2012. Vol.3, No.12A, 1254-1258
Published Online December 2012 in SciRes (http://www.SciRP.org/journal/psych) http://dx.doi.org/10.4236/psych.2012.312A186
Copyright © 2012 SciRes.
Relations between Positive and Negative Attributional Styles and
Sales Performance as Moderated by Length of Insurance Sales
Experience among Japanese Life Insurance Sales Agents
Yukie Tsuzuki1, Tamao Matsui2, Takashi Kakuyama3
1Seijo University, Tokyo, Japan
2Surugadai University, Saitama, Japan
3Tokyo Future University, Tokyo, Japan
Received August 10th, 2012; revised September 9th, 2012; accepted October 5th, 2012
Past studies have shown that a sales agent’s attributional style for positive and negative events is related
to sales performance. The aim of the present study was to examine an agent’s length of sales experience
as a moderator of the relation between sales performance and attributional style for positive/negative
events. 360 Japanese life insurance sales agents were assessed with attributional styles for positive and
negative events (would be referred to as positive attributional style and negative attributional style, re-
spectively), sales performance and their lengths of sales experience. It was found that relationship be-
tween sales performance and the two types of attributional styles was largely different depending upon
length of an agent’s insurance sales experience. Among “novices” (n = 183) whose lengths of insurance
sales experience were less than three years, sales performance was related significantly to negative at-
tributional style, whereas it was not related to positive attributional style. By contrast, among “veterans”
(n = 177) whose lengths of insurance sales experience were three years or more, sales performance was
related significantly to positive attributional style, whereas it was not related to negative attributional style.
Theoretical and practical implications are discussed.
Keywords: Attributional Style; Sales Performance; Japan
Attributional style is a cognitive personality variable that re-
fers to the habitual ways people explain their positive and nega-
tive life experiences (Abramson, Seligman, & Teasdale, 1978).
Abramson et al. suggested that the three attributional dimen-
sions of internality, stability and globality are crucial for ex-
plaining human helplessness and depression.
The internal-external dimension refers to the extent to which
an individual sees his or her life experiences being caused by
something about himself or herself (internal attributions), as
opposed to something about the situation (external attributions).
The stable-unstable attribution refers to the extent to which an
individual perceives his or her life experiences being caused by
nontransient factors (stable attributions), as opposed to transient
ones (unstable attributions). The global-specific dimension
refers to the extent to which an individual sees cause of his or
her life experience as being present in a variety of situations
(global attributions), as opposed to more circumscribed (spe-
cific attributions) (Peterson, Semmel, Baeyer, Abrahamson,
Metalsky, & Seligman, 1982).
It was also suggested that depression-prone individuals tend
to attribute negative life experiences to internal, stable, and
global factors and to attribute positive life experiences to ex-
ternal, unstable and specific factors (Abramson et al., 1978).
Seligman et al. (1979) examined attributional styles for de-
pressed and non-depressed individuals and found that depressed
individuals make attributions systematically different from
non-depressed individuals on all of the three attributional di-
mensions. Specifically, relative to non-depressed individuals,
depressed individuals attributed positive life events to external,
unstable, and specific causes, and negative life events to inter-
nal, stable and global causes.
Seligman (1990) defined optimistic attributional style as the
tendency to attribute positive events to internal, stable and
global factors and negative events to external, unstable and
specific factors. Pessimistic attributional style, on the other
hand, was defined as the tendency to attribute positive events to
external, unstable and specific factors and negative events to
internal, stable and global factors. Individuals with an optimis-
tic attributional style are more resilient when faced with unfa-
vorable events than individuals with a pessimistic attributional
style (Abramson, Seligman, & Teasdale, 1978).
Past studies have found that an individual’s attributional
style is also linked to work-related outcomes such as work per-
formance and turnover. In their study of attributions made by
newly recruited US insurance sales agents, Seligman and
Schulman (1986) reported that attributional style for negative
events (would be referred to as “negative attributional style”,
hereafter) correlated significantly with sales performance. Spe-
cifically, tendency to explain negative events with internal,
stable, and global causes predicted lower sales performance.
Agents who scored in the good (optimistic) half of negative
attributional style sold 29% more insurance in their first year
and sold 130% more insurance in their second year than did
agents who scored in the bad (pessimistic) half of negative
Y. TSUZUKI ET AL.
Corr and Gray (1996) examined the relationship of attribu-
tional style and work performance among 130 senior male sales
agents in a leading UK insurance company. In their study, it
was attributional style for positive events (would be referred to
as “positive attributional style”, hereafter), rather than attribu-
tional style for negative events, which was related to sales per-
formance. Specifically, tendency to explain positive events with
internal, stable, and global causes predicted higher sales per-
Proudfoot, Corr, Guest, & Gray (2001) reported that positive
attributional style was more strongly related to job motivation,
intention to quit, learned resourcefulness and psychological
strain, than was the negative attributional style. Furnham,
Sadka, & Brewin (1992) found that positive attributional style
was related to job motivation and satisfaction, whereas negative
attributional style was related neither to job motivation nor to
satisfaction. Using job-rated outcomes as events, Silvester,
Patterson, & Ferguson (2003) reported that positive attribu-
tional style was a better predictor of performance ratings and
job satisfaction than was negative attributional style.
Xenikou (2005) administered the Occupational Attributional
Style Questionnaire (Furnham, Sadka, & Brewin, 1992) to em-
ployees of various organizations such as banks, hospitals, in-
surance companies, and examined the relation between attribu-
tional style and job motivation. It was found that not only the
two attributional styles, but their interaction also was a predic-
tor of job motivation. It was also suggested that length of ex-
perience at a given organization had an impact on the relation-
ship between attributional style and job motivation. Specifically,
only in the case of employees with more than four years of
experience at their organization, negative attributional style was
related to lower levels of job motivation. Xenikou suggested
that becoming a “veteran” employee in an organization was
likely to make people more vulnerable to the negative effect of
the pessimistic negative attributional style on job motivation.
In summary, results of past studies have been inconsistent in
whether attributional style for positive events or that for nega-
tive events is more critical in predicting sales performance.
Whereas Seligman & Schulman (1986) found a good (optimis-
tic) attributional style for “negative” events to be implicated in
successful insurance sales, other studies (e.g., Corr & Gray,
1996) suggested that a good (optimistic) attributional style for
“positive” events was most predictive of high sales perform-
ance. This inconsistency may suggest an existence of a mod-
erator variable, where in some situations having a good attribu-
tional style for positive events is critical and in others having a
good attributional style for negative events is critical.
The present study was aimed to explore the impact of an
agent’s length of sales experience on the relation between posi-
tive/negative attributional styles and work performance among
Japanese life insurance sales agents. It was hypothesized that an
agent’s length of sales experience would moderate the rela-
tionship between attributional style and sales performance.
Specifically, among novices with a relatively short experience
in insurance sales, negative attributional style would be more
strongly related to sales performance than positive attributional
style would. Among veterans with a relatively long experience
in insurance sales, positive attributional style would be more
strongly related to sales performance than negative attributional
style would. The reason for these hypotheses is as follows.
By contrast to Xenikou’s observation, in Japan’s life insur-
ance sales, being a new sales agent would be more vulnerable
to the impact of negative attributional style. Novice sales agents
are new in the field and do not have “regulars”, and thus are
destined to repeatedly encounter failure, rejection, and indif-
ference from prospective clients. Reflecting these situations, the
turnover rate among life insurance sales agents in an early pe-
riod of employment is very high in Japan. About 80% of the
life insurance sales agents hired in Japan quit within two years
of sales experience, and only less than 10% remain in 10 years.
Under these circumstances, new sales agents whose work life
tends to be full of rejection and failure are likely to be more
vulnerable to the impact of attributional style for negative
events. Thus, for novices to survive in life insurance sales,
having a pessimistic negative attributional style puts them at a
vast disadvantage in terms of their work performance. There-
fore, among novices, work performance would be related pre-
dominantly to negative, but not positive, attributional style.
Situations are different for Japan’s veteran sales agents.
Typically, veterans do not have to encounter as frequent fail-
ures and rejections as novices do. Veterans tend to have their
own regular customers. Those regulars refer their sales agents
to their friends and acquaintances. Therefore, veterans may
obtain new clients through introduction by their old customers.
This reduces their chances to feel hurt and rejected by prospec-
tive customers. In addition, veterans usually have a thorough
knowledge of insurance commodities they sell and can offer
commodities that meet their customers’ needs, thereby resulting
in higher sales performance in general. Attributional style for
positive events would influence the degree to which they savor
their success. Thus, positive attributional style, rather than
negative attributional style, would be more relevant to the con-
tinuation of their success as veterans. Therefore, among veter-
ans, their performance would be related predominantly to their
positive, but not negative, attributional style.
In summary, the present study examines an agent’s length of
experience as a moderator of the relation between sales per-
formance and positive/negative attributional styles. More spe-
cifically, it was hypothesized that positive and negative attribu-
tional styles would differentially affect sales performance de-
pending on an agent’s length of experience in insurance sales.
Participants of this study consisted of 360 Japanese female
life insurance route sales agents of a nation-wide life insurance
company. Participants worked on a complex commission sys-
tem that was basically determined by their performance but is
complicated by changes in the rate of conversion based on the
number and value sold by the participants over the last three
months. Participants ranged in age from 22 to 78 years with a
mean of 47.2 years (SD = 11.09). The length of sales experi-
ence on their present jobs ranged from 1 to 40 years with a
mean of 8.6 years. About 80% of participants were married, the
remainder being unmarried or widowed.
Attributional style measure. Based on the Attributional Style
Questionnaire (ASQ; Peterson, Semmel, Von Baeyer, Abram-
son, Metalsky & Seligman, 1982), a Japanese version of at-
tributional style questionnaire was developed by the authors (A
written permission of Dr. Seligman was obtained). Majority of
Copyright © 2012 SciRes. 1255
Y. TSUZUKI ET AL.
items were based on the ASQ except for a few items that were
modified to adjust to Japanese settings. Two of the authors who
are Japanese natives but familiar with English language and
culture independently translated the items of the English origin-
nal ASQ. They decided on the best translation by joint consen-
sus. Next, a bilingual English native speaker was asked to in-
dependently translate the Japanese version back to English.
When the back translation differed from the English original,
the final translation was chosen by consensus between the au-
The questionnaire consisted of 12 hypothetical situations, six
positive (e.g., “You apply for a position that you want very
badly, and you get it”) and six negative (e.g., “You give an
important talk in front of a group and the audience reacts nega-
tively.”). Participants were asked to imagine the event happen-
ing to them and to fill in the most likely cause. They were then
asked to rate the cause they supplied, on a one-to-seven scale,
for internal (7) vs external (1), stable (7) vs unstable (1), and
global (7) vs specific (1). This measure yields two sets of scores:
composite positive attributional style and composite negative
attributional style. The composite positive attributional style
was the composite score for the six positive events, summing
across internal, stable and global dimensions. The score was
higher to the extent that participants attributed positive events
to more internal, stable, and global factors. The composite
negative attributional style was the composite score for the six
negative events summing across internal, stable and global
dimensions. The score was higher to the extent that participants
attributed negative events to more internal, stable, and global
Performance data. Performance was measured by the num-
ber of policies sold by participants for the last three months.
The data was obtained from the company records.
Participants were presented with a questionnaire consisting
of the attributional style measure and a few demographic ques-
tions (e.g., age, length of sales experience). They were asked to
write their individual code number assigned by the company on
the cover of the questionnaire so that their performance data
could be obtained from the company records afterward. Par-
ticipants were assured that their individual responses to the
questionnaire would not be disclosed to the company. After
completing the questionnaire, participants were asked to en-
close the questionnaire into the attached posted-envelope and
send it directly to the research team.
Table 1 presents the means and standard deviations of the
research variables and their intercorrelations. The means and
standard deviations of attributional styles were M = 14.8, SD =
2.33 for positive attributional style and M = 11.4, SD = 2.17 for
negative attributional style. The reliabilities, as estimated by
Cronbach’s alpha were modest: .70 for positive attributional
style and, .69 for negative attributional style, which were com-
parable to Seligman & Schulman’s (1986) data. There were no
significant differences in attributional style scores by length of
sales experience (positive attributional style r = .04, ns; nega-
tive attributional style r = –.07, ns), indicating that experienced
sales agents did not have a better attributional style compared to
Means, standard deviations, and intercorrelations of research variables.
M SD 1 2 3 4
1) Positive attributional
stylea 14.82.33 (.70)
2) Negative attributional
styleb 11.4 2.17 –.02 (.69)
3) Sales experiencec 9.2 9.53 .04 –.07
4) Performanced 7.3 3.78 .16** –.11* .52** (.65)
Note: N = 360. *p < .05. **p < .01. The numerals within parentheses are cronbach
reliabililty coefficients. aExplaining positive events with internal, stable, and
global causes; bExplaining negative events with internal, stable, and global causes;
cNumber of years of sales experience; dNumber of policies sold for the last three
relatively new agents. This was also consistent with data in
Seligman & Schulman’s study.
The correlation between positive attributional style and nega-
tive attributional style was –.02 (ns), thereby indicating that the
two attributional styles are independent. Some previous studies
(e.g., Peterson, 1991) also reported no relation between positive
and negative attributional styles. Length of sales experience and
performance were strongly related (r = .52, p < .01).
Attributional style for positive events was related signifi-
cantly and positively to sales performance (r = .16, p < .01).
Thus, insurance sales agents who attributed positive events to
internal, global and stable factors were more successful than
those who attributed positive events to external, specific, and
unstable factors. Using the median cutoff, agents were divided
into two groups according to their scores for positive attribu-
tional style. Agents who were in the optimistic half sold 11.5%
more insurance policies for the last three months than agents
who were in the less-optimistic half (t = 2.10, p < .05). This
result was consistent with that of Corr & Gray’s (1996) study.
Attributional style for negative events was related signifi-
cantly and negatively to related sales performance (r = –.11, p
< .05). Thus, insurance sales agents who attributed negative
events to external, specific, unstable factors were more suc-
cessful than those sales agents who attributed negative events to
internal, global and stable factors. Using the median cutoff,
agents who were in the optimistic half of the negative attribu-
tional style, sold 11.1% more insurance policies for the last
three months than agents who were in the less optimistic half (t
= 2.00, p < .05). This result was consistent with Seligman &
Schulman (1986)’s findings.
As the present study was concerned with a moderating effect
of length of insurance sales experience on relations between
work performance and two types of attributional style, partici-
pants were classified into two groups based on their length of
life insurance sales experience. One was named as a “novice”
group (n = 183), and the other was named as a “veteran” group
(n = 177). The novice group consisted of those whose length of
life insurance sales experience was less than three years, while
the veteran group consisted of those whose length of insurance
sales experience was three years or more. The company records
indicated that about 80% of newly hired employees quitted
their jobs within two years of their employment, whereas the
reminder tended to continue their jobs for very long years (M =
To determine the moderating effect of the length of life in-
surance sales experience on the relation between the two types
Copyright © 2012 SciRes.
Y. TSUZUKI ET AL.
of attributional style and performance, three-step multiple re-
gression was performed. In Step 1, performance was regressed
on positive and negative attributional styles and length of in-
surance sales experience, and then on the three variables and
their two-way interactions in Step 2. In Step 3, performance
was regressed on all the variables above entered and the three-
way interaction of two attributional styles and length of insur-
ance sales experience. Increments in R2 from Steps 1 to 3 were
estimated. Table 2 presents the results. It is indicated that R2
significantly increased from Step 1 to 3. These findings sug-
gest the possibility that positive and negative attributional
styles were related to sales performance differentially, depend-
ing upon the length of insurance sales experience.
As the three-step multiple regression supported length of
sales experience as a moderator of the relation between the two
types of attributional style and performance, subgroup correla-
tional analyses were performed. Correlations of positive and
negative attributional styles to performance were computed for
each of the two groups.
Table 3 indicates that for the novice group, performance was
correlated significantly and negatively with negative attribu-
tional style, r = −.22, p < .01, whereas it was not related to posi-
tive attributional style, r = .05, ns. By contrast, for the veteran
group, performance was correlated significantly and positively
to positive attributional style, r = .26, p < .01, whereas it was
not related to negative attributional style, r = −.04, ns. These
findings support the hypotheses that relations between per-
formance and positive and negative attributional styles would
be largely different depending upon an agent’s length of life
insurance sales experience.
The findings of the present study showed that the positive
and negative attributional styles differentially impact sales per-
formance between “novice” sales agents and “veteran” sales
agents. Specifically, among novices, negative attributing style
Three-step multiple regression predicting performance among japanese
life insurance sales agents.
Steps Variables added R2 ΔR2 df F p
Step 1 .33
Length of sales
Step 2 .35.02 1.352 7.64.001
A × B interaction
B × C interaction
A × C interaction
Step 3 .36.01 1.350 4.62.001
A × B × C interaction
Note: N = 360. *p < .05. **p < .01; aExplaining positive events with internal,
stable, and global causes; bExplaining negative events with internal, stable, and
global causes; c1 = novices whose life insurance sales experience was less than
three years. 2 = veterans whose life insurance sales experience was three years or
Correlations of positive and negative attributional styles to sales perfor-
mance for novice and veteran groups.
Correlation with performance (r)
(n = 183)
(n = 177)
Positive attributional stylea .05 .26**
Negative attributional styleb−.22** −.04
Note: N = 360. **p < .01. aExplaining positive events with internal, stable, and
global causes; bExplaining negative events with internal, stable, and global
impaired sales performance, whereas positive attributional style
had no relationship with sales performance. By contrast, among
veterans, positive attributing style was related to higher sales
performance, but negative attributional style had no relationship
with sales performance.
These findings seem to make sense considering situations in
which life insurance sales agents find themselves in Japan. In
life insurance sales, it is likely that new sales agents repeatedly
encounter failure, rejection, and indifference from prospective
clients. Therefore, new sales agents are vulnerable to a variety
of emotionally painful events. Under such circumstances, an
attributional style for negative events, rather than for positive
events, would exert strong influence on their sales performance.
If novice agents explain bad events with internal, stable and
global causes they would likely be too discouraged to live up to
their full potential.
For veteran sales agents, situations are likely to be quite dif-
ferent. Veterans usually have a number of regular customers.
Veterans are likely to get referrals from their old customers and
it gets much easier for veterans to contact prospective clients
through introduction by their regular customers. Thus, for vet-
erans there may be much smaller chances of outright failures
and rejections. In addition, the company records indicated that
the number of policies sold by veterans for recent three months
(M = 9.51) was about twice as much as the number of policies
sold by novices (M = 5.28). Under these circumstances, an
attributional style for positive events, rather than that for nega-
tive events, would be more relevant to their work performance
for veterans. Specifically, those with an attributional style that
fully appreciates positive events by explaining them with inter-
nal, stable, and global causes, would be at advantage. Internal-
izing and generalizing success would increase self-efficacy and
thus would be the key to continuous success for veterans.
The finding of this study may account for inconsistencies
observed in past studies. Whereas Seligman & Schulman (1986)
found attributional style for negative events to be critical in
predicting sales performance, some other studies (e.g., Corr &
Gray, 1996) suggested that attributional style for positive
events was most predictive of sales performance. This study has
shown that there is a third variable (i.e., length of sales experi-
ence) that moderates the relationship between sales perform-
ance and positive/negative attributional style.
The present study’s findings are in conflict with those of
Xenikou’s (2005) study. Xenikou found that an attributional
style explaining negative events with stable and global causes
had impaired job motivation only in the case of employees with
longer than four years’ employment, which was equivalent to
“veterans” in the present study. In contrast, the present study
Copyright © 2012 SciRes. 1257
Y. TSUZUKI ET AL.
Copyright © 2012 SciRes.
found that an attributional style internalizing and generalizing
negative events impaired sales performance only in the case of
novices with less than three years employment.
Although it is difficult to determine the exact cause of these
conflicting findings, at least two points should be taken into
consideration. First, there are differences in participants’ occu-
pations between the two studies. Participants of Xenikou’s
study were employees from various organizations such as banks,
hospitals, and insurance company, presumably including nurses
and office workers. The participants of the present study were
all female life insurance sales agents. Thus, work situations of
the participants are likely to be vastly different between
Xenikou’s study and the present study. For example, Xenikou
suggests that the longer employees work at the same organiza-
tion, the more difficult and harsh their challenges and obstacles
posed on them tend to become. By contrast, the longer Japanese
sales agents work, the less harsh their work tends to become as
they have already obtained their own regular customers. An-
other point to be considered is the difference in criterion mea-
sures used between Xenikou’s and the present study. Xenikou
(2005) used seven-item self-ratings of “intrinsic job motive-
tion,” whereas the present study used “the number of policies
actually sold” by agents for recent three months.
The result of the present study seems to suggest the impor-
tance of distinguishing attributional style for positive and nega-
tive events when applying the concept of attributional style in
real-life contexts. As positive and negative attributional styles
are not significantly correlated, they are relatively independent
dimensions. Thus, having an optimistic attributional style about
positive events does not necessarily mean having an optimistic
attributional style in terms of negative events. Someone who
thinks and acts like an optimist when he or she is in normal and
decent life conditions, may fall apart and start to think like a
total pessimist when he or she finds him/herself in difficult
situations. Thus, managers should not assume that someone is
either an optimist or pessimist. People’s attributional style can
change depending on their situations. For novice life insurance
sales agents, having a pessimistic attributional style about nega-
tive events has a detrimental effect on their work performance,
as their work life is likely to challenge them with many inci-
dents of failures and rejections. It is advisable that managers
should carefully monitor how their subordinates are attributing
their sales experiences, success or failure, and pay special at-
tention to their subordinates’ attributional style about negative
events. For those subordinates who tend to attribute their fail-
ures with internal, stable and global causes, managers should
encourage them to examine if their attribution is realistic and if
their attributional style can be changed into a more adaptive,
optimistic one. For veteran workers, it is advisable that manag-
ers encourage them to “own” and savor their success by attri-
buting their achievement with internal, stable and global causes.
Abramsom, L. Y., Seligman, M. E. P., & Teasdale, J. (1978). Learned
helplessness in humans: Critique and reformulation. Journal of Ab-
normal Psychology, 87, 49-74. doi:10.1037/0021-843X.87.1.49
Corr, P. J., & Gray, J. A. (1996). Attributional style as a personality
factor in insurance sales performance in the UK. Journal of Occupa-
tional and Organizational Psychology, 69, 83-87.
Furnham, A., Sadka, V., & Brewin, C. R. (1992). The development of
an occupational attributional style questionnaire. Journal of Organ-
izational Behavior, 13, 27-39. doi:10.1002/job.4030130104
Peterson, C. (1991). The meaning and measurement of explanatory
style. Psychological Inquiry, 2, 1-10.
Peterson, C., Semmel, A., Von Baeyer, C., Abramson, L. Y., Metalsky,
G. I., & Seligman, M. E. P. (1982). The attributional style question-
naire. Cognitive Therapy and Research, 6, 287-300.
Proudfoot, J. G., Corr, P. J., Guest, D. E., & Gray, J. A. (2001). The
development and evaluation of a scale to measure occupational at-
tributional style in the financial services sector. Personality and In-
dividual Differences, 30, 259-270.
Seligman, M. E. P. (1990). Learned optimism. New York: Knopf.
Seligman, M. E. P., Abramson, L. Y., Semmel, A., & Von Baeyer, C.
(1979). Depressive attributional style. Journal of Abnormal Psy-
chology, 88, 242-247. doi:10.1037/0021-843X.88.3.242
Seligman, M. E. P., & Schulman, P. (1986). Explanatory style as a
predictor of productivity and quitting among life insurance sales
agents. Journal of Personality a n d Social Psychology, 50, 832-838.
Silvester, J., Patterson, F., & Ferguson, E. (2003). Comparing two
attributional models of job performance in retail sales: A field study.
Journal of Occupational and Organizational Psychology, 76, 115-
Xenikou, A. (2005). The interactive effect of positive and negative
occupational attributional styles on job motivation. European Jour-
nal of work and organizational psycho logy, 14, 43-58.