8/10/2019 The Influence of General Perceptions Of
1/26
http://jom.sagepub.com/Journal of Management
http://jom.sagepub.com/content/21/1/1
The online version of this article can be found at:
DOI: 10.1177/0149206395021001011995 21: 1Journal of Management
KudischJeffrey D. Facteau, Gregory H. Dobbins, Joyce E.A. Russell, Robert T. Ladd and Jeffrey D.
Motivation and Perceived Training TransferThe influence of General Perceptions of the Training Environment on Pretraining
Published by:
http://www.sagepublications.com
On behalf of:
Southern Management Association
can be found at:Journal of ManagementAdditional services and information for
http://jom.sagepub.com/cgi/alertsEmail Alerts:
http://jom.sagepub.com/subscriptionsSubscriptions:
http://www.sagepub.com/journalsReprints.navReprints:
http://www.sagepub.com/journalsPermissions.navPermissions:
http://jom.sagepub.com/content/21/1/1.refs.htmlCitations:
What is This?
- Feb 1, 1995Version of Record>>
at SAGE Publications on March 27, 2014jom.sagepub.comDownloaded from at SAGE Publications on March 27, 2014jom.sagepub.comDownloaded from
http://jom.sagepub.com/http://jom.sagepub.com/http://jom.sagepub.com/http://jom.sagepub.com/content/21/1/1http://jom.sagepub.com/content/21/1/1http://jom.sagepub.com/cgi/alertshttp://jom.sagepub.com/subscriptionshttp://www.sagepub.com/journalsReprints.navhttp://www.sagepub.com/journalsReprints.navhttp://jom.sagepub.com/content/21/1/1.refs.htmlhttp://jom.sagepub.com/cgi/alertshttp://jom.sagepub.com/cgi/alertshttp://jom.sagepub.com/content/21/1/1.full.pdfhttp://jom.sagepub.com/subscriptionshttp://jom.sagepub.com/content/21/1/1.full.pdfhttp://www.sagepub.com/journalsReprints.navhttp://online.sagepub.com/site/sphelp/vorhelp.xhtmlhttp://www.sagepub.com/journalsReprints.navhttp://www.sagepub.com/journalsReprints.navhttp://www.sagepub.com/journalsPermissions.navhttp://jom.sagepub.com/content/21/1/1.refs.htmlhttp://jom.sagepub.com/content/21/1/1.refs.htmlhttp://online.sagepub.com/site/sphelp/vorhelp.xhtmlhttp://online.sagepub.com/site/sphelp/vorhelp.xhtmlhttp://jom.sagepub.com/content/21/1/1.full.pdfhttp://jom.sagepub.com/content/21/1/1.full.pdfhttp://jom.sagepub.com/http://jom.sagepub.com/http://jom.sagepub.com/http://jom.sagepub.com/http://jom.sagepub.com/http://jom.sagepub.com/http://jom.sagepub.com/http://jom.sagepub.com/http://online.sagepub.com/site/sphelp/vorhelp.xhtmlhttp://online.sagepub.com/site/sphelp/vorhelp.xhtmlhttp://jom.sagepub.com/content/21/1/1.full.pdfhttp://jom.sagepub.com/content/21/1/1.full.pdfhttp://jom.sagepub.com/content/21/1/1.refs.htmlhttp://jom.sagepub.com/content/21/1/1.refs.htmlhttp://www.sagepub.com/journalsPermissions.navhttp://www.sagepub.com/journalsPermissions.navhttp://www.sagepub.com/journalsReprints.navhttp://www.sagepub.com/journalsReprints.navhttp://jom.sagepub.com/subscriptionshttp://jom.sagepub.com/subscriptionshttp://jom.sagepub.com/cgi/alertshttp://jom.sagepub.com/cgi/alertshttp://southernmanagement.org/http://southernmanagement.org/http://www.sagepublications.com/http://jom.sagepub.com/content/21/1/1http://jom.sagepub.com/8/10/2019 The Influence of General Perceptions Of
2/26
Journal of Management
1995, Vol. 21, No. I, 1-25
The nf luence of General Percept ions of
the Train ing nvironment on Retrain ing
Motivat ion and Perceived Training Trans fer
Jeffrey D. Facteau
Gregory H. Dobbins
Joyce E. A. Russell
Robert T. Ladd
Jeffrey D. Kudisch
l e Un iversity of Tennessee
The present study was conducted to determine whether trainees
general beliefs about training affect
pretraining
motivation and transfer
of traini ng in a large-scale traini ng curr iculum. I n addition, the
inf luence of social support for traini ng from four organizational
constituents (top management, supervisors, peers, and subordinates)
and task constraints in the work environment on pretrain ing
motivation and training transfer were evaluated. Nine hundred six ty-
seven managers and supervisors completed a questionnaire that
assessed 14 constructs. Structural equations analysis with L ISREL VI I
indicated that the overall r eputation of traini ng, intr insic and
compliance incentives, organizational commitment, and three social
support variables (subordinate, supervisor, and top management
support) were predictive of pretraini ng motivation. I n additi on,
pretrain ing motivation and subordinate, peer, and supervisor support
were predictive of managers perceived train ing transfer. Thesef indings
suggests that previous theory and research (e.g., Noe, 1986; Noe &
Schmitt, 1986) serve as a useful heur istic for predicti ng the effects of
general beli efs about traini ng on training eff ectiveness. Implications of
the-findings for futur e research and practice are discussed.
Recently, it has been estimated that organizations spend as much as 200 billion
annually on workforce training (McKenna, 1990). In the 1970s, the average
training expenditure per employee was approximately 75 to 100. Today, some
large corporations (e.g., IBM) budget as much as 2,000 annually per employee
on formal education and training (Wexley & Latham, 1991).
Not only do organizations today rely extensively on training, there is
considerable reason to believe that the significance of training for organizations
will continue to increase in the future. Several projected changes in the U. S.
Direct all correspondence to: Jeffrey D. Facteau, The University of Tennessee, Department of Management,
408 Stokely Management Center, Knoxville, TN 379964545.
Copyright Q 1995 by J AI Press Inc. 0149 2063
I
at SAGE Publications on March 27, 2014jom.sagepub.comDownloaded from
http://jom.sagepub.com/http://jom.sagepub.com/http://jom.sagepub.com/http://jom.sagepub.com/8/10/2019 The Influence of General Perceptions Of
3/26
2
FACTEAU DOBBINS RUSSELL LADD AND KUDISCH
labor force and work place are expected to account for this increased
importance. First, it is estimated that the proportion of entry-level youth will
decline and that middle-aged workers will represent a greater percentage of the
U. S. workforce. Second, there will continue to be a decline in the manufacturing
sector of the economy with a corresponding increase in the high technology,
service, and information sectors. Also, rapid technological development will
place greater demands on workers. In addition, the increased emphasis on
improving quality of services and products due to global competition is expected
to continue to change the way work is accomplished (e.g., more self-managed
work teams). Taken together, these changes will result in greater competition
for qualified entry-level employees and a need for organizations to better utilize
available labor. As such, organizations will need to enhance their recruitment
efforts to attract qualified entry-level employees and continually train and
retrain their current employees (c.f. Goldstein & Associates, 1989; Johnston &
Packer, 1987; London, 1989). Recently, some organizations have begun to takes
these initiatives (Ford, Major & Seaton, 1991).
Training Effectiveness
Given the significance of formal training programs for organizational
effectiveness, it is imperative that organizations design and implement training
programs in the most effective manner, and that they understand the factors
that contribute to training effectiveness. This is especially important given the
finding that only ten percent of training expenditures have been shown to result
in behavioral changes back on the job (Georgenson, 1982). In addition, a
frequent criticism of training practice and research is its susceptibility to fads
and its generally atheoretical approach (c.f. Campbell, 1971). For example,
some practitioners have attempted to increase the effectiveness of training by
utilizing sophisticated training techniques without regard for their actual need.
Others have focused on arranging the best training environment (e.g., Wexley,
1984). Without a theoretical basis for studying these techniques and training
environments, however, practitioners and researchers are often at a loss to
explain why training programs are effective or to predict their effectiveness in
other settings or for other trainees (Baldwin & Ford, 1988; Noe, 1986; Wexley
& Baldwin, 1986). Recognizing this situation in the training and development
literature, Tannenbaum and Yukl called for a paradigm shift to research
designed to understand
. . .why, when, and for whom a particular type of
training is effective (1992, p. 433).
Recent attempts to understand the factors that influence training
effectiveness have suggested that pretraining motivation may influence
important training outcomes. For example, Baldwin, Magjuka and Loher
(1991) found that pretraining motivation was related to actual learning in a
training program designed to improve skills in conducting performance
appraisals and in providing feedback. Another study revealed that pretraining
motivation was related to performance in a classroom environment (Baldwin
& Karl, 1987). Finally, Mathieu, Tannenbaum, and Salas (1992) found that
JOURNAL OF MANAGEMENT VOL. 21 NO. I 1995
at SAGE Publications on March 27, 2014jom.sagepub.comDownloaded from
http://jom.sagepub.com/http://jom.sagepub.com/http://jom.sagepub.com/http://jom.sagepub.com/8/10/2019 The Influence of General Perceptions Of
4/26
PERCEPTIONS OF THE TRAINING ENVIRONMENT
3
pretraining motivation for a proofreading skills program predicted learning and
subsequent performance on a work sample test.
These studies indicate that pretraining motivation has an important
influence on the extent to which trainees actually learn the material presented
to them during a training program. Furthermore, the amount of learning that
occurs during training may influence other indicators of training effectiveness,
such as trainees behavioral changes on the job and other organizational criteria
(e.g., absenteeism, productivity) (c.f. Goldstein, 1993). Thus, because of its
relationship to these training outcomes, pretraining motivation appears to be
an important antecedent to training effectiveness.
While pretraining motivation may influence training effectiveness,
relatively little research has examined the factors which contribute to trainees
pretraining motivation. The purpose of the present study was to examine the
extent to which employees attitudes and beliefs about training influence their
pretraining motivation, and their perceptions of the extent to which they are
able to transfer training back to their job. Several predictors of these variables
were examined in the study, including training attitudes, individual attitudes,
and social support for learning and transfer.
Training Attitudes
Perceived training reputation.
Prior to actually taking a training course,
an employee often has an expectation about the quality of the course and its
job relevance. Such expectations may be based upon past experiences with a
specific training program or may actually come from comments made by
coworkers who have already completed the training. If training is perceived as
a waste of time, employees may lack pretraining motivation irrespective of the
actual quality of the training program. In other words, the reputation of a
training program or training department should affect an employees pretraining
motivation. While other researchers have acknowledged the importance of
developing an understanding of factors that affect training motivation (e.g.,
Noe, 1986; Tannenbaum, Mathieu, Salas & Cannon-Bowers, 1991), research
has not examined the influence of the general reputation of training on
pretraining motivation. The first purpose of the present study was to test the
prediction that training reputation affects pretraining motivation.
Training incentives.
Training motivation should be affected by the extent
to which training is rewarded in the organization. Noes (1986) model of training
effectiveness proposed that rewards resulting from successful completion of
training influence individuals motivation to attend training and to learn from
it. In the present study, we assessed three types of training incentives: intrinsic
incentives (the extent to which training meets internal needs or provides
employees with growth opportunities), extrinsic incentives (the extent to which
training results in tangible external rewards such as promotions, pay raises, and
higher performance evaluations), and compliance (the extent to which training
is taken because it is mandated by the organization).
Based upon the work by Noe (1986), we expected that intrinsic incentives
and extrinsic incentives would be positively associated with pretraining
JOURNAL OF MANAGEMENT VOL. 21 NO. I 1995
at SAGE Publications on March 27, 2014jom.sagepub.comDownloaded from
http://jom.sagepub.com/http://jom.sagepub.com/http://jom.sagepub.com/http://jom.sagepub.com/8/10/2019 The Influence of General Perceptions Of
5/26
8/10/2019 The Influence of General Perceptions Of
6/26
PERCEPTIONS OF THE TRAINING ENVIRONMENT
5
Support for Learni ng and Trai ni ng Transfer
No (1986)
model proposed that environmental favorability should
affect pretraining motivation and transfer of training skills. Environmental
favorability is comprised of task constraints (e.g., lack of resources) and the
perceived social support for training. The social component is concerned with
employees perceptions of the extent to which the social context at work
supports training transfer. According to Noe, a supportive social context is
one in which employees believe that others provide them with opportunities
and reinforcement for practicing skills or for using knowledge acquired in
training.
Rouillier and Goldstein (1990) (cited in Goldstein, 1991) conceptualized
transfer climate in a similar manner by suggesting that it consists of two
components. The first, situational cues, refers to the extent to which aspects of
a situation encourage employees to use what has been learned in training (e.g.,
goal-setting, reducing employees workload so that they may become familiar with
new skills). The second, consequences, refers to the degree to which employees
are rewarded for applying what has been learned in training. These two views
(i.e., Noe, 1986; Rouillier & Goldstein, 1990) suggest that the social context at
work might support training in two ways, either by eliciting trained skills via a
host of antecedent variables (e.g., opportunities, situational cues), or by rewarding
these behaviors via a host of consequent variables (i.e., reinforcement).
In addition to the type of social support that exists for transferring trained
behaviors (i.e., antecedents, consequences), a variety of different sources of
social support may exist. An examination of the training literature suggests that
there are several potentially important sources of social support for training.
These include top management, supervisors, peers, and subordinates (c.f.
Baldwin & Ford, 1988; Goldstein & Musicante, 1986; Noe, 1986; Noe &
Schmitt, 1986). It seems likely, however, that the various sources of social
support may have differential effects on important training outcomes. For
example, peer support may be more important for predicting training transfer
than is top management support. Hence, the present study will determine the
independent and possibly differential effects of subordinate support, peer
support, supervisor support, and top management support on training transfer.
The task component of environmental favorability refers to the extent to
which characteristics of the work environment (e.g., tools and equipment,
materials and supplies, financial resources, etc.) facilitate or constrain
employees ability to transfer the skills learned in training back to their job.
Mathieu et al. (1992) found that task constraints were marginally related to
training motivation. Other researchers (e.g., Noe & Schmitt, 1986; Williams et
al., 1991) have employed separate measures of social support and task
constraints, but collapsed the scales to form a single variable, which was labeled
environmental favorability. From a theoretical standpoint, it may be that task
constraints have a different effect on transfer than social constraints. For this
reason, measures of task constraints and the different facets of social support
were analyzed separately in the present study.
JOURNAL OF MANAGEMENT VOL. 21 NO. I 1995
at SAGE Publications on March 27, 2014jom.sagepub.comDownloaded from
http://jom.sagepub.com/http://jom.sagepub.com/http://jom.sagepub.com/http://jom.sagepub.com/8/10/2019 The Influence of General Perceptions Of
7/26
6 FACTEAU DOBBINS RUSSELL LADD AND KUDISCH
The social and task support measures may also affect pretraining
motivation. If employees believe that training will not transfer back to the
job due to either nonsupportive peers, subordinates, supervisors,
management, or a lack of resources (task constraints), they should be less
motivated to attend and learn from training. Hence, the present study
examined the effects of task constraints and the four social support variables
(subordinate, peer, supervisor and top management) on training motivation
and perceived training transfer.
Overview of the Study
In summary, the present study was conducted to determine the effects of
employees attitudes and beliefs about training in general on pretraining
motivation and perceived training transfer. Training motivation was the
primary criteria since recent research has shown that trainees who enter training
with high levels of pretraining motivation learn more and are more likely to
complete training than their less motivated peers (e.g., Baldwin et al., 1991;
Hicks & Klimoski, 1987). Figure 1 summarizes the predictions of the study.
While researchers have not recognized that the general reputation of training
may exert strong effects of training motivation, we expected such a relationship
in the present study. In addition, consistent with past research, we expected
incentives, career exploration and planning, and organizational commitment
to affect pretraining motivation. Furthermore, we differentiated the social
support construct into four predictors. Specifically, new scales were developed
for examining employees perceptions of different sources of social support (i.e.,
top management, supervisors, peers, subordinates) within the organization.
These scales allowed for an examination of the possibility that different sources
of social support have unique or differential effects on employees pretraining
motivation and training transfer.
The results of the study should have important theoretical and applied
significance. From the theoretical perspective, the findings may serve to
support several links in Noes (1986) model of training effectiveness. In
addition, several additional variables (e.g., training reputation) were postulated
to predict training motivation. Finally, the study explored whether the various
social support factors differentially affect training motivation and training
transfer.
The study also has implications for organizational practices. For example,
if training reputation affects training motivation, then organizations should be
concerned about general impressions of available training programs and the
training department prior to actual training. Irrespective of the actual quality
of training, employees may not be motivated if they perceive training programs
as ineffective and irrelevant to theirjobs. Similarly, if the various support factors
have independent influences on pretraining motivation and training transfer,
then organizations may be able to target organizational development efforts
at the support factors that are most critical.
JOURNAL OF MANAGEMENT VOL. 21 NO. 1 1995
at SAGE Publications on March 27, 2014jom.sagepub.comDownloaded from
http://jom.sagepub.com/http://jom.sagepub.com/http://jom.sagepub.com/http://jom.sagepub.com/8/10/2019 The Influence of General Perceptions Of
8/26
PERCEPTIONS OF THE TRAINING ENVIRONMENT
Note:
The nature of the relationship expected among constructs is indicated by a -t or -.
Figure 1. Hypothesized Model
Methods
Subj ects
Subjects consisted of 967 managers and supervisors employed by a
southeastern state government. These subjects, 61% of whom were male, held
management positions in 37 different departments within state government (e.g.,
Education, Finance and Administration, Labor, etc.). Eighty seven percent of
the subjects were Caucasian and nine percent were African-American. The
remaining four percent either represented another racial group (e.g., Native
American, Asian) or did not report their racial background. Fifty one percent
of the sample held first-line supervisory positions, while 31% and 15% of the
sample held middle-level and upper-level management positions, respectively.
Subjects average tenure with state government was 16.34 years (SIX8.16) and
average job tenure was 6.51 years (SD=5.44). Subjects average age was 45.5
years (SDx8.7). Finally, subjects responses to an item asking them to report
their highest educational attainment were as follows: high school degree (150/o),
trade/technical degree (5%), some college (20 @1), ssociate degree (5%), four
JOURNAL OF MANAGEMENT VOL. 21 NO. 1 1995
at SAGE Publications on March 27, 2014jom.sagepub.comDownloaded from
http://jom.sagepub.com/http://jom.sagepub.com/http://jom.sagepub.com/http://jom.sagepub.com/8/10/2019 The Influence of General Perceptions Of
9/26
8 FACTEAU DOBBINS RUSSELL LADD AND KUDISCH
year college degree (23%), some graduate work (90/o), nd graduate degree (18%).
The remaining subjects (5%) did not respond to this item. The median
educational level was the completion of a four year college degree.
Procedure
The current study was conducted as part of a large scale training needs
assessment to identify the training needs and curricula for all levels of
management within state government. As part of the needs assessment, a
comprehensive survey was distributed to a random sample of 1,700 supervisors
and managers. The survey requested managers to provide information about
their perceived training needs, training preferences, and their attitudes about
training, the organization, and their job. Managers completed the surveys at
work and were assured of anonymity and confidentiality of their responses.
Completed surveys were returned directly to the researchers. Of those sampled,
1,089 managers returned usable surveys (64% response rate). Because the current
research focused upon managers attitudes and perceptions of the states
management training curriculum, only those who had taken one or more of
the courses in the curriculum were included in the analyses. Based upon this
criterion, 122 managers who had not yet taken any training in the curriculum
(because they were new to their position) were eliminated from the sample.
Hence, the results of this study were based on a total sample of 967 supervisors
and managers.
Measur es
In order to test the hypotheses of interest in this study, measures of fourteen
distinct theoretical constructs were included on the questionnaire. These
constructs were measured with a total of 85 items. The number of items
comprising each scale and a sample item are presented in Table 1. Responses
to all of the items on the questionnaire were made on five point Likert-type
scales (l=Strongly Disagree,
5=Strongly Agree). Unless indicated
otherwise, higher scores indicated greater levels of the variable measured.
Many of the constructs were assessed with scales that have been used in
previous research. Four items from Porter and Smiths (1970) scale were used
to assess
Organizational Commitment. Career Exploration was
measured with
scales from Stumpf, Colarelli and Hartmans (1983) career exploration survey.
Career PZanning was
assessed with a modified version of Goulds (1979) career
planning scale.
Pretraining Motivation
was assessed with items drawn from
several scales used in previous research (Baldwin & Karl, 1987; Hicks, 1984;
Noe & Schmitt, 1986). Finally,
Task Constraints
were assessed with a scale
based upon Peters and OConnors (1980) taxonomy of the situational factors
that may constrain individual performance.
Several constructs used were measured with scales that were designed
specifically for the current study. These include measures of training reputation,
training incentives, social support for training, and perceived training transfer.
Training Reputation.
It was expected that the overall reputation of state
supervisory and managerial training would influence managers pretraining
JOURNAL OF MANAGEMENT VOL. 21 NO. 1 1995
at SAGE Publications on March 27, 2014jom.sagepub.comDownloaded from
http://jom.sagepub.com/http://jom.sagepub.com/http://jom.sagepub.com/http://jom.sagepub.com/8/10/2019 The Influence of General Perceptions Of
10/26
PERCEPTIONS OF THE TRAINING ENVIRONMENT
Table 1. Description of Scales Used in the Study
9
Scale
Number of
I tems
Sample I tem
Pretraining Motivation
Perceived Training Transfer
Training Reputation 6
Intrinsic Incentives 9
Compliance
Extrinsic Incentives
Career Exploration
Career Planning
Organizational Commitment
Subordinate Support
Peer Support
Supervisor Support
Top Management Support
Task Constraints
9
9
10
5
10
I try to learn as much as I can from training
courses.
The productivity of my subordinates has
improved due to the skills that I learned in
training courses.
Most supervisory and managerial training
courses are of the highest quality.
I take training because it provides me with
skills that allow me to be more effective on
the job.
I take training because it is mandated by state
government.
I take training because it increases my job
opportunities outside of state government.
I have often tried specific work roles just to
assess my skills and abilities.
I have a strategy for achieving my career goals.
I find that my values and the organizations
values are very similar.
My subordinates allow me to get accustomed
to using my new training skills on the job.
My peers encourage my efforts to incorporate
new procedures that I have learned in
training.
My supervisor is tolerant of changes that I
initiate as a result of learning new skills in
training.
Top management believes in the importance of
training for supervisors and managers.
Inadequate financial resources hamper my abil-
ity to apply new skills learned in training
back to my job.
motivation. A six item scale was developed to assess training reputation. The
items asked subjects to rate: (1) the overall quality of supervisory and managerial
training courses; (2) whether they would recommend these courses to their peers;
and (3) the extent to which these courses developed skills necessary for success
as a supervisor or manager in state government.
Training incentives.
Trainees may perceive a number of reasons for
attending training programs (Nordhaug, 1989). Three constructs thought to be
particularly relevant in the context of this study included intrinsic incentives,
extrinsic incentives, and compliance.
Intrinsic Incentives
measured the extent
to which managers believed that a variety of intrinsic benefits or outcomes were
likely to result from successful completion of the training.
Extrinsic Incentives
measured the extent to which managers believed that training would result in
extrinsic rewards and benefits. Finally,
Compliance
assessed the extent to which
JOURNAL OF MANAGEMENT VOL. 21 NO. 1 1995
at SAGE Publications on March 27, 2014jom.sagepub.comDownloaded from
http://jom.sagepub.com/http://jom.sagepub.com/http://jom.sagepub.com/http://jom.sagepub.com/8/10/2019 The Influence of General Perceptions Of
11/26
10
FACTEAU DOBBINS RUSSELL LADD AND KUDISCH
managers attended supervisory and management training in order to satisfy
organizational mandates or requirements.
Social support for training.
Social support for training and transfer of
learned skills was viewed as a multidimensional construct. The extent to which
four different organizational groups were supportive of managers training
efforts was measured. These groups included the managerssubordinates, peers,
supervisors, and top management. Items were developed based upon a review
of relevant literature (Baumgartel & Jeanpierre, 1972; Baldwin & Ford, 1988;
Clark, 1990; Goldstein & Musicante, 1986; Noe, 1986). In general, these items
assessed the extent to which these sources: (1) provided opportunities for
managers to utilize trained skills; (2) were supportive of managers efforts to
apply trained skills back on the job (e.g., were tolerant of mistakes); and (3)
reinforced managers efforts to transfer skills to their job situation.
Perceived Training Transfer.
This scale measured the extent to which
managers believed that a variety of desirable outcomes (e.g., improved
performance, reduced turnover, etc.) have occurred as a result of their ability
to transfer the skills they have learned in supervisory and management training
back to the job. Items for this measure were based upon a review of the relevant
training and development literature (Noe, 1986; Tziner, Haccoun, & Kadish,
1991; Wexley & Baldwin, 1986).
The items in the transfer measure are presented in the Appendix. They
were constructed to maximize the validity of self-reports of training transfer.
For example, Dunning, Meyerowitz and Holzberg (1989) found that self-ratings
made on ambiguous scale anchors were more inflated and less accurate than
self-ratings made on specific scale anchors. Farh and Dobbins (1989) reported
similar findings. Thus, self-report transfer items were constructed to be as
specific as possible (e.g.,
subordinates have told me that my behavior has
improved following a training course). Similarly, in order to improve the
accuracy of self-reports (and minimize potential inflation), all responses were
anonymous and all completed questionnaires were returned directly to the
researchers.
Analyses
The current research utilized a latent variable model to test the
hypothesized relationships among the constructs of interest. An advantage of
latent variable models is that they involve tests of the relationships among
constructs which are not attenuated by measurement error (c.f. Loehlin, 1987).
Analysis proceeded in two stages (Lance, Cornwell & Mulaik, 1988). First,
confirmatory factor analysis (CFA) was used to assess the fit of the
measurement model to the data. Second, the relationships among the constructs
of interest were estimated (i.e., structural model). Both of these analyses were
conducted using the LISREL VII computer program (Joreskog & S&born,
1989a).
Our primary purpose in this study was to examine the magnitude of the
relationships among the constructs of interest (i.e., path coefficients). A
secondary purpose was to determine the extent to which the data, in general,
JOURNAL OF MANAGEMENT VOL. 21 NO. 1 1995
at SAGE Publications on March 27, 2014jom.sagepub.comDownloaded from
http://jom.sagepub.com/http://jom.sagepub.com/http://jom.sagepub.com/http://jom.sagepub.com/8/10/2019 The Influence of General Perceptions Of
12/26
PERCEPTIONS OF THE TRAINING ENVIRONMENT
II
conformed to our predictions. To do this, a variety of goodness-of-fit indices
were utilized. Three indices calculated by the LISREL VII program, including
chi-square, the goodness-of-fit index (GFI), and the root mean square residual
(RMSR) were used. These indices provide a measure of the extent to which
the covariance matrix estimated by the hypothesized model reproduces the
observed covariance matrix (James, Mulaik & Brett, 1982).
In addition to these indices, the normed-fit index (NFI; Bentler & Bonett,
1980), and the relative normed-fit index (RNFI; Mulaik, James, Van Alstine,
Bennett, Lind & Stilwell, 1989) were used because they have certain desirable
characteristics relative to the other fit indices. First, both NFI and RNFI assess
a models fit in relation to the worst and best fit attainable rather than as a
sole function of the difference between the reproduced and observed covariance
matrices. Second, RNFI was used because it assesses the relative lit of a
structural model independent of assessing the fit of the measurement model,
thus providing a purer indication of the fit of a structural model. This is desirable
because the other fit indices are heavily influenced by the fit of the measurement
model and, to a lesser degree, the fit of the structural model (Mulaik et al.,
1989). In general, a good fit to the data is indicated by non-significant values
of chi-square, RMSR values which approach zero, and values of GFI, NFI,
and RNFI which approach unity.
Results
M easurement M odel
Before considering the hypothesized relationships among the constructs
of interest, the accuracy of the measurement model with respect to the data
was examined. A confirmatory factor analysis of the a priori measurement
model was conducted. In this analysis, the variances for all of the factors were
fixed at unity and all of the constructs were allowed to intercorrelate.
Computation of NFI for this model required the estimation of two comparison
models. These included a most-restrictive null model in which no factors were
assumed to underlie the questionnaire items, and a least-restrictive model in
which each of the items was allowed to load on each of the hypothesized
constructs.
Table 2 presents the goodness of fit indices for the a priori measurement
model. The results reported in this table indicate that while this model provided
Table 2. Goodness-of-Fit Indices for the A Priori and Revised Measurement Models
Measurement
M odel Chi -Square
df
GFI
RMSR
NFI
A Priori
11,633.60 3,394
.I4 ,062
.86
Revised
8,294.88* 2,983
.79 .055
.92
Notes:
a GFI=goodness-of-fit index; RMSR=root mean square residual; NFI=normed-fit index.
* p < ,001
JOURNAL OF MANAGEMENT, VOL. 21, NO. 1, 1995
at SAGE Publications on March 27, 2014jom.sagepub.comDownloaded from
http://jom.sagepub.com/http://jom.sagepub.com/http://jom.sagepub.com/http://jom.sagepub.com/8/10/2019 The Influence of General Perceptions Of
13/26
T
e
3
I
n
e
c
e
a
o
A
S
u
V
a
e
V
a
e
1
2
3
4
5
6
7
8
9
1
I
1
1
1
E
V
a
e
1
P
e
a
n
n
M
o
v
o
1
o
2
P
c
v
T
a
n
n
T
a
e
.
4
1
0
E
V
a
e
3
T
a
n
n
R
a
o
.
6
.
7
1
o
4
n
n
c
I
n
v
.
7
.
5
.
6
1
0
5
C
m
a
-
4
-
2
-
3
-
5
1
0
6
E
n
c
I
n
v
.
3
.
3
.
3
.
4
.
0
1
o
7
C
e
E
o
a
o
.
1
.
1
.
1
.
1
-
0
.
3
1
0
8
C
e
P
a
n
.
2
.
2
.
2
.
2
-
1
.
1
.
2
1
0
9
O
g
C
m
m
.
4
.
3
.
3
.
3
-
2
.
1
.
0
.
3
1
0
1
S
d
n
e
S
.
2
.
5
.
4
.
4
-
1
.
3
.
1
.
1
.
3
1
0
1
P
S
.
3
.
5
.
4
.
4
-
1
.
3
.
1
.
2
.
4
.
7
1
0
1
S
v
s
o
y
S
,
2
.
3
.
3
.
3
-
.
1
.
0
.
0
.
4
.
5
.
5
1
0
1
T
M
g
S
.
2
.
3
.
3
.
3
-
1
.
1
.
o
.
1
.
5
.
4
.
6
.
6
1
0
1
T
k
C
a
n
s
.
o
-
0
-
1
.
o
.
0
.
0
.
2
.
0
-
2
-
O
-
1
-
2
-
2
1
0
N
o
p