistent with that of other gainsharing plans. A recent survey William M. Mercer, Inc. reports
that firms with gainsharing plans were paying bonuses of about 5 percent of base pay to their
non-exempt employees in 1997 (1998/99 Compensation Planning Survey; Reported in Gross &
Duncan, 1998: 49).
An obvious extension to this line of research would be to compare the results from this
study with data from other gainsharing plans and organizational contexts. For example, we
would hypothesize that plans with larger rewards would increase the volume of suggestions and
thus decrease the time required before the second-order learning suggestions become a majority
of the total suggestions. The logic is that there is a finite amount of savings that can be gained
from first-order learning suggestions. The higher the pay-out, the more incentive there is for
employees to find cost savings, and the more quickly the saving from first-order learning would
be used up. In contrast we would expect, all else equal, that more frequent introductions of
new production processes into the plant would increase the amount of time before second-order

learning suggestions became the dominant type of suggestion. The reasoning is that each new
production process creates a certain amount of excess material and product defects that provide
additional opportunities for cost savings through first-order learning suggestions.
Finally, the impact of variations in employee and union involvement on volume and types
of employee suggestions is of particular interest to both researchers and practitioners. As noted
above, the union in this case was significantly involved in both the development and
implementation of the gainsharing program. A number of studies have found the level of union
support and involvement to be a significant factor in determining the success of employee
involvement programs in general (Cooke, 1992; McMahan & Lawler, 1995) and gainsharing in
particular (Cooper et al., 1992; Kim and Voos, 1997). It is possible that this involvement is also
a key to explaining the patterns of suggestion-making found in this case. It may be that
employees willingness to submit second-order learning suggestions is conditional on employee
perceptions of gainsharing fairness, which is impacted by union involvement.
An additional limitation is that we have no direct measure of organizational changes that
resulted from suggestions over time. As noted above, our evidence relates more directly to
changes in knowledge and values rather than to changes in routines and structures. Indirect
evidence, however, of the source of gains from mature gainsharing plans comes from a recent
study by Kim (1996). Kim surveyed 269 establishments with gainsharing and found that
program age was negatively associated with perceived annual improvements in cost reduction
and labor productivity, but positively associated with perceived improvements in a firms
production process. Interestingly, Kim also found that use of employee involvement and
Scanlon Plans were significantly associated with improved quality and production processes but
unrelated to annual improvements in labor productivity and cost reduction. These results,

combined with the results from this study, suggest that the exclusive reliance on measures of
operational efficiency associated more with first-order learning may underestimate the impact
and potential contribution of gainsharing programs that focus on employee involvement and
organizational transformation.
Despite the limitations indicated above, the pattern of results found in this case point to a
number of implications and insights for both gainsharing and organizational learning research
and practice. A central question in gainsharing research concerns how these programs impact
performance (Bullock & Lawler, 1984; Hammer, 1988; Lawler, 1988). Results from this study
suggest that the answer to this question may change over the life of the program--gains stemming
from the elimination of waste in the early stages of the program appear to give way to process
and design improvements as the program matures. In addition, the finding that second-order
learning suggestions become dominant as the gainsharing program matures may help to explain
the long term success of a number of well-known gainsharing plans which have been cited in the
gainsharing literature (e.g., Davenport, 1950; Graham-Moore, 1983).
Another implication of this analysis concerns the role that power and interest group
(coalitional) politics play in the organizational learning process. Although integral to some of
the early discussions of organizational search processes and learning (Cyert & March, 1963, pp.
120-125; March, 1962), the effect of organizational power and interest group politics has
generally not been developed in more recent organizational learning literature. Yet, as Miner &
Mezias (1996, p.97) point out, the process and outcomes of organizational learning are rarely
politically neutral. In this study, we observed that the shift from first- to second-order learning
implied a re-negotiation of the wage-and-effort bargain between employees and employer. This
re-negotiation is embedded in the power dynamics between labor and management stakeholders

(Collins, 1995). Thus, in addition to cognitive and social barriers to engaging in second order
learning, employees and organizations must also grapple with the political ramifications of
engaging in these activities (Argyris & Schn, 1996; Levitt & March, 1988).
The recognition of the importance of these power and interest group issues in the
organizational learning process calls attention to the impact of perceptions of procedural and
distributive justice. These topics have recently emerged in the gainsharing literature as
important determinants of gainsharing success (Cooper et al, 1992; Welborne, 1998; Welborne et
al., 1995). From an organizational learning perspective, it may be that the willingness of
organizational members to translate individual learning into a form that would become the
property of the organization (such as employee suggestions) is conditioned in part on members
justice perceptions. This may be particularly true in the case of second-order learning, which
may de-stabilize the power balance between interest groups.
CONCLUSIONS
This study addresses the call for theory-based studies of how gainsharing works
(Hammer, 1988; Lawler, 1988). It has been shown that it is possible to distinguish between two
types of employee suggestions resulting from a gainsharing program which represent two
different types of learning in organizations. The data from this organization suggests that,
consistent with the pent-up ideas explanation for gainsharing effectiveness, the first-order
learning suggestions increase initially after gainsharing is implemented. However, there appears
to be a limit to the ability of these types of suggestions to generate continued gainsharing
payouts and, as a proportion of total suggestions, these suggestions decline over time. In
contrast, the proportion of second-order learning suggestions begin at a lower level, then become
a relatively larger proportion of total employee suggestions over time. These findings are seen

as consistent with findings in the organizational learning literature which suggest that, although
more difficult to achieve, second order learning provides a means for organizations to sustain
superior performance over time.
The observed transition in the proportion of first- and second-order learning suggestions
in this case may help to explain the longevity and sustained performance of some well-known
gainsharing programs such as Adamson Co., Lincoln Electric, DeSoto, and Nucor Steel. This
transition is also consistent with the position held by early proponents of the Scanlon Plan who
viewed the Plans main contribution as helping to transform the way employees perceived their
work and their role in the organization. As Frost et al. (1974: 91-92) observed: In the
application of the Scanlon Plan, the job situation would be opened up so that the worker would
be stimulated to learn by asking why, how, when, what, and who. The procedure invites the
employee...to be open in exploring and experimenting, to risk and fail and not be punished. The
Scanlon Plan philosophy and format provide one method of initiating, developing, and
reinforcing learning throughout the employees vocational career.

FOOTNOTES
1
We gratefully acknowledge an anonymous reviewer for pointing this out to us.
The results using this method are essentially identical to the results of the conditional least
squares estimates produced using the ARIMA procedure (SAS Institute, 1993). Details of
results of these analyses are available upon request from the first author.
Page 36
REFERENCES
Alexander Consulting Group. 1992. Heath care costs, quality top list of human resource
concerns. Employee Benefit Plan Review, 47: 38-39.
Ashby, W. 1960. Design for a brain; the origin of adaptive behavior. New York, Wiley
Argyris, C., & Schn, D.A. 1978. Organizational learning. Reading, MA: Addison-Wesley.
Argyis, C., & Schn, D.A. 1996. Organizational learning II: Theory method and practice.
Reading, MA: Addison-Wesley.
Barley, S. 1986. Technology as an occasion for structuring: Evidence from observations of CT
scanners and the social order of radiology departments. Administrative Science Quarterly,
31: 78, 31.
Bowie-McCoy, S.W., Wendt, A.C., & Chope, R. 1993. Gainsharing in public accounting:
Working harder and smarter. Industrial Relations, 32: 432-445.
Bullock, R.J., & Lawler. E.E. 1984. Gainsharing: A few questions and fewer answers. Human
Resource Management, 23: 23-40.
Chamberlain, J. 1946. Every man a capitalist. Life, December 23: 93-94.
Cohen, J. 1960. A coefficient of agreement for nominal scales. Educational and Psychological
Measurement, 20: 37-46.
Collins, D., Hatcher, L, & Ross, T.L. 1993. The decision to implement gainsharing: The role of
work climate, expected outcomes, and union status. Personnel Psychology, 46: 77-104.
Collins, D. 1995. Self-interests and group-interests in employee involvement programs: A case
study. Journal of Labor Research, 16: 57-79.
Cooke, W.N. 1989. Improving productivity and quality through collaboration. Industrial
Relations, 28: 299-319.
Cooke, W.N. 1992. Product quality improvement through employee participation: The effects of
unionization and joint union-management administration. Industrial and Labor Relations
Review, 46: 119-134.
Cooper, C.L., Dyck, B., & Frohlich, N. 1992. Improving the effectiveness of gainsharing: The
role of fairness and participation. Administrative Science Quarterly, 37: 471-490.
Cutcher-Gershenfeld, J., & McHugh, P.P. 1994. Competition and divergence: Collective
bargaining in the North American auto supply industry. In P.Voos (Ed.), Contemporary
collective bargaining in the private sector: 225-258. Madison, WI: IRRA.
Page 37
Cyert, R.M., & March, J.G. 1963. A behavioral theory of the firm. Englewood Cliffs, NJ:
Prentice-Hall.
Davenport, R.W. 1950. Enterprise for everyman. Fortune (January): 50-58.
Duncan, R, & Weiss, A. 1979. Organizational learning: Implications for organizational design.
In B.M. Staw (Ed.), Research in organizational behavior, 1: 75-123. Greenwich, CT: JAI
Press.
Easterby-Smith, M. 1997. Disciplines of organizational learning: Contributions and critiques.
Human Relations; 50: 1085-1113.
Fiol, C.M., & Lyles, M.A.1985. Organizational learning. Academy of Management Review, 10:
803-813.
Frost, C.F., Wakeley, J.H., & Ruh, R.H.1974. The Scanlon plan for organization development:
identity, participation, and equity. East Lansing, MI: Michigan State University Press.
Garwin, D.A. 1993. Building a learning organization. Harvard Business Review (July-August):
73-91.
Geare, A.J. 1976. Productivity from Scanlon-type plans. Academy of Management Review, 1:
99-108.
Gerhart, B., & Milkovich, G.T. 1992. Employee compensation: Research and practice. In M.D.
Dunnette & L. Hough (Eds.), Handbook of industrial and organizational psychology, 2nd
edition: 481-569. Palo Alto, CA: Consulting Psychologists Press.
Goodman, P.S. 1980. Realities of improving the quality of work life. Labor Law Journal, 31:
487-494.
Goodman, P.S., & Moore, B.E.1976. Factors affecting acquisition of beliefs about a new reward
system. Human Relations, 29: 571-588.
Gowen, C.R., & Jennings, S.A.1991. The effects of changes in participation and group size on
gainsharing success: A case study. Journal of Organizational Behavior Management, 11:
147-169.
Graham-Moore, B.E. 1983. Ten years of experience with Scanlon plan: DeSoto revisited. In B.E.
Graham-Moore & T.L. Ross (Eds.), Productivity gainsharing: 62-88. Englewood Cliffs,
NJ: Prentice Hall.
Gray, R.B. 1971. The Scanlon plan: A case study. British Journal of Industrial Relations, 9: 291313.
Page 38
Griffin, R.W. 1988. Consequences of quality circles in an industrial setting: A longitudinal

assessment. Academy of Management Journal, 31: 338-358.
Gross, S. E. & Duncan, D. 1998. Gainsharing plan spurs record productivity and payouts at
AmeriSteel. Compensation and Benefits Review, (November December): 46-50.
Hammer, T.H. 1988. New developments in profit sharing, gainsharing, and employee ownership.
In J.P. Campbell, & R.J. Campbell (Eds.), Productivity in organizations: New perspectives
from industrial and organizational psychology: 328-368. San Francisco: Jossey-Bass.
Hanlon, S.C., & Taylor, R.R. 1991. An examination of changes in work group communication
behaviors following installation of a gainsharing plan. Group and Organization Studies, 16:
238-267.
Hanlon, S.C., Meyer, D.G., & Taylor, R.R.. 1994. Consequences of gainsharing. Group and
Organization Management, 19: 87-111.
Harvey, A.C. 1981. The econometric analysis of time series. New York: John Wiley.
Hatcher, L., & Ross, T.L. 1991. From individual incentives to organization-wide gainsharing
plan: Effects on teamwork and product quality. Journal of Organizational Behavior, 12:
169-183.
Hatcher, L., Ross, T.L., & Collins, D. 1991. Attributions for participation and nonparticipation in
gainsharing plan involvement systems. Group and Organization Studies, 16: 25-43.
Hayes, R.H., Wheelwright, S.C. & Clark, K.B. 1988. Dynamic manufacturing: Creating the
learning organization. New York: Free Press.
Hedberg, B.L. 1981. How organizations learn and unlearn. In P.C.Nystron & W.H. Starbuck
(Eds.), Handbook of organizational design. Volume1. New York: Oxford University Press.
Hedberg, B.L., Nystrom, P.C., & Starbuck, W.H. 1976. Camping on seesaws: Prescription for a
self-designing organization. Administrative Science Quarterly, 21: 41-65.
Huber, G.P. 1991. Organizational learning: The contributing processes and the literatures.
Organizational Science, 2: 88-115.
Katz, D., & Kahn, R.L. 1966. The social psychology of organizations. New York: John Wiley.
Kaufman, R.T. 1992. The effects of improshare on productivity. Industrial and Labor Relations
Review, 45: 311-322.
Kim, D. 1993. The link between individual and organizational learning. Sloan Management
Review, 35: 37-50.
Page 39
Kim, D. 1996. Factors influencing organizational performance in gainsharing programs.

Industrial Relations, 35: 227-244.
Kim, D. 1999. Determinants of the survival of gainsharing programs. Industrial and Labor
Relations Review, 53: 21-42.
Kim, D. and Voos, P.B. 1997. Unionization, union involvement, and the performance of
gainsharing programs. Relations Industrielles (Canadian) 52: 304-333.
Kochan, T.A., Katz, H.C., & McKersie, R.B. 1986. The transformation of American industrial
relations. New York: Basic Books.
Lant, T.K., & Mezias, S.J. 1993. Organizational learning model of convergence. Organizational
Science, 3: 47-71.
Lawler, E.E. III. 1988. Gainsharing theory and research: findings and future directions. In W.A.
Pasmore & R.W. Woodman (Eds.), Research in Organizational Change and Development:
323-344. Greenwich, CT: JAI Press.
Lawler, E.E. III, & Cohen, S.G. 1992. Designing a pay system for teams. American
Compensation Association Journal, 1: 6-19.
Leana, C.R., & Florkowski, G.W. 1992. Employee involvement programs: Integrating
psychological theory and management practice. Research in Personnel and Human
Resource Management, 10: 233-270. New York: JAI Press.
Leroy, F. & Ramanantsoa, B. 1997. The cognitive and behavioral dimensions of organizational
learning in a merger: an empirical study. Journal of Management Studies, 34: 871-895.
Levitt, B. & March, J. 1988. Organizational learning. Annual Review of Sociology, 14: 319340.
Levitt, B. & March, J. 1990. Chester I. Barnard and the intelligence of learning. In O. E.
Williamson (Ed.), Organization Theory: 11-37. New York: Oxford University Press.
Lundberg, C.C. 1995. Learning in and by organizations: Three conceptual issues. International
Journal of Organizational Analysis, 3: 10-23.
March, J.G. 1962. The business firm as a political coalition. Journal of Politics, 24: 662-678.
March, J.G. 1991. Exploration and exploitation in organizational learning. Organizational
Science, 2: 71-87.
March, J.G., & Simon, H.A. 1958. Organizations. New York: John Wiley.
Mawhinney, T.C. and C.R. Gowen. 1991. Gainsharing and the law of effect as the matching law:
A theoretical framework. Journal of Organizational Behavior Management, 11: 61-75.
Page 40
McGregor, D. 1960. The human side of enterprise. New York: McGraw-Hill.

McKersie, R.B. 1963. Wage payment methods of the future. British Journal of Industrial
Relations, 1: 191-212.
McMahan, G.C. & Lawler, E.E. 1995. Effects of union status on employee involvement:
Diffusion and effectiveness. Research in Organizational Change and Development, 8: 4776.
Mezias, S. & Glynn, M. 1993. The three faces of corporate renewal: Institution, revolution, and
evolution. Strategic Management Journal, 14, 77-101.
Miner, A.S., & Mezias, S.J. 1996. Ugly duckling no more: Pasts and futures of organizational
learning research. Organizational Science, 7: 88-99.
Pentland, B.T. 1992. Organizing moves in software support hotlines. Administrative Science
Quarterly, 37: 527-548.
Robinson & Schroeder. 1993. Training, continuous improvement, and human relations: The U.S.
TWI programs and the Japanese management style. California Management Review
(Winter): 35-57.
Rosenberg, R.D., & Rosenstein, E. 1980. Participation and productivity: An empirical study.
Industrial and Labor Relations Review, 33: 355-367.
Ross, T.L., & Ross, R.A. 1984. Productivity gainsharing: Resolving some of the measurement
issues. National Productivity Review: 382-394.
SAS Institute, Inc. 1993. SAS/ETS users guide, version 6 (2nd ed). Cary, NC: SAS Institute, Inc.
Scanlon, J.N. 1948. Profit sharing under collective bargaining: Three case studies. Industrial and
Labor Relations Review, 2: 58-75.
Schuster, M. 1983. The impact of union-management cooperation on productivity and
employment. Industrial and Labor Relations Review, 36: 415-430.
Schuster, M. 1984. The Scanlon plan: A longitudinal analysis. The Journal of Applied
Behavioral Science, 20: 23-38.
Schuster, M. 1985. Models of cooperation and change in union settings. Industrial Relations, 24:
382-394.
Senge, P.M. 1990. The fifth discipline: The art and practice of the learning organization. New
York: Doubleday.
Page 41
Snyder, W.M. & Cummings, T.G. 1998. Organizational learning disorders: Conceptual model
and intervention hypotheses. Human Relations, 51: 873-895.
Tsang, E.W.K. 1997. Organizational learning and the learning organization: A dichotomy
between descriptive and prescriptive research. Human Relations, 50: 73-89.
Vroom, V. 1964. Work and motivation. New York: Wiley.
Wagner, JA. III, Rubin, P.A., & Callahan, T.J. 1988. Incentive payment and nonmanagerial
productivity: An interrupted time series analysis of magnitude and trend. Organizational
Behavior and Human Decision Processes, 42: 47-74.
Welborne, T.M. 1998. Untangling procedural and distributive justice. Group & Organization
Management, 23: 325-339.
Welborne, T.M., & Gomez-Mejia, L.R. 1995. Gainsharing: A critical review and future research
agenda. Journal of Management, 21: 559-609.
Welborne, T.M., Balkin, D.B., & Gomez-Mejia, L.R. 1995. Gainsharing and mutual monitoring:
A combined agency-organizational justice interpretation. Academy of Management
Journal, 38: 881-899.
White, J. K. 1979. The Scanlon plan: Causes and correlates of success. Academy of Management
Journal, 22: 292-312.
page 42
TABLE 1
Categories of Gainsharing Suggestions
Suggestion Type
Description
Examples from Actual Suggestions
A. First-Order Suggestions
1. Material Savings
Savings come primarily from saving on the cost of Purchase tires for our internal carts from [specified] dealer. We are
materials at the plant
currently paying $130/tire, and [another specified dealer] will sell them for
$84.00.
2. New Worka
These suggestions justify creating additional work

for the plant employees due to lower costs in
house.
B. Second-Order Suggestions
Suggestions for changing the process sequence,
1. Process Changesb
including scheduling changes, set-up changes,
improved material handling, and changes to the
content of production tasks. Savings identified to
come primarily from improved labor efficiency
(reduced labor costs).
Regrind cut-off tool [part number] used in [specified] assembly area.

New cost [of tool] is $11.50. Regrind in house.
...the plates are stacked one on top of each other in wire baskets in the
punch press area. After plates are degreased and sent to the [next
assembly] area, an assembler removes them two at a time and flips one so
that the plate tabs face each other in pairs...my suggestion is to have the
punch press operator load the parts in the basket the way they will be used
in the [next assembly area]...
Use the Burr Oak machine in [area] 9215 to produce the header for the
new heater core rather than use the Cutoff Vail machine and new tube
cutter in [area] 9227. This process can be done on one machine and save
24 or more hours per shift
2. Design Changesc
Suggestions that would change the design of the

product or the materials used to manufacture the
product. Saving expected to come from improved
labor efficiency or improved product performance
(less returns and rework).
Replace screw used on housing to plastic linkage with a solder screw.

Presently operator drives the screw with a hand tool and then backs it
offThis operation is very time consumingand could affect quality.
_________________
Used in conjunction with Material savings category.

Suggestions that identified both Process and Material savings were included in this group.
c
Used in conjunction with both Material Savings, Process Change suggestion categories
b
page 43
TABLE 2
Descriptive Statistics and Correlation
Total
(n=48
months)
Average
per
Month
Standard
Deviation
1) Payout per
employee (paid
every 6 months)
$4,038
$511
491
----
2) Employment
Level
n/a
1484.4
206.4
.008
----
3) Time (Months)
48
N/A
14.0
-.262
-904*
----
4) Total Suggestions
436
9.08
6.22
.575*
.005
-.098
----
5) First-order
Learning
Suggestions
204
4.25
3.67
.531*
.084
-.182
.854*
----
6) Second-order
learning suggestions
214
4.45
3.56
.592*
-.098
.020
.863*
.490*
VARIABLE
* p<.01
----
page 44
TABLE 3
Results of Regression Analyses for the Total Number of Employee Suggestions
Over Time (n=48 months)
Variables
beta
Intercept
-37.15
Time
Time2
Std.
Error
19.89
0.90*
0.37
-0.12*
0.06
Payoutt
0.06**
0.02
Employmentt
0.02*
0.01
R2
Durbin-Watson
.41**
1.85
* p < .05;
** p < .01; two-tailed tests
page 45
TABLE 4
Results of Transformed Regression Analysisa Showing Decline in First Order
Learning Suggestions Over Time (n=48 months).
Variables
Std.
Error
beta
Intercept
.378
.499
Time
-.027*
.013
Total
Suggestions
.521**
.034
R2
Durbin-Watson
.81**
1.90
* p < .05
** p < .01; two-tailed tests.
a
Transformed using Yule-Walker method to control for 2nd and 9th order autoregressive
errors.
page 46
CHART 1:
Number of Suggestions by Month (6-month moving averages, 1/89-12/92)
18
16
12
10
8
6
4
2
Months
43
41
39
37
35
33
31
29
27
25
23
21
19
17
15
13
11
0
1
Number of Suggestions
14
page 47
CHART 2:
Percentage of First-order Learning Suggestions by Month (6-month moving
averages, 1/89-12/92)
0.7
0.6
0.4
0.3
0.2
0.1
Months
48
46
44
42
40
38
36
34
32
30
28
26
24
22
20
18
16
14
12
10
0
6
Percent
0.5
page 48
Jeffrey B. Arthur earned his Ph.D. at Cornell University. He is an assistant professor of

management at the R. B. Pamplin College of Business at Virginia Tech University. His current
research interests include understanding how the introduction of human resource management
practices lead to organizational performance outcomes.
Lynda Aiman-Smith is an assistant professor of business management at the College of
Management at NC State University. She received her Ph.D. in organizational behavior and
technology management from Purdue University. She researches implementation issues,
especially issues of bringing new technology, work systems, or new technical personnel into
organizations.

Arthur PDF

Uploaded by

Copyright:

Available Formats

Arthur PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Arthur PDF

Uploaded by

Copyright:

Available Formats

Academy of Management Journal - In Press

GAINSHARING AND ORGANIZATIONAL LEARNING: AN ANALYSIS

Academy of Management Journal - In Press

Academy of Management Journal - In Press

Academy of Management Journal - In Press

Academy of Management Journal - In Press

Academy of Management Journal - In Press

Academy of Management Journal - In Press

Academy of Management Journal - In Press

Academy of Management Journal - In Press

Academy of Management Journal - In Press

Academy of Management Journal - In Press

Academy of Management Journal - In Press

Academy of Management Journal - In Press

Academy of Management Journal - In Press

Academy of Management Journal - In Press

Academy of Management Journal - In Press

Academy of Management Journal - In Press

Academy of Management Journal - In Press

Academy of Management Journal - In Press

Academy of Management Journal - In Press

Academy of Management Journal - In Press

Academy of Management Journal - In Press

Academy of Management Journal - In Press

Academy of Management Journal - In Press

Academy of Management Journal - In Press

Academy of Management Journal - In Press

Academy of Management Journal - In Press

Academy of Management Journal - In Press

Academy of Management Journal - In Press

Academy of Management Journal - In Press

Academy of Management Journal - In Press

Academy of Management Journal - In Press

Academy of Management Journal - In Press

Academy of Management Journal - In Press

Academy of Management Journal - In Press

Academy of Management Journal - In Press

We gratefully acknowledge an anonymous reviewer for pointing this out to us.

Griffin, R.W. 1988. Consequences of quality circles in an industrial setting: A longitudinal

Kim, D. 1996. Factors influencing organizational performance in gainsharing programs.

McGregor, D. 1960. The human side of enterprise. New York: McGraw-Hill.

These suggestions justify creating additional work

Regrind cut-off tool [part number] used in [specified] assembly area.

Suggestions that would change the design of the

Replace screw used on housing to plastic linkage with a solder screw.

Used in conjunction with Material savings category.

Jeffrey B. Arthur earned his Ph.D. at Cornell University. He is an assistant professor of

You might also like