06 brugeres forventninger

Abstract
The Role A field study of MBA students enrolled in a

of Work, Play, microcomputer software training class was con-
ducted. Trainees' expectations of the extent to
which the training would be like work and play

and Fun in were collected prior to the training and their per-
ceptions of the extent to which the training was
like work and play were collected after the train-
Microcomputer ing. In addition, trainees' perceptions of the
extent to which the training was fun were also
measured. Results indicated that play percep-
Software tions operated as a suppressor variable. Post-
training play perceptions had a marginally signifi-
cant and negative effect on learning and #7-
Training creased the positive and significant effect of
post-training fun perceptions on learning. Re-
suits also indicated that despite the fact that
Elissa L. Perry much of the training occurred on the computer,
University of Illinois, Urbana-Champaign trainers influenced trainees' perceptions that the
training seemed like work, play, and fun. Finally,
Deborah J. Ballou learning was negatively affected when trainees
with high pre-training play expectations per-
University of Notre Dame ceived trainers to have a high work orientation in
the training. These results suggest that play
and fun perceptions have potentially important
consequences for learning and that trainers play
an important role in influencing these percep-
tions in microcomputer training.

Keywords: Computer-based training, human-
computer interaction, work, play, fun, learning

ACM Categories: H.4.1, K3.1, K.81

Introduction

The use of microcomputer technology in
organizations is widespread (Ballou & Rush,
1996; Turnage, 1990). Training and retraining
demands increase with the implementation of
technology. In addition, some have suggested
that the primary causes for the failure of office
technologies (e.g., computer systems) are hu-
Acknowledgement man and organizational, including the lack of
employee training (Turnage, 1990). As a result,
The authors would like to thank Carol Kulik and interest in training individuals in the use of mi-
Joe Martocchio for their assistance and helpful crocomputer technology has risen (Allan, 1993;
comments at various stages of this research. Ballou & Rush, 1996; Geber, 1994; and Klein-
We would also like to thank Ruth Reingold, schrod, 1988). At the same time, a growing
Manager of System Services, for her cooper- body of research has developed which explores
ation in this research endeavor. microcomputer training issues in the workplace

The DATA BASE for Advances in Information Systems - - Spring 1997 (Vol. 28, No. 2) 93

(e.g., Gist, Schwoerer, & Rosen, 1989; Martoc- Glynn (1994) has suggested that "...research is
chio, 1992; Martocchio & Webster, 1992; and n e e d e d . . . o n how task interpretations may
Webster & Martocchio, 1993). occur spontaneously, even in the absence of
experimental cues that send the message that
Although the structural design of microcomputer 'this is work' or 'this is play.'"
software is likely to influence how easy it is to
learn, a recent stream of microcomputer training Both the social contextual and individual differ-
research suggests that an individual's approach ence approaches to work-versus-play orienta-
to a task as work-versus-play may also influ- tions suggest that play has multiple dimensions.
ence computer training effectiveness (Martoc- Martocchio and Webster (1992) suggest that
chio & Webster, 1992; Webster, Heian, & while the most relevant aspect of individual
Michelman, 1990; Webster & Martocchio, 1992; playfulness in human-computer interactions is
and Webster & Martocchio, 1993). This re- cognitive playfulness, other dimensions of play-
search suggests that a work-versus-play orien- fulness (e.g., manifest joy) may be relevant and
tation may be the result of individual attributes should be explored. The literature is less clear
(e.g., a playfulness trait) (Martocchio & Web- about what aspects of labeling a computer task
ster, 1992; Webster & Martocchio, 1992) or as "play" may contribute to its beneficial effects.
social contextual cues (e.g., labeling) (Webster, Some research suggests that one dimension of
1990; Webster & Martocchio, 1993). It has play is fun (Bamett, 1990; Glynn & Webster,
been argued that individuals demonstrate 1992).
greater creativity and develop skills through ex-
ploratory behaviors during playful interactions Other research suggests that play and fun are
with tasks. Therefore, those who approach a closely related but distinct concepts (e.g.,
task as play are expected to learn more than Abramis, 1990). Although "fun" is likely to be
those who approach a task in less playful terms associated with play more than work (Dan-
or who approach it as work. Consistent with dridge, 1986) this is not always the case (Burke,
this, research suggests that perceptions of com- 1971). Work can sometimes be fun and en-
puter training as play as opposed to work result joyable while play may not be. Currently, little
in more positive training outcomes. research has explored the extent to which com-
puter training tasks may be perceived as fun
The little research that has explored the influ- and the extent to which this perception is related
ence of labeling computer training tasks as work to perceptions of the task as work versus play.
versus play (Webster et al., 1990; Webster & It is also unclear how perceptions of a task (e.g.,
Martocchio, 1993) has for the most part done so computer training) as work, play, and fun influ-
in the context of an experimental research de- ence training outcomes relative to one another.
sign which explicitly manipulates task labels. As The current paper presents exploratory re-
a result, these research efforts have high inter- search that addresses four issues. First, this
nal validity and have thus been able to assess paper explores the extent to which individuals in
causal relationships. However, this research an actual computer training context spontane-
provides little infonTlation about how individuals ously perceive the training as work versus play
naturally and spontaneously perceive actual and the influence of these perceptions on
computer training contexts. It is also unclear to learning. A field study approach is adopted in
what extent actua~ trainers influence trainees' order to determine the extent to which previous
perceptions of a task as work versus play when research results which indicate that task label-
trainers are not explicitly instructed to label ing (work versus play) influences training out-
tasks. It is possible that individuals do not comes in a field experiment generalize to a non-
spontaneously perceive computer training in
experimental context.
work versus play terms in the course of an
actual training program. Further, even if indi- Second, this paper explores the extent to which
viduals do perceive computer training in work- individuals' work and play perceptions of the
versus-play terms, these perceptions may not training are influenced by the trainer. Although
influence learning in a training context that does other social contextual factors (e.g., classmates)
not explicitly and clearly prime these concepts may influence individuals' perceptions and train-
(e.g., using task labels). Consistent with this, ing outcomes, the focus in the current research

94 The DATA BASE for Advances in Information Systems - - Spring 1997 (Vol. 28, No. 2)

is on the role of the trainer for two reasons. ence work related (e.g., training) outcomes
First, previous experimental research suggests (e.g., Abramis, 1990; Cellar & Barrett, 1987;
that the social context (i.e., the trainer) can influ- Glynn, 1994; Sandelands, 1988; Tang & Bau-
ence the extent to which individuals perceive meister, 1984; Webster et al., 1990; and Web-
training as work versus play (e.g., Webster et ster & Martocchio, 1992, 1993). This is
al., 1990; Webster & Martocchio, 1993). It is hypothesized when a work-versus-play ap-
informative to assess the extent to which trainer proach is a function of individual attributes (e.g.,
effects also operate in a non-experimental set- Glynn & Webster, 1992; Webster & Martocchio,
ting. Second, trainers constitute an element of 1992), contextual factors (Abramis, 1990; Glynn,
the social context over which training designers 1994; Sandelands, 1988; Webster et al., 1990;
have relatively more control. As a result, it is and Webster & Martocchio, 1993), or both (e.g.,
particularly important to focus on this aspect of Tang & Baumeister, 1984). Research has
the training context. The current paper assess- generally hypothesized and found positive ef-
es the extent to which actual trainers influence fects for individuals who approach a task as
work, play, and fun perceptions in a context play (Abramis, 1990; Cellar & Barrett, 1987;
where they are left to their own devices and are Csikszentmihalyi, 1975; Martocchio & Webster,
not instructed to label the training in any 1992; Miller, 1973; Webster et al., 1990; and
particular way. Webster & Martocchio, 1992, 1993). For ex-
ample, play has been found to be positively
Third, we examine the extent to which indi- associated with mood, satisfaction, learning,
viduals' pre-training work and play expectations performance, and intrinsic motivation.
interact with their perceptions of the trainers'
work and play orientation to influence training
Research that has taken an individual difference
outcomes. Research suggests that it is impor-
approach to work-versus-play orientation sug-
tant to explore the effects of individual differ-
gests that one of the most important aspects of
ence factors on work-and-play labeling (e.g.,
playfulness in human computer interaction is
Tang & Baumeister, 1984). In addition, re-
cognitive spontaneity (Martocchio & Webster,
search suggests that pre-training expectations
1992; Webster & Martocchio, 1992). This as-
of a task (e.g., as work or play) are important to
pect of playfulness refers to an individual's
statistical control (Martocchio, 1992; Webster &
tendency to interact spontaneously, inventively,
Martocchio, 1993). However, little research has
and imaginatively with microcomputers (Web-
explored the extent to which pre-training work- ster & Martocchio, 1992). It is hypothesized that
and-play expectations interact with perceptions, those who are higher in microcomputer playful-
which are not experimentally manipulated, that
ness will view computer interactions more
the trainer made the task seem like work or positively, be more motivated to engage in com-
play.
puter interactions in the future, engage in more
Fourth, we explore the relationship between exploratory behaviors, and thus learn more than
perceptions of training as work, play, and fun less playful people.
and determine their effects on learning in order
to understand better the types of task percep- In their review of multiple microcomputer
tions that are likely to have beneficial effects. studies, Webster and Martocchio (1992) found
To date, research has primarily focused on the that the individual attribute of microcomputer
role of work and play perceptions in the context playfulness was positively related to computer
of computer training. In addition, little research attitudes, computer competence, computer effi-
has explored dimensions of play other than cacy, involvement, positive mood, satisfaction,
cognitive playfulness in computer training (Mar- and learning. In addition, they found that play-
tocchio & Webster, 1992; Webster & Mar- fulness influenced learning more than other
tocchio, 1992). attitude factors (e.g., computer anxiety). Com-
puter training study results are consistent with
Theoretical Basis for Work-Versus- research that finds that measures of general
Play Effects playfulness are positively related to creativity,
and more exploratory behaviors during inter-
Research suggests that whether individuals actions with tasks (e.g., Glynn & Webster,
approach a task as work versus play may influ- 1992).


Similarly, those taking a social contextual training may prime trainees' work or play cate-
approach to work-versus-play orientation sug- gories which in turn influence trainees' evalu-
gest that labeling a task as work or play will ations of the task and the extent to which they
affect the evaluation of a task, which in turn, will learn in training.
influence learning in training (Webster &
Martocchio, 1993). The theoretical basis for this Consistent with this, Glynn (1994) found that the
approach is social information processing theory influence of task labeling on performance
developed by Salancik and Pfeffer (1977, 1978), outcomes was mediated by cognitive processes
cognitive categorization theory (Fiske & Taylor, (i.e., means-end orientation). Specifically, she
1991), and play research and theory. Speci- found that the label play induced a means
fically, social information processing theory orientation which, in turn, resulted in higher task
suggests that task attitudes may be influenced performance outcomes. In contrast, the label
by the labeling of tasks by others (Salancik & work induced an ends orientation which, in turn,
Pfeffer, 1977, 1978; Staw, 1984). The social resulted in lower task performance outcomes.
context makes certain information salient to the Webster and Martocchio (1993) found that
individual. Consistent with this, organizational younger employees who received training
research finds that labeling a task as work- labeled as play showed higher motivation to
versus-play influences a variety of work related learn and learned more than older employees.
outcomes (e.g., learning, positive affect, task Finally, Webster et al. (1990) found that stu-
evaluation, motivation) with more positive out- dents experienced higher mood and involve-
comes typically found in the play condition ment and learned more in computer training
(Cellar & Barrett, 1987; Sandelands, 1988; classes labeled as play rather than work.
Webster et al., 1990; and Webster & Martoc-
chio, 1993). While we have suggested that a play orientation
is frequently found to have positive effects,
One mechanism that has been used to explain research and theory suggests that it can also
these labeling effects is cognitive categor- have potentially negative effects. For example,
ization. Research and theory suggests that Sandelands (1988) found that when a task was
individuals have and use different types of job- labeled play, individuals took longer on the task
related categories under certain conditions (Ku- than when the task was labeled work.
Ilk & Perry, 1994). For example, there is some Csikszentmihalyi (1975) suggested that playful-
evidence that individuals have work-and-play ness may lead to over-involvement. Finally,
cognitive categories which are more or less Schuck (1985) argued that play can have
available to the perceiver (e.g., Cellar & Barrett, negative consequences for work equipment and
1987; Porac & Meindl, 1982; and Webster & processes. However, there is little research
Martocchio, 1993). evidence that a play orientation has negative
consequences in the context of computer train-
Play research and theory suggests that the play ing (e.g., Martocchio & Webster, 1992; Webster
category is likely to include a greater emphasis et al. 1990; and Webster & Martocchio, 1992,
on means than ends and feelings of both plea- 1993).
sure and involvement (Sandelands, 1988;
Sandelands & Buckner, 1989; and Webster & There is some evidence that perceptions of play
Martocchio, 1993). Researchers studying play may have a greater influence on training out-
(e.g., Csikszentmihalyi, 1975; Miller, 1973; comes than perceptions of work. Much of the
Sandelands, 1988; Sandelands & Buckner, research that has explored the effects of task
1989; and Schuck, 1985) suggest that during labeling on work outcomes has manipulated
more playful interactions with tasks, people en- work and play task perceptions in an experi-
gage in exploratory behaviors, spend more time mental between-subject factorial design. How-
and effort on task: performance, enjoy what they ever, Cellar and Barrett (1987) assessed the
are doing more, and learn more effectively. effects of a within-subject measure of work and
Therefore, labeling a task as play should have play perceptions which enabled them to
more positive implications for learning than determine the relative influence of these percep-
labeling a task as work. This research and tions on task outcomes. They found that two
theory suggests that the social context of play measures significantly predicted intrinsic

96 The DATA BASE for Advances in Information Systems--Spring 1997 (Vol. 28, No. 2)

motivation, one work measure was non- ing as work should be high when the trainer is
significant, and the second work measure only perceived to have a high work orientation.
marginally significantly predicted intrinsic moti-
We suggested earlier that individuals have
vation. Additionally, work-and-play research cognitive categories for work and play. These
and theory tends to emphasize the positive
categories provide expectations that guide the
impact of a playful approach to tasks, sug-
understanding of new information (Fiske &
gesting that a play orientation may have more Taylor, 1991). As a result, individuals are likely
important consequences for work outcomes
to have expectations about the extent to which
than a work orientation. This suggests that play training will be like work or play prior to the start
perceptions may have a greater impact on work of the training. These expectations result from
outcomes than work perceptions.- the priming of one category or the other. Work-
and-play categories may be activated by the so-
Based on the above research, we expect that
cial context. For example, individuals who have
individuals who perceive training as play will ex-
previously taken the training course may tell
perience greater positive outcomes than those
current trainees what they thought of the train-
who perceive it as work. This positive effect is
ing. Alternatively, work-and-play categories
expected regardless of whether perceptions are
may be activated by individual level attributes.
a function of individual attributes or social con-
For example, those who have a higher level of
text. In addition, play perceptions are expected
the playfulness trait may have a play category
to have a greater influence on training outcomes
that is chronically activated. Research suggests
than work perceptions. Therefore we hypothe-
that pre-training expectations are likely to influ-
size the following:
ence training outcomes and therefore are often
statistically controlled (Martocchio, 1992; Web-
Hypothesis 1: Trainees' perceptions of com-
ster & Martocchio, 1993).
puter training as play will more positively and
significantly affect training outcomes (e.g.,
However, while researchers have called for
learning) than their perceptions of training as
research exploring the moderating effects of
work.
individual differences on work-and-play labeling
(Cellar & Barrett, 1987; Tang & Baumeister,
We have suggested that individuals' percep- 1984), little research has directly explored the
tions of a task as work or play may be the result extent to which pre-training work-and-play
of individual difference characteristics as well as expectations are moderated by trainer work-
social contextual factors. Specifically, previous and- play orientation. We suggest that indivi-
research has found that having a trainer label a duals' pre-training expectations about the extent
task as work-versus-play influenced training to which the task will be like work or play will
outcomes (Webster et al., 1990; Webster & interact with perceptions of the trainer's
Martocchio, 1993). Therefore, it is likely that orientation (perceptions that the trainer made
trainers can and often do influence individuals' the training seem like work or play) to influence
perceptions of the task as work-versus-play. training outcomes such as learning. It seems
This might occur because of the trainer's label- logical to predict that individuals whose expec-
ing of the task as well as his or her presentation tations are consistent with perceptions of the
of the material and behavior during the training. trainer's work and play orientation will have
For example, some research suggests that more positive outcomes than those whose
behavior modeling can have particularly positive expectations are inconsistent.
implications for learning computer software
(e.g., Gist et al., 1989). Therefore, we hypothe- Consistent with this, Tang and Baumeister
size the following: (1984) found that individuals with a high work
ethic showed greater intrinsic motivation on a
Hypothesis 2: Trainees' perceptions of the task labeled work-versus-play than those with a
training as work and play will be influenced by lower work ethic. The effect was eliminated or
the trainer. Specifically, trainees' perceptions reversed when the task was labeled leisure (i.e.,
of the training as play should be high when play). To the extent that worker values (e.g.,
the trainer is perceived to have a high play work ethic) are correlated with task expec-
orientation. Trainees' perceptions of the train- tations, this suggests that exl-Jctations that are


fulfilled will have more positive consequences If perceptions of a task as work, play, and fun
than expectations that are unfulfilled. In are distinct, it is not entirely clear how they may
addition, previous research suggests that play be related to one another and how they influ-
perceptions may, have more positive and ence training outcomes relative to one another.
significant effects on training outcomes than Although no firm hypotheses are offered, we
work perceptions (e.g., Cellar & Barrett, 1987). expect that individuals form perceptions of
Therefore, the extent to which trainees' play training tasks as work, play, and fun and that
expectations are fulfilled or unfulfilled should these perceptions independently affect learning.
have a greater effect on training outcomes than
work expectations. We therefore posit the The Study
following hypothesis:
The current research was a field study of MBA
Hypothesis 3: Training outcomes (e.g., learn-
students enrolled in a compulsory software-
ing) will be more positive when trainees' pre- training module. Questionnaires asking train-
training work-and-play expectations are con-
ees about their work-and-play related expecta-
sistent with perceptions of the trainer's work- tions and perceptions of the training were ad-
and-play orientation than when they are ministered prior to and following the training
inconsistent. The extent to which play expec- module respectively. A field study was used
tations are fulfilled or unfulfilled will have a because we were primarily interested in describ-
stronger influence on training outcomes than ing how individuals spontaneously and naturally
the extent to which work expectations are perceive training contexts, the extent to which
fulfilled or unfulfilled. these perceptions are influenced by trainers,
and the extent to which these perceptions
Research suggests that playfulness encom- influence learning in training.
passes cognitive (e.g., cognitive playfulness),
affective (e.g., manifest joy), and behavioral
(e.g., physical spontaneity) components (Bar- The Method
nett, 1990; Glynn & Webster, 1992). Webster
and Martocchio (1992) argued that cognitive Sample
playfulness represents the most relevant aspect Participants in this study were 75 first-year MBA
of playfulness in human-computer interactions students at a large midwestern university. This
and have studied it to the exclusion of other sample consisted of 52 males and 23 females
dimensions. However, they suggest that further with an average age of 25.48 years and an
investigation should be made to determine average of 2.46 years of work experience.
whether other dimensions of playfulness are These individuals were enrolled in a compulsory
relevant in human-computer interactions. software-training module which covered Micro-
soft Excel 5.0.1 Although all first-year MBA
Some theorists have suggested that play and students were required to attend this training
fun are not synonymous. Specifically, Biesty module, students could exempt out if they
(1986) notes that recent work has confused play passed a competency test.
with fun. Because something is fun does not
mean that it is play. The notion that play and Procedure
fun are related but not necessarily completely First-year MBA students at this university are
overlapping concepts is suggested in research required to participate in a series of compulsory
by Glynn and Webster (1992) and Abramis microcomputer software-training modules cover-
(1990). Glynn and Webster (1992) found that a ing a variety of software packages (e.g., word-
measure of playfulness was comprised of five processing, presentation graphics, spread-
factors, one of which was fun. Abramis (1990) sheet). The MBA training coordinator was con-
found that one measure of play was positively, tacted by the experimenters and agreed to allow
significantly, and highly correlated with six mea- the experimenters access to first-year MBA
sures of various aspects of fun. This research students prior to and following their participation
and theory suggests that perceptions of a task
as fun may be significantly correlated with, but 1Microsoft Excel 5.0, Copyright by Microsoft Corporation,
distinct from, perceptions of a task as work or 1985-1993. All rights reserved.
play.

98 The DATA BASE for Advances in Information Systems-- Spring 1997 (Vol. 28, No. 2)

in one training module. The Excel training research. Learning was assessed by an Excel
module was selected primarily because the quiz which was included in both the pre- and
training coordinator believed that this module post-training questionnaires. Two versions of
would allow the most time for the study and the pre-training and post-training questionnaires
because the experimenters' presence would be were given to trainees. These versions differed
least disruptive in this module. Subjects partici- only in the order in which measures were col-
pating in this module were assigned to one of lected. In addition, the order of the items on the
seven sections each of which was taught by a Excel quiz was randomly determined and was
different "head" trainer. A total of 273 students different in the pre-training and post-training
was enrolled in the Excel training module. questionnaires.

Prior to the start of the Excel training module, Measures
two assistants administered a questionnaire to
all of the individuals present in each of the Pre-training Play Expectations
seven training sections. A cover letter from the Two items measured the extent to which sub-
experimenters was attached to the question- jects expected the computer training to be like
naires stating that the purpose of the study was "play." These items included: "To what extent
to "better understand how people approach do you expect the Excel training to be like
training and learn in training programs." The 'play'?" (1 = Not at all like play; 7 = Very much
letter also indicated that participation in this like play) and "To what extent do you expect it to
study was voluntary and that all requested infor- feel like you are playing in the Excel training?"
mation would be kept confidential. (1 = Not at all; 7 = Very much). These items
were averaged to form a pre-training play ex-
One hundred and sixty-eight individuals (61% of pectations scale (r = .65, io < .01).
the total enrolled) returned completed pre-
training questionnaires. The Excel training Pre-training Work Expectations
module consisted of approximately five sessions Two items measured the extent to which sub-
which took place over a five-week period. At the jects expected the computer training to be like
completion of this module, the same two assis- "work." These items included: "To what extent
tants administered a questionnaire to all of the do you expect the Excel training to be like
individuals present in each of the seven training 'work'?" (1 = Not at all like work; 7 = Very much
sections. One hundred and two individuals like work) and "To what extent do you expect it
completed the post-training questionnaire (37% to feel like you are working in the Excel
of the total enrolled). Only those trainees who training?" (1 = Not at all; 7 = Very much). These
returned both the pre- and post-training ques- items were averaged to form a pre-training work
tionnaires were included in the study analyses. expectations scale (_r = .46, 12 < .01 ).
The total number of trainees who completed
both the pre- and post-training questionnaires Post-training Play Perceptions
was 77 (28% of the total enrolled). Two sub- Two items, parallel to those used in the pre-
jects' data were eliminated because they made training play expectations scalel were asked
ratings off of the rating scales, resulting in a following the completion of the Excel training
final sample size of 75. module. These items included: "To what extent
was the Excel training like 'play'.'?" (1 = Not at all
The MBA training coordinator requested that like play; 7 = Very much like play) and "To what
administration of the questionnaires take no extent did it feel like you were playing in the
more than fifteen minutes of the training class. Excel training?" (1 = Not at all; 7 = Very much).
Therefore, there were limitations on the number These items were averaged to form a post-
of questions that trainees could be asked. The training play perceptions scale (_r = .69, 12 < .01 ).
questionnaires collected information about work,
play, and fun expectations and perceptions, Post-training Work Perceptions
control variables (e.g., microcomputer experi- Two items, parallel to those used in the pre-
ence, computer anxiety), and the dependent training work expectations scale, were asked
measure. Because the primary objective of this following the completion of the Excel training
training module was to increase learning, learn- module. These items included: "To what extent
ing was the dependent measure used in this was the Excel training like 'work'?" (1 = Not at


all like work; 7 = Very much like work) and "To at all knowledgeable; 7 = Extremely knowledge-
what extent did it feel like you were working in able). These items were averaged to form an
the Excel training?" (1 = Netat all; 7 = Vey nuah). Excel knowledge scale (Cronbach's alpha =
These items were averaged to form a post- .96).
training work perceptions scale (.[ = .75, 1o< .01).
Computer Anxiety
Post-training Fun Perceptions Ten items adapted from Heinssen, Glass, and
Two items measured the extent to which Knight (1987) were used to assess computer
subjects perceived the training to be fun. These anxiety. A typical item on this scale was, "1 feel
items included: "ro what extent was the Excel insecure about my ability to interpret a computer
training fun?" (1 = Not at all fun; 7 = Very fun) printout" (1 = Strongly disagree; 5 = Strongly
and "How enjoyable was the Excel training?" (1 Agree). These items were averaged to form a
= Not at all enjoyable; 7 = Extremely enjoyable). computer anxiety scale (Cronbach's alpha =
These items were averaged to form a post- .83). Higher values indicate more computer
training fun perceptions scale (_r = .84, 12< .01 ). anxiety.

Trainer Work-Versus-Play Orientation Motivation
Two questions on the post-training question- Seven items were used to assess pre-training
naire asked subjects to indicate the extent to motivation for the Excel training. These items
which they perceived that the training instructors were adapted from Baldwin and Karl (1987). A
made the Excel training seem like work and play sample item is, "1 am willing to exert consider-
respectively. One item assessed trainer play able effort to improve my skills in the upcoming
orientation, "To what extent did your instructors Excel training module" (1 = Strongly disagree; 7
make the Excel training seem like play?" (1 = Not = Strongly agree). These items were averaged
at all; 7 = Very much). The other item assessed to form a motivation scale (Cronbach's alpha =
trainer work orientation, "To what extent did .85). Higher values indicate more motivation.
your instructors make the Excel training seem
like work?" (1 = Not at all; 7 = Very much).
Demographic Information
Demographic information was collected for des-
A number of control measures were also col- criptive purposes only. Subjects were asked to
lected on the pre-training questionnaire. These indicate their age, sex, and years of full-time
measures are consistent with those used in re- work experience.
lated research (e.g., Martocchio & Webster,
1992; Webster & Martocchio, 1993). These Learning
measures are described next. The same ten-item, multiple-choice quiz asses-
sing Excel knowledge was administered prior to
PC knowledge and following the Excel training module. The
Two items were used to assess the extent to content of the quiz was developed by the
which individuals were knowledgeable about experimenters with the help of the MBA training
personal computers. These items included: coordinator based on actual competency tests
"How skilled are you at using PCs?" (1 = Not at used by the training program in previous years.
all skilled; 7 = Extremely skilled) and "How much Items on the quiz were scored 0 = incorrect or 1
experience have ,.you had using PCs~ (1 = No = correct. The total number of items answered
experience at all; 7 = A great deal of experi- correctly was summed to form a measure of quiz
ence). These items were averaged to form a performance at both pre- and post-training.
PC-knowledge scale (r = .86, 12< .01). Learning was measured as post-training quiz
performance controlling for pre-training quiz
Excel Knowledge performance.
Three items were used to assess the extent to
which individuals were knowledgeable about Ex- Results
cel. These items included: "How skilled are you
at using Excel?" (.1 = Not at all skilled; 7 = Ex- Correlations among Study Measures
tremely skilled); "How much experience have
you had using Excel?" (1 = No experience at all; A correlation matrix which includes the study
7 = A great deal of experience); and "How measures is located in Table 1. This table re-
knowledgeable are you about Excel?" (1 = Not veals several significant and interesting


Variables 1 2 3 4 5 6 7 8 9 10 11 12 13

1. Play - .14 .63"* .64"* .10 .27" .06 .07 -.03 -.16 .09 -.03 -.04
Perceptions

2. Work .25" .21 .67** -.02 .53** .23* -.23* -.41"* .40** -.21 -.12
Perceptions

3. Fun - .55"* .22 .02 .08 .06 .01 -.12 .22 -.10 .13
Perceptions

4. Trainer -.03 .05 .03 -.05 .04 -.12 .10 -.06 .13
Play
Orientation

5. Trainer -.01 .43** .36** -.18 -.24* .33** -.19 -.32**
Work
Orientation

6. Play -,18 -.12 .13 .16 .18 .17 -.03
Expectations

7. Work .25" -.16 -.43"* -.43" -,23" -.16
Expectations

8. Computer -.54"* -.49"* .15 -.39"* -.40"*
Anxiety

9. Pc .55** -.06 .48** .20
Knowledge

10. Excel -.36"* .64** .20
Knowledge

11. Motivation -.03 .01

12. Pre-Training .30"*
Quiz
Performance

13, Post-Training
Quiz
Performance

** 12< .01, * 12 < .05, two-tailed
Table 1. Correlations among Study Measures

correlations. First, post-training fun perceptions correlated with PC and Excel knowledge and
were significantly and positively correlated with pre- and post-training Excel quiz performance.
both post-training play and post-training work This is consistent with research that finds that
perceptions although the former correlation is computer anxiety can have negative implica-
higher and more significant than the latter. This tions for learning (e.g., Heinssen et al., 1987;
is consistent with research and theory that sug- Martocchio, 1992; and Martocchio & Webster,
gests that fun and play are not necessarily 1992). Third, a positive and significant correla-
equivalent concepts and that both work and play tion was found between computer anxiety and
may be fun. Second, this matrix reveals that post-training work perceptions and pre-training
computer anxiety is negatively and significantly work expectations. This suggests that indi-


viduals who werE; more anxious about com- In the first two-step hierarchical regression
puters were more likely to perceive computer analysis, the first step in the analysis was
training as work. This finding is consistent with significant, indicating a significant main effect of
research suggesting that work perceptions are the control variables on learning. Specifically,
less likely to be .associated with positive out- computer anxiety was found to have a signifi-
comes than play perceptions. cant and negative effect on learning. However,
the second step in this analysis did not account
Effects of Work-and-Play Perceptions on for a significant increase in variance. Post-train-
Learning ing perceptions of the extent to which the train-
To test Hypothesis; 1 and examine the effects of ing was like play did not influence learning in
perceptions of work compared to play on train- the training beyond the effects of the control
variables.
ing program learning, two, two-step hierarchical
regression analyses were conducted. In the
first hierarchical regression analysis, all control In the second, two-step hierarchical regression
variables were entered in the first step and the analysis, the first step was the same as that
post-training play perceptions scale was entered reported for the first hierarchical regression
in the second step. The control variables in- analysis. Of particular interest is the second
cluded: computer anxiety, PC knowledge, Excel step which did not account for a significant in-
knowledge, motivation, and pre-training quiz crease in variance. Post-training perceptions of
performance. the extent to which the training was like work did
not influence learning in training beyond the
In the second, two-step hierarchical regression effects of the control variables. These results
analysis, the control variables were entered in do not support Hypothesis 1 which predicted
the first step and the post-training work per- that perceptions of the training as play would
ceptions scale was entered in the second step. positively and significantly influence learning
The increase in r2 contributed by each step is of and would do so to a greater extent than
particular interest (Horn, Griffith, & Sellaro, perceptions of the training as work. These
1984). if the r2 change is significant, it suggests analyses suggest that while the model appears
that the associated step accounts for a to account for a significant amount of variance
substantial amount of additional variance in the in learning, neither play nor work perceptions
dependent variable. These analyses are sum- accounted for additional variance in learning
marized in Table 2. beyond the effects of the control variables.

Step Independent Measure " beta r2 &r2
(1) Computer Anxiety -.40"*
PC Knowledge -.09
Excel Knowledge -.09
Motivation .04
Pre-Training Quiz Performance .24t .20"* .20"*

(2) Play Perceptions -.03 .20" .00

(2) Work Perceptions -.07 .20" .00

(2) Play Perceptions -.23t
Fun Pemeptions .32" .26"* .06t

**]2<.01, *D<.05, t l ~ < . 1 0
Table 2. Hierarchical Regression Analyses on Leaming


Work, Play, and Fun Perceptions play perceptions (beta = -.24, 12 < .10) and a
significant and positive effect of post training fun
Analyses were conducted to determine the perceptions on learning (beta = .35, p < .05).
relationship between work, play, and fun Post-training work perceptions did not have a
perceptions and how these perceptions, in turn, significant effect on learning.
influenced ]earning in training. First, correlation
analyses indicated that the correlation between This and the correlation analyses indicate that
post-training pray and fun perceptions was the post-training play perceptions measure
positive and significant (_r = .63, p < .01 ) as was operated as a suppressor variable (Cascio,
the correlation between post-training work and 1991; Pedhazur, 1982). First, post-training play
fun perceptions (r = .25, p < .05). This pattern and fun perceptions were positively and sig-
of results is consistent with the contention that nificantly correlated with each other. Second,
fun is more highly correlated with play compared post-training perceptions of play had a near
to work (Dandridge, 1986). Results are also zero correlation with the criterion (post-training
consistent with research by Abramis (1990) quiz performance). Third, when the post-
which found that a measure of play was signi- training play perceptions scale was entered into
ficentiy and highly correlated with six measures the regression equation, the coefficient on the
of various aspects of fun (_r= .37 - .52). post-training fun perceptions scale increased
(from .19 to .35), r2 increased (from .23 to .27),
Second, a regression analysis was conducted and the coefficient on the post-training play per-
which regressed post-training quiz performance captions scale was negative and marginally
on the five control variables (computer anxiety, significant (beta = -.24, p < .10).
PC knowledge, Excel knowledge, motivation,
and pre-training quiz performance), and the This series of analyses suggests that fun and
post-training work, play, and fun perception play perceptions are related but not necessarily
scales. The results of this regression analysis equivalent concepts. In light of these results, a
are summarized in Table 3. This analysis re- two-step hierarchical regression analysis was
vealed that the variables explained a significant conducted to determine whether post-training
amount of variance in learning (r2= .27, .o < .01 ). play and fun perceptions together accounted for
a significant amount of additional variance in
Computer anxiety was found to have a signifi- learning beyond the effects of the control vari-
cant and negative effect on learning, while pre- ables. In the first step, the control variables
training quiz performance had a positive and were entered (computer anxiety, PC knowledge,
significant effect on post-training quiz perfor- Excel knowledge, motivation, and pre-training
mance. In addition, this analysis revealed a quiz performance). In the second step, both the
marginal and negative effect of post-training post-training play and fun perceptions scales
were entered. The results are reported in Table
2. The analyses indicate that the second step
accounted for a marginally significant amount of
Variable beta
additional variance in learning. In addition, con-
Computer Anxiety -.40"* sistent with earlier analyses, post-training play
PC Knowledge -. 11 perceptions had a marginally significant but
Excel Knowledge -. 19 negative effect on learning while post-training
Motivation .00 fun perceptions had a significant and positive
Pre-Training Quiz Performance .32*
effect on learning.
Play Perceptions -.24t
Work Perceptions -. 12
Trainer Effects
Fun Perceptions .35"

r2 .27** Hypothesis 2 predicted that trainees' per-
ceptions of the training as work and play would
be influenced by the trainer. We predicted that
**1::)<.01, * p < . 0 5 , t P < . 1 0 trainees' play perceptions would be high when
Table 3. Regression Analysis Exploring their trainer was perceived to have a high play
the Effects of Post-training Work, Play, orientation and trainees' work perceptions would
and Fun Perceptions on Learning be high when their trainer was perceived to

The DATA BASE for Advances in Information S y s t e m s - Spring 1997 (Vol. 28, No. 2) 103

Variable Play Perceptions
beta r2
Trainer Play Orientation .64"*
Trainer Work Orientation •12 .42.*

Work Perceptions
Trainer Play Orientation .23**
Trainer Work Orientation .67"* .49**

Fun Perceptions
Trainer Play Orientation .55**
Trainer Work Orientation .23" .35"*

*'12<.01, *10<.05, 1"12<.10
Table 4. Regression Analyses Exploring the Effects of Trainer Work-and-Play
Orientation on Post-training Work, Play, and Fun Perceptions
have a high work orientation. In order to test ations, post-training play perceptions, and post-
this hypothesis, a number of analyses were con- training work perceptions respectively on the
ducted. training section, which was dummy coded. Re-
call that subjects were assigned to one of seven
First, post-training work-and-play perceptions training sessions each of which had a different
scales respectively were regressed on the two head trainer. Analyses revealed that there was
items that measured perceptions of the trainer's no main effect of section on the pre-training play
work and play orientation. The first question expectations F(6, 68) = 1.30, 2 = ns, r2 = .10, or
asked trainees to indicate the extent to which pre-training work expectations scales F(6, 68) =
the trainers made the Excel training seem like .58, 12 = ns, r2 = .05. However, a main effect of
work and the second question asked trainees to section was found on the post-training play
indicate the extent to which the trainers made perceptions scale F(6, 68) = 2.90, 13 < .05, r ~ =
the Excel training seem like play. Results of this .20, but not the post-training work perceptions
analysis are presented in Table 4. Consistent scale F(6, 68) = 1.18, p = ns, r2 = .09. Third,
with Hypothesis 2, results indicated that post- additional correlation analyses suggested that
training play perceptions were significantly influ- training section was significantly correlated with
enced by perceptions of the trainer's play trainer play orientation (r = .31, 12< .01), but not
orientation. Also consistent with Hypothesis 2, trainer work orientation (r = .00, 1o= ns).
post-training work perceptions were significantly
influenced by perceptions of the trainer's work Together these analyses suggest that social
orientation. Although unanticipated, this ana- context (e.g., trainers) can influence trainees'
lysis also revealed that post-training work work and play perceptions even in the absence
perceptions were significantly influenced by of explicit instructions to label tasks as work-
perceptions of the trainer's play orientation. versus-play.
Finally, even though we did not hypothesize that
instructor orientation would influence post- Trainee by Trainer Interaction Effects
training perceptions of fun, we conducted addi-
tional analyses to explore this possibility. Hypotheses 3 predicted that trainees' pre-
Analyses revealed that post-training fun per- training expectations of the task as work or play
ceptions were significantly influenced by per- would interact with trainer work and play
ceptions of both the trainer's work and play orientation to influence training outcomes. Spe-
orientation. cifically, we predicted that training outcomes
(learning) would be more positive when
Second, four additional regression analyses trainees' expectations about whether the task
were conducted which regressed pre-training would be like work and play were consistent
play expectations, pre-training work expect- with their perceptions of whether the trainer in


fact made the task seem like work and play.
Variable Beta
Additionally, we suggested that the fulfillment or
Computer Anxiety -.25?
lack of fulfillment of play expectations was likely
PC Knowledge -. 11
to have a greater effect on learning than the
Excel Knowledge -.03
fulfillment or lack of fulfillment of work expec-
tations. Motivation .14
Pre-Training Quiz Performance .24
In order to test this hypothesis, a regression Play Expectations (PE) .13
analysis was conducted in which post-training Work Expectations (WE) -.28
quiz performance was regressed on the control Trainer Play Orientation (TP) -. 12
variables (computer anxiety, PC knowledge, Ex- Trainer Work Orientation (TW) .29
cel knowledge, motivation, and pre-training quiz PE x TP .63
performance), pre-training work and play expec- PE x TW -1.11"
tations, trainer work-and-play orientation, and WE x TP -.08
the four relevant two-way interactions between WE x TW .43
trainees' work-and-play expectations and trainer R2 .34"
work-and-play orientation. This regression *P < .05, t P < . 1 0
analysis is reported in Table 5.
Table 5. Regression Analyses Exploring the
Analyses suggest that these variables account- Independent and Combined Effects of Pre-
ed for a significant amount of variance in training Work and Play Expectations and
learning (r2 = .34, P < .05). Results revealed a Trainer Work and Play Orientation on
marginally significant and negative effect of Learning
computer anxiety on learning. In addition, a
significant and negative effect for one of the learning in a field study. The main findings of
two-way interactions was found (pre-training this research are reviewed next.
play expectations x trainer-work orientation).
Following procedures recommended by Peters, Effects of Work-and-Play Perceptions on
O'Connor, and Wise (1984) we used within- Leaming
subgroup regression equations to examine this
interaction. First, high play expectation and low We predicted that perceptions of the training as
play expectation conditions were created by play would more positively and significantly
doing a median split on the pre-training play affect learning than perceptions of the training
expectations scale. Next, equations regressing as work after controlling for a number of factors
post-training quiz performance on trainer work (Hypothesis 1). However, contrary to prediction,
orientation were plotted for the high and low perceptions of the computer training as play did
play expectation conditions using unstan- not account for a significant amount of addi-
dardized regression coefficients. This plot is tional variance in learning after controlling for
shown in Figure 1. computer anxiety, PC knowledge, Excel know-
ledge, motivation, and pre-training quiz per-
Consistent with Hypothesis 3, this figure formance. As expected, perceptions of the
indicates that subjects learned less when their training as work did not account for a significant
pre-training play expectations were high and amount of additional variance in learning be-
they perceived that the trainer made the training yond the effects of the control variables. These
seem more like work. However, the interaction findings are inconsistent with prior research
between pre-training play expectations and which finds that labeling computer training play
trainer play orientation was not significant. positively influences training outcomes (Web-
Consequently, only partial support was found for ster et al., 1990; Webster & Martocchio, 1993).
Hypothesis 3. These results suggest that un-
fulfilled play expectations have a greater impact One explanation for the lack of effect for play
on learning than fulfilled play expectations. perceptions in the current study is that indi-
viduals' perceptions of play may be especially
Discussion predictive when trainers explicitly label com-
puter training tasks. Prior research that has
This research explored the effects of work, play, explored the effects of labeling in computer
and fun perceptions of computer training on training has explicitly manipulated task labels.


Explicit manipulations of tasks as work or play influence on training outcomes. Learning was
are likely to prime individuals' work and play regressed on a number of control variables, and
categories and influence training outcomes. In post-training work, play, and fun perceptions.
the current study, perceptions were measured Results of this regression analysis and cor-
and not manipulalled. This suggests that per- relation analyses indicated that the post-training
ceptions of a task as play may be less influential play perceptions measure operated as a sup-
when the social context does not consistently pressor variable (Cascio, 1991; Pedhazur,
and explicitly m~=ke task labels salient. A 1982). Specifically, post-training perceptions of
second explanation for the lack of effect for play fun appeared to have a significant and positive
perceptions is suggested by the research of effect on learning while post-training percep-
Webster and Martocchio (1993) which found tions of play had a marginally significant and
that the effects ,of labeling training play on negative effect on learning. No significant effect
learning were moderated by an individual was found for post-training work perceptions.
difference characteristic (i.e., age). This re-
search suggests that play perceptions may only The current results suggest that perceptions of
have beneficial el~ects for certain types of indi- the task as play had a negative effect on
viduals obscuring a main effect in the present learning in the training. This may be because
study. individuals who perceived the training as play
Work, Play, and Fun Perceptions perceived it to be less serious or important and
therefore performed less well. On the other
Further analysis of the data suggested that play hand, those who perceived the training as fun
perceptions did in fact influence learning but did found the training enjoyable, may have been
so in a complicated manner. The current study more motivated to perform in the training, and
measured the extent to which trainees per- learned more. This interpretation is consistent
ceived the training as work, play, and fun in or- with the pattern of correlations that reveals that
der to determine the relationship between these pre-training motivation was more highly and
perceptions and to assess their relative positively correlated with post-training fun

10-
9 - ...

8
Post-training
Quiz 7
Performance 6

(Numberof 5
correct~tem~) 4"

3
2
1
0 I I '1 ' I I I : :
0 1 2 3 4 5 6 ? 8
Low Orientation Trainer Work Orientation High Orientation

Figure 1. Interaction Between Pre-training Play Expectations
and Trainer Work Orientation on Learning


perceptions (r = .22, 12 = ns) than post-training primarily driven by differences between trainers
play perceptions (r = .09, 12 = ns) although appears reasonable in light of the first set of
neither of these correlations is significant. This analyses showing the direct effect of trainer
interpretation is also consistent with the finding orientation on trainees' work, play, and fun per-
that post-training work perceptions were more ceptions. This assumption is also consistent
highly correlated with post-training fun per- with the fact that the training module material
ceptions (r = .25, 12< .05) than post-training play and computer resources were standardized
perceptions (r = .14, 12 = ns) and that post- across sections.
training work perceptions were highly and
positively correlated with motivation (r = .40, p_<
.01). Trainee by Trainer Interaction Effects

The suppressor effect suggests that play and Hypothesis 3 suggested that trainees' pre-
fun perceptions were highly related but percep- training expectations about whether the task
tions of the training as fun were beneficial while would be like work or play would interact with
perceptions of the training as play were not. trainer orientation, whether trainees perceived
This pattern of results is consistent with re- the trainer made the task seem like work or
search and theory that suggests that play and play, to influence learning. Specifically, we sug-
fun are not necessarily equivalent concepts and gested that pre-training work-and-play expec-
that work as well as play may be fun (Abramis, tations that were consistent with perceptions of
1990; Biesty, 1986; Burke, 1971; Glynn and the trainer's work-and-play orientation would
Webster, 1992). It is also consistent with the have a positive influence on learning, while
finding that a play orientation can in some cir- expectations that were inconsistent with the
cumstances have negative effects (e.g., San- trainers orientation would have a negative influ-
delands, 1988; Schuck, 1985). ence on learning. Further, we anticipated that
the fulfillment or lack of fulfillment of play ex-
Trainer Effects pectations would have a greater effect on
learning than the fulfillment or lack of fulfillment
Results suggest that trainers influenced
of work expectations. Partial support for this
trainees' work-and-play perceptions. This was
hypothesis was found.
indicated by a number of analyses. First, work
and play perceptions respectively were re-
gressed on perceptions that the trainer made A regression analysis was conducted which re-
the training seem like work and like play. These gressed learning on the control variables, work-
analyses indicated that trainer work orientation and-play expectations, trainer work-and-play
influenced trainees' perceptions of the training orientation, and the interactions between trainee
as work and trainer play orientation influenced work-and-play expectations and trainer work-
trainees' perceptions of the training as play. and-play orientation. The regression accounted
Additional analyses also revealed that trainer for a significant amount of variance in learning
orientation significantly influenced post-training and indicated that one interaction term had a
fun perceptions. significant and negative effect on learning. Spe-
cifically, trainees who had higher play
Second, results revealed no differences be- expectations learned less when they perceived
tween training sections on work-and-play ex- that their trainer made the training seem more
pectations prior to the start of training. How- like work. Contrary to expectation, no evidence
ever, following the training, differences in play was found for a play expectations x trainer play
perceptions among training sections were orientation interaction, suggesting that unful-
found. Third, the training section was sig- filled play expectations had a greater influence
nificantly correlated with trainer play orientation. on learning than fulfilled expectations. Finally,
Although training- section effects may have also as expected, there was little evidence that the
resulted from other aspects of the training fulfillment or lack of fulfillment of trainees' work
context (e.g., classmates), the pattern of results expectations influenced learning. These ana-
presented suggests that trainers influenced lyses suggest that the trainer may play an
trainees' perceptions of the task. The important role in the context of computer
assumption that section differences are training.


Implications label training contexts. This may suggest that
trainers need to be aware of how they influence
First, current results suggest that perceptions of trainees' work, play, and fun perceptions and
a task as play may, not necessarily have a posi- learn how to intentionally influence these
tive effect and may even have a negative effect perceptions.
on learning. Our results suggest that per-
ceptions of training as fun may have more Third, current results revealed that the extent to
positive and beneficial effects on learning than which there was congruence between trainees'
perceptions of the training as play or work. The expectations that the training would be like work
post-training play perceptions measure ap- or play and perceptions that the trainer made
peared to operate as a suppressor variable, in- the training seem like work or play influenced
creasing the positiive effect of post-training fun learning. Specifically, results indicated that
perceptions on learning. Perceptions of the when individuals expected the training to be like
training as play may have been accompanied by play and they perceived that the trainer made it
perceptions that the training was tess serious or seem more like work, learning was negatively
important. In contrast, perceptions of the train- affected. This suggests that trainers should
ing as fun appeared to be more highly and attempt to assess and in some cases influence
positively correlated with pre-training motivation trainees' pre-training expectations. It is likely
and work perceptions. that many trainees have inappropriate and
unrealistic expectations about the training. In
The current results suggest that it is particularly addition, this finding suggests that trainers may
important that trainees perceive training as fun need to alter their behavior (where appropriate)
but not necessarily as play. This could be to reduce the potential for unfulfilled training
accomplished by having trainers explicitly label expectations. Therefore, these results suggest
training situations as fun. Research by Webster that it is important that trainers understand
and Martocchio (1993) suggests that this might trainees' expectations, make sure that they are
be accomplished in a fairly straightforward accurate and realistic, and behave in a manner
manner. Fun and enjoyable aspects of the that is consistent with these expectations.
training could be highlighted for trainees by the
trainer or training material. Finally, while trainers appeared to influence
trainees' perceptions and learning, the current
Second, study results suggest that social study also found some evidence that individual
context appears to have an influence on train- difference factors played an important role in
ees' perceptions of the training as work or play. learning. Specifically, subjects who experi-
Specifically, trainers appeared to influence enced a greater amount of computer anxiety did
trainees' work, play, and fun perceptions. To less well on the Excel quizzes. This finding is
the extent that these perceptions affect learning, consistent with previous research that finds that
trainer influence is important. These findings computer anxiety can have detrimental effects
suggest that trainers may be able to shape on learning (e.g., Heinssen et al., 1987; Martoc-
perceptions in a way that can have a significant chio, 1992; and Martocchio & Webster, 1992).
and positive influence on learning. For ex- This result suggests that it is important to
ample, current results suggest that trainers can consider the role of individual difference factors
shape the extent to which trainees perceive as well as social contextual factors in
training as fun and thus positively influence microcomputer training. This might suggest that
learning. This can be done by having trainers trainers assess the extent to which individuals
behave in particular ways (e.g., model certain are anxious about computer training and take
types of behavior), or by specifically labeling steps t o alleviate this anxiety (e.g., showing
training fun. The influence of trainers in the trainees that they are unlikely to break the
context of training is not always acknowledged computer).
nor is it aT~vays intentional. Current results
suggest that trainers are likely to differ in the Limitations and Future Research
extent to which they influence trainees'
perceptions of the training (as work, play, and These research results should be interpreted in
fun) in the absence of specific instructions to the context of several of the study's limitations.


First, and most importantly, this research was a was small in the context of some of our analyses
field study. Consequently, the external validity which included a relatively large number of
of this study is quite high, while internal validity variables. However, while a relatively small
is likely to be lower due to the lack of ex- sample may limit our ability to detect small
perimental control. Because much of the effects, it also increases our confidence in the
previous research on work-and-play orientation effects we did find. Effects must be quite large
in computer training has been experimental in to be detected in small samples using con-
nature, we chose to conduct a field study. ventional levels of statistical significance (Co-
However, this research was relational and ex- hen, 1988).
ploratory in nature and therefore few causal
claims can be made. For example, we cannot Future research should explore work, play, and
conclude that work and play perceptions were fun perceptions of more and different types of
influenced by trainer work-and- play orientation subjects. For example, current study subjects
based solely on the regression analyses had some familiarity with micro-computers be-
reported in Table 4. cause they had already taken several training
modules prior to the Excel module. Thus, ef-
One could argue that work and play perceptions fects found in this study may not generalize to
influenced perceptions of the extent to which less experienced trainees. However, we believe
trainers made the task seem like work or play. that these findings have implications for organi-
Therefore, wherever possible, we included zations. Study subjects had some work exper-
additional analyses to help isolate the nature of ience and organizations offer their employees
the effects that were found. For example, in software training similar to that provided in the
order to provide additional evidence that train- current study (e.g., Ballou & Rush, 1996).
ers influenced trainees' work-and-play per-
ceptions, we looked for differences in these The current study explored work, play, and fun
perceptions by training section before and after perceptions following subjects' participation in
the training. In addition, we believe that it is an Excel training module. However, this training
reasonable to conclude that perceptions of play module was the fourth module to be covered in
and fun influenced learning and not the reverse a series of modules. It is possible that the ef-
for two reasons. First, subjects did not know fects found in the current study might not be
how well they scored on either the pre- or the found in other types of computer-based training.
post-training Excel quiz. It is difficult to imagine Certain types of software may require more
how learning would influence perceptions of the creativity, spontaneity, and imagination to use
task if subjects were not told how much they (e.g., graphics software packages). We would
learned. Second, previous experimental re- predict that perceptions of the task as play
search provides evidence that task perceptions might be even more important in these training
(e.g., play perceptions) influence learning. contexts.

Although 273 individuals were enrolled in the In addition, computer training on non-computer
Excel training module, only 75 individuals pro- related topics (e.g., problem-solving) may differ
vided complete and usable pre- and post-train- from training on computer applications such as
ing questionnaires. This participation rate was Excel. It is also unclear how work-and-play
low and is likely due to two factors. First, perceptions may have differed across the
trainees believed that they could skip a training training modules. It might be that perceptions of
class when they were familiar with the material the training changed as the content of the
covered on that day. Second, the post-training training changed and as trainees' exposure to
questionnaire was administered close to stu- the trainer increased. Consequently, future re-
dents' midterm exams. This was unavoidable search should explore work, play, and fun
because the timing of both the training and perceptions in the context of different types of
exams was determined by the MBA department. microcomputer training and observe how these
perceptions change over time as a function of
It is not clear how representative these 75 the content of the training and the nature of the
individuals are of the entire MBA program or of trainees' relationship with the trainer. It is also
other individuals who are likely to take computer important to explore the extent to which work,
training classes. Additionally, the sample size play, and fun perceptions at different points


during the training are predictive of important software that shape trainees' perceptions and
training outcomes. learning much as human trainers do? Future
research is necessary to assess the differences
The current research used a multiple-item quiz between human-led and computer-based train-
administered prior' to and following the Excel ing and to determine the effect of incorporating
training to measure learning. However, it would the role of the trainer into the computer itself on
be useful to expllore other types of learning training outcomes.
measures in futu~re research. For example,
future research might explore the speed with References
which learned skills are executed in a simulated
work task. In addition, it may also be important Abramis, D. J. (1990). "Play in Work: Childish
to explore the number of errors made in the Hedonism or Adult Enthusiasm," American
execution of learned skills. Processing speed Behavioral Scientist, Vol. 33, pp. 353-373.
and errors measured in human-computer inter- Allan, K. (1993). "Computer Courses Ensure
action research can provide additional infor- Uniform Training," Personnel Journal, Vol.
mation about the extent to which computer skills 72, No. 6, pp. 65-71.
and knowledge are well-learned (e.g., Card, Baldwin, T. T., and Karl, K.A. (1987). "The De-
Moran, & Newell, 1983; Rubinstein & Hersh, velopment and Empirical Test of a Measure of
1984). Assessing Motivation to Learn in Management
Education," In F. Hoy (Ed.), Academy of
In addition, future research should develop and Management Best Paper Proceedings, New
assess construct valid measures of work, play, Orleans, LA, pp. 117-121.
and fun perceptions in the context of computer Ballou, D. J., and Rush, C. A. (1996). "A
training. The current research was exploratory Survey of Computer-Based Training Practices
in nature and suggests that work, play, and fun in a Sample of Canadian Financial Post 500
perceptions are distinct but correlated. How- Companies," Proceedings of the Human Re-
ever, this study does not provide detailed infor- sources Division of the Administrative
mation regarding the basis of these perceptions Sciences Association of Canada, Montreal,
or the nature of the relationship between these Quebec, pp. 53-63.
perceptions. In light of current study results, it Barnett, L.A. (1990). "Playfulness: Definition,
is important to explore the nature of work, play, Design, and Measurement," Play and Culture,
and fun perceptions in some detail and to deter- Vol. 3, pp. 319-336.
mine whether the concepts of play and fun are Biesty, P. (1986). "If It's Fun, Is It Play? A
in fact separable entities that influence learning Median Analysis," In B. Mergen (Ed.), Cul-
differently. This requires greater attention to the tural Dimensions of Play, Games, and Sport.
measurement of work, play, and fun perceptions Champaign, IL: Human Kinetics, pp. 61-72.
in future research. Burke, R. (1971). "'Work' and 'Play'," Ethics,
Vol. 82, pp. 33-47.
The current study suggests that trainers can Card, S. K., Moran, T. P., and Newell, A.
have a strong influence on trainees' perceptions (1983). The Psychology of Human-Computer
of training as work, play, or fun, with a resulting Interaction, Hillsdale, N J: Lawrence Erlbaum.
impact on learning. However, it would be inter- Cascio, W. F. (1991). Applied Psychology in
esting to explore the effects of computer-based Personnel Management (4th ed.), Englewood
training which effectively eliminates the role of Cliffs, N J: Prentice Hall.
the trainer on trainees' work, play, and fun Cellar, D. F., and Barrett, G.V. (1987). "Script
perceptions and learning. It may be the case Processing and Intrinsic Motivation: The Cog-
that software-based training is more or less like nitive Sets Underlying Cognitive Labels," Or-
work, play, or fun than human-led training. The ganizational Behavior and Human Decision
extent to which there are differences between Processes, Vol. 40, pp. 115-135.
the two types of training suggests the following Cohen, J. (1988). Statistical Power Analysis for
questions: Are individual differences in pre- the Behavioral Sciences, Hillsdale, N J: Law-
dispositions toward work, play, and fun more rence Erlbaum.
influential in human-led or computer-based Csikszentmihalyi, M. (1975). "Play and Intrinsic
training? Can computer-based training de- Rewards," Humanistic Psychology, Vol. 15,
velopers incorporate cues into the training pp. 41-63.


Dandridge, C. (1986). "Ceremony as Integra- Pedhazur, E. J. (1982). Multiple Regression in
tion of Work and Play," Organization Studies, Behavioral Research (2nd ed.), New York:
Vol. 7, No. 2, pp. 159-170. Holt, Rinehart and Winston.
Fiske, S. T., and Taylor, S. E. (1991). Social Peters, L. H., O'Connor, E. J., and Wise, S. L.
Cognition (2nd ed.), New York: McGraw Hill. (1984). "The Specification and Testing of
Geber, B. (1994). "Re-Engineering the Train- Useful Moderator Variable Hypotheses," In T.
ing Department," Training, Vol. 31, No. 5, pp. S. Bateman and G. R. Ferris (Eds.), Method
27-34. and Analysis in Organizational Research.
Gist, M. E., Schwoerer, C., and Rosen, B. Reston, VA: Reston Publishing, pp: 128-139.
(1989). "Effects of Alternative Training Porac, J. F, and Meindl, J. (1982). "Under-
Methods on Self-Efficacy and Performance in mining Overjustification: Inducing Intrinsic and
Computer Software Training," Journal of Ap- Extrinsic Task Representations," Organiza-
plied Psychology, Vol. 74, pp. 884-891. tional Behavior and Human Performance, Vol.
Glynn, M.A. (1994). "Effects of Work Task 29, pp. 208-226.
Cues and Play Task Cues on Information Pro- Rubinstein, R., and Hersh, H. (1984). The Hu-
cessing, Judgment, And Motivation," Journal man Factor'. Designing Computer Systems for
of Applied Psychology, Vol. 79, pp. 34-45. People, Burlington, MA: Digital Press.
Glynn, M. A., and Webster, J. (1992). "The Salancik, G. R., and Pfeffer, J. (1977). "An
Adult Playfulness Scale: An Initial Assess- Examination of Need-Satisfaction Models of
ment," Psychological Reports, Vol. 71, pp. 83- Job Attitudes," Administrative Science Quar-
103. terly, Vol. 22, pp. 427-456.
Heinssen, R., Glass, C., and Knight, L. (1987). Salancik, G. R., and Pfeffer, J. (1978). "A
"Assessing Computer Anxiety: Development Social Information Processing Approach to
and Validation of the Computer Anxiety Rating Job Attitudes and Task Design," Administra-
Scale," Computers in Human Behavior, Vol. tive Science Quarterly, Vol. 23, pp. 224-243.
3, pp. 49-59. Sandelands, L. E. (1988). "Effects of Work and
Hom, P. W., Griffith, R. D., and Sellaro, C. L. Play Signals on Task Evaluation," Journal of
(1984). "The Validity of Mobley's (1977) Applied Social Psychology, Vol. 18, pp. 1032-
Model of Employee Turnover," Organizational 1048.
Behavior and Decision Processes, Vol. 34, pp. Sandelands, L. E., and Buckner, G. C. (1989).
141-174. "Of Art and Work: Aesthetic Experience and
Kleinschrod, W. (1988). "The Trend to Elec- the Psychology of Work Feelings," Research
tronic Training," Administrative Management, in Organizational Behavior, Vol. 11, pp. 105-
Vol. 49, pp. 29-33. 131.
Kulik, C. T., and Perry, E. L. (1994). "Heuristic Schuck, G. (1985). "Intelligent Technology, In-
Processing in Organizational Judgments," In telligent Workers: A New Pedagogy for the
L. Heath, R. S. Tindale, J. Edwards, E. J. High-Tech Work Place," Organizational Dy-
Posavac, F. B. Bryant, E. Henderson-King, Y. namics, Vol. 14, pp. 66-79.
Suarez-Balcazar, and J. Myers (Eds.), Appli- Staw, B. M. (1984). "Organizational Behavior:
cations of Heuristics and Biases to Social A Review and Reformulation of the Field's
Issues, Vol. 3. New York: Plenum Press, pp. Outcome Variables," Annual Review of
185-204. Psychology, Vol. 35, pp. 627-666.
Martocchio, J. J. (1992). "Microcomputer Us- Tang Li-Ping, T., and Baumeister, R. F. (1984).
age as Opportunity: The Influence of Context "Effects of Personal Values, Perceived Sur-
in Employee Training," Personnel Psycho- veillance, and Task Labels on Task Prefer-
logy, Vol. 45, pp. 529-552. ence: The Ideology of Turning Play into
Martocchio, J. J., and Webster, J. (1992). Work," Journal of Applied Psychology, Vol.
"Effects of Feedback and Cognitive Playful- 69, pp. 99-105.
ness on Performance in Microcomputer Soft- Turnage, J. J. (1990). "The Challenge of New
ware Training," Personnel Psychology, Vol. Workplace Technology for Psychology,"
45, pp. 553-578. American Psychologist, Vol. 45, pp. 171-178.
Miller, S. (1973). "Ends, Means, and Galumph- Webster, J., Heian, J. B., and Michelman, J. E.
ing: Some Leitmotifs of Play," American Anth- (1990). "Computer Training and Computer
ropologist, Vol. 75, pp. 87-98. Anxiety in the Educational Process: An Ex-


perimental Analysis," Proceedings of the
Eleventh International Conference on Informa-
tion Systems, Copenhagen, Denmark, pp.
171-182.
Webster, J., and Martocchio, J. J. (1992).
"Microcomputer Playfulness: Development of
a Measure with 'Workplace Implications," MIS
Quarterly, Vol. 16, pp. 201-226.
Webster, J., and Martocchio, J. J. (1993).
"Turning Work into Play: Implications for Mi-
crocomputer Software Training," Journal of
Management, Vol. 19, pp. 127-146.

About the Authors

Elissa L. Perry is an assistant professor of
labor and industrial relations at the University of
Illinois at Urbana-Champaign. She earned her
M.S. and Ph.D. in organizational behavior and
theory from Carnegie Mellon University. Pro-
fessor Perry has research interests in the role of
demographic variables in human resource
judgments, social ,cognition and human resource
decision making, individual differences and the
training process, and sexual harassment. She
has published articles in journals such as
Academy of Management Review, Journal of
Applied Psychology, and Journal of Applied
Social Psychology'.
E-mail: e-perry@uiuc.edu

Deborah J. Ballou is an assistant professor in
the Department of Management at the Univer-
sity of Notre Dame. She earned her M.S. and
Ph.D. in information systems from Carnegie
Mellon University. Professor Ballou has re-
search interests in the study of information
acquisition and decision making as the basis for
designing more effective computer-based sup-
port. The applied areas in which she has con-
ducted most of her work are computer-based
performance monitoring, and decision making in
time-pressured, dynamic work environments.
She has published articles in journals such as
the Annals of Software Engineering, SIGCHI
Bulletin, and Journal of Applied Social Psycho-
Iogy.
E-mail: Deborah.J.Ballou.l@nd.edu


06 brugeres forventninger

More Related Content

Similar to 06 brugeres forventninger (20)

More from Nika Stuard (20)

06 brugeres forventninger