Journal of Personality and Social Psychology

1993. Vol.64, No. 6.906-920 Copyright 1993 by the American Psychological Association, lnc
0022-3514/93/S3.00

Distinguishing the Experiences of Envy and Jealousy

W Gerrod Parrott and Richard H. Smith

This article describes the nature and significance of the distinction between the emotions of envy
and jealousy and reports 2 experiments that empirically investigated it. In Experiment 1, Ss re-
called a personal experience of either envy or jealousy. In Experiment 2, Ss read 1 of a set of stories
in which circumstances producing envy and jealousy were manipulated independently in a facto-
rial design. Both experiments introduced new methodologies to enhance their sensitivity, and both
revealed qualitative differences between the 2 emotions. Envy was characterized by feelings of
inferiority, longing, resentment, and disapproval of the emotion. Jealousy was characterized by fear
of loss, distrust, anxiety, and anger. The practical importance of this distinction, the reasons for its
confusion, and general issues regarding the empirical differentiation of emotions are discussed.

Envy and jealousy have long been regarded as distinct emo- The Conceptual Distinction Between Envy and Jealousy
tions. Philosophers from ancient times to the present have ar-
gued that envy and jealousy have distinct causes and experi- According to the traditional distinction, envy occurs when a
ences (Cicero, 45 B.c/1927; Neu, 1980; Russell, 1930; Solomon, person lacks another's superior quality, achievement, or posses-
1976; Spinoza, 1678/1985; Taylor, 1988). This distinction be- sion and either desires it or wishes that the other lacked it. It
tween envy and jealousy has been maintained by some contem- occurs when this shortcoming exists in a domain that is self-
porary social psychologists (e.g., Hupka et al., 1985; Silver & definitional (Salovey & Rodin, 1984), and it intensifies with
Sabini, 1978b; Wegner & Vallacher, 1977) as well as by other self-dissatisfaction and when superiority in alternative compari-
social scientists (e.g., Barker, 1987; Foster, 1972; Schoeck, 1969; son domains is lacking (R. H. Smith, Diener, & Garonzik,
Sullivan, 1953). Recent research suggests that this distinction is 1990). These circumstances do not produce a single, simple
maintained (at least implicitly) by laypersons as well (R. H. affect; rather, the experience of envy is best characterized as a
Smith, Kim, & Parrott, 1988). constellation of several distinguishable affective elements that
Increasingly, however, envy and jealousy are confounded by typically occur during episodes of envy. These affective ele-
laypersons and scholars alike (Schoeck, 1969). Furthermore, ments may include feelings of inferiority, longing, resentment
recent research has failed to find any qualitative differences in of the circumstances, and ill will toward the envied person,
the experience of these emotions (Salovey & Rodin, 1986). In sometimes accompanied by guilt, denial, or awareness of the
this article, we reexamine the distinction between envy and inappropriateness of the ill will (Parrott, 1991). Understanding
jealousy and argue for its importance to numerous domains of envy has been important to research on social comparison (Sa-
social psychology as well as to theories of emotion. The empiri- lovey & Rothman, 1991; Tesser, Millar, & Moore, 1988), life
cal differentiation of the emotional experiences of envy and satisfaction (R. H. Smith, Parrott, & Diener, 1991), self-evalua-
jealousy is undertaken in two studies that explore issues and tion maintenance (Tesser, 1988), self-presentation (Silver & Sa-
methods relevant to the general issue of empirically distinguish- bini, 1978a), and social justice (Nozick, 1974; Rawls, 1971;
ing emotions. R.H. Smith, 1991).
Jealousy, by contrast, necessarily occurs in the context of
relationships. It occurs when a person fears losing an important
W Gerrod Parrott, Department of Psychology, Georgetown Univer-
sity; Richard H. Smith, Department of Psychology, University of Ken- relationship with another person to a rival—in particular, los-
tucky. ing a relationship that is formative to one's sense of self (Hupka,
These findings have been reported at the 95th, 96th, and 100th An- 1991; Rorty, 1988, chap. 7). As with envy, episodes of jealousy
nual Conventions of the American Psychological Association (August- may involve a number of specific affects. For jealousy these
September 1987, New York; August 1988, Atlanta, Georgia; and Au- include fear of loss, anxiety, and suspiciousness and anger about
gust 1992, Washington, DC; respectively)—those of Experiment 1 in betrayal (Hupka, 1984; Mathes, Adams, & Davies, 1985;
1987 and 1988 and those of Experiment 2 in 1992. Parrott, 1991; White, 1981 a). An understanding of jealousy has
Preparation of this article was supported in part by a Georgetown been a concern in the study of relationships (Buunk, 1991;
University Junior Faculty Research Fellowship to W Gerrod Parrott Mathes et al., 1985), romantic love (Mathes, 1986), personality
and by National Institute of Mental Health Grant R03 MH50044-01 to
Richard H. Smith.
(Bringle & Buunk, 1985; White, 198 lb), and clinical psychology
We gratefully acknowledge the helpful comments of Peter Salovey (Barker, 1987; Baumgart, 1990; White & Mullen, 1989).
and two anonymous reviewers on a draft of this article and the assis- Several factors have promoted the confounding of envy and
tance of Sung Hee Kim in designing Experiment 2. jealousy. These emotions are sometimes confused because of a
Correspondence concerning this article should be addressed to W linguistic ambiguity in English, namely, that the word jealousy
Gerrod Parrott, Department of Psychology, Georgetown University, can mean either jealousy or envy, whereas the word envy is fairly
Washington, DC 20057. unambiguous (see Smith et al., 1988). Some authors have main-
906
 ENVY AND JEALOUSY 907

tained that the preference in this century for using the term expected to be characterized by anxiety, distrust, and anger.
jealousy to refer to envy is motivated by discomfort with envy's Thus, there are theoretical reasons to expect envy and jealousy
moral connotations (Schoeck, 1969). to have the qualitative differences described in the traditional
A second cause of conflation of envy and jealousy is the distinction.
frequency of their co-occurrence. When a person's romantic Yet, these complex emotions do appear to have some overlap:
partner gives attention to an attractive rival, that person may both can involve some form of hostility (envy may produce
both be jealous of the special relationship with the partner and resentment and rancor; jealousy may produce anger over
also envious of the rival for being so attractive (Schmitt, 1988). betrayal), and both can involve some form of lowered self-es-
Although envy may easily occur without jealousy, jealousy of- teem and sadness (envy because of inferiority and longing and
ten is accompanied by envy. jealousy because of rejection and loss). Still, even these shared
The linguistic ambiguity of envy and jealousy and the fre- aspects may lead to distinguishable experiences. Envious hostil-
quency of their co-occurrence are pervasive sources of confu- ity is rarely socially sanctioned, whereas jealous hostility has
sion for both researchers and laypersons. Once these two prob- usually been accorded greater legitimacy (Schoeck, 1969; R. H.
lems are appreciated, however, the differences between envy Smith, Parrott, Ozer, & Moniz, 1992; Stearns, 1989). Envious
and jealousy become quite clear. Envy occurs when another has hostility is therefore less likely to entail righteous anger and
what one lacks, whereas jealousy is usually concerned with the more likely to be accompanied by a sense of disapproval by
loss of a relationship that one already possesses. Envy involves others. Likewise with lowered self-esteem—inferiority may re-
two elements (oneself and a person to whom one compares sult in a subjective experience different from that of rejection.
poorly), whereas jealousy requires three (oneself, a partner with The lack of evidence for such differences in Salovey and Ro-
whom one has a relationship, and a rival to whom one fears that din's (1986) findings may be attributable to any offiveaspects
this relationship will be lost).1 Envy involves comparing poorly of their research. First, the dependent variables may have been
with others on characteristics that are important to oneself, less than optimal for capturing the differences between these
whereas jealousy involves fear of rejection by another in prefer- emotions. In their first experiment, subjects were asked to rate
ence to someone who may be one's inferior in all other respects situations on only four variables, one being the confounded
(Neu, 1980). item "jealousy/envy" and the others being anger, sadness, and
embarrassment. In the second and third experiments more de-
pendent variables were included, but even these items may not
Empirical Differentiation of Experiences of Envy and have fully captured the hypothesized affective differences be-
Jealousy tween envy and jealousy; items measuring envy's most distinc-
The different conditions under which envy and jealousy tive features (longing, inferiority, and unsanctioned ill will)
arise would lead one to expect that they are experienced differ- were not included in the item pool, for example. Second, the
ently and that researchers should carefully distinguish between authors' choice of variable-reduction methods tended to ob-
them. It is therefore quite surprising that a recent investigation scure whatever subtlety the measures might have afforded. For
failed to discover any qualitative difference between them. Sa- example, in the third experiment, factor analysis combined all
lovey and Rodin (1986) reported three experiments in which variables measuring the subject's emotions into a single factor.
subjects read descriptions of situations eliciting envy or jeal- Third, the fact that jealousy is frequently accompanied by envy
ousy and then reported the thoughts and feelings that they imag- could have weakened the contrast between the envy and jeal-
ined would occur. In all three studies the authors concluded ousy conditions. Fourth, the manipulations of envy and jeal-
that the differences between the experiences of envy and jeal- ousy may have obscured the differences between the two emo-
ousy were more quantitative than qualitative—jealousy seemed tions. In thefirstand second studies, for example, subjects were
the more intense of the two, but different patterns of emotion presented with terse, abstract descriptions of situations that
and thought did not emerge in the data analysis. The authors may not have given rise to vicarious experience of the emotions
concluded that it was more useful to consider envy and jealousy involved.
as different types of situations that both produced essentially Finally, it may be that the issue of intensity is important for
the same experience than to distinguish envy and jealousy as distinguishing envy and jealousy and, more generally, for the
affective experiences. issue of how one emotion should be differentiated from an-
There are reasons to be puzzled about this conclusion and other. Previous research has found that jealousy is often more
concerned about its implications. Many theories of emotion intense than envy (Salovey & Rodin, 1986; R. H. Smith et al,
propose that cognitive focus is an important determinant of 1988). This difference in intensity may tend to obscure differ-
emotional experience (e.g., Frijda, 1986; Ortony, Clore, & Col- ences in quality. We contend that the most appropriate measure
lins, 1988; Parrott, 1988; Roseman, 1984; C. A. Smith & Ells- of the quality of affective experience is the relative salience of
worth, 1985). Inasmuch as situations inducing envy and jeal- the components of each. That is, if the experience of jealousy is
ousy give rise to different cognitive assessments, these theories characterized by a high level of certain affects and beliefs rela-
would predict that different affective responses should result.
Envy, because of its focus on the existence of and reasons for
1
inferiority, would be expected to be characterized by longing, Both emotions, however, can involve nonhuman entities, as when
self-diminishment, and resentment, whereas jealousy, because one is jealous because of the attention one's spouse devotes to a hobby
of its focus on the threat of loss of another'sfidelity,would be or when one is envious of the carefree life of one's dog.
 908 W GERROD PARROTT AND RICHARD H. SMITH

tive to others, then these relative differences should be present of loss, distrust, righteous anger over betrayal, and uncertainty
at different intensities of the emotion. Thus, a sensitive compari- about the circumstances.
son between the experiences of envy and jealousy can be made
only when the emotions are of similar intensities, or when the
Method
data have been adjusted to compensate for differences in over-
all intensity. Subjects. Subjects were 78 male and 78 female undergraduates at
the University of Illinois at Urbana-Champaign. They received credit
toward an introductory psychology course requirement for participat-
Overview of the Research ing.
Procedure and materials. The experiment was conducted in a small
The following two experiments investigated envy and jeal- classroom in which groups of 3 to 8 subjects participated at a time.
ousy to determine the affective character of each, to find Subjects were instructed to write a description of a situation in which
whether they differ in the ways predicted by theories of emo- they had felt either "strong envy" or "strong romantic jealousy." Equal
tion, and to examine the role of overall intensity in determining numbers of subjects were assigned at random to one or the other condi-
the similarity of two emotions. Manipulations, measures, and tion, and subjects were aware of only their own condition. Because of
analyses were designed to provide a sensitive and accurate as- the ambiguity of the word jealousy, romantic jealousy was specified to
sessment of the two emotions. In Experiment 1, subjects' recol- ensure that subjects in this condition did not recall an experience of
envy instead. Research has shown that virtually all subjects recalling
lections of actual personal experiences of envy and jealousy cases of jealousy recall romantic jealousy rather than some other type,
were examined, and a statistical adjustment for differences in such as sibling jealousy (R. H. Smith et al., 1988); therefore, specifying
overall affective intensity was introduced. In Experiment 2, vi- romantic jealousy should only have served to clarify the instructions,
gnette methodology was used that, for the first time, manipu- not to have affected the type of jealousy recalled.
lated envy and jealousy factorially, a technique designed to Subjects were told the importance of recalling the situation vividly
overcome insensitivity caused by the frequent co-occurrence of and of providing details of the emotion and the circumstances that led
envy and jealousy. This technique also permitted manipulation up to it. The anonymity of their descriptions was emphasized. Most
of these affective states while avoiding any use of their seman- subjects completed writing their accounts in about 20 min.
ticly imprecise verbal labels. Subjects then completed 59 randomly ordered questionnaire items
about the experience they had described. These items were con-
structed to describe those feelings and emotional thoughts proposed
to be characteristic of jealousy and envy; care was taken to include
Experiment 1: Recalled Experiences of Envy and items theorized to distinguish between these affective states, although
Jealousy some items applied to both emotions. The items are listed in the first
column of Table 1. Subjects rated on a 9-point scale how characteristic
Subjects were asked to recall an experience either of envy or each item was of their particular experience. The lowest point of this
of jealousy and to write a detailed description of it before an- scale was labeled not at all characteristic, the midpoint moderately char-
swering questionnaire items. Although this method is suscepti- acteristic, and the highest point very characteristic.
ble to the limitations inherent to retrospective accounts, it has
the advantage of asking subjects to report on an actual emo-
tional episode rather than a hypothetical one. The question-
Results
naire items were constructed to tap feelings and emotional Almost all subjects wrote detailed, evocative accounts of
thoughts theorized to be characteristic of envy and jealousy. events of the type requested, suggesting that subjects had in
These data were analyzed in a manner that preserved the dis- mind a detailed recollection when they rated their experience
tinctions that these items made. on the questionnaire. Five subjects, however, wrote short, vague
To address the issue of how to compare emotions that differ accounts, and their data (4 from the jealousy condition and 1
in overall intensity, the data were examined both in their raw from the envy condition) were excluded from the analysis. Also
form and after a transformation. Because it is the relative sa- excluded were 2 subjects who claimed never to have experi-
lience of affective components that is hypothesized to deter- enced the requested emotion. Of the 149 remaining subjects, 73
mine the quality of a complex affective experience, the desired had been assigned to the jealousy condition and 76 to the envy
transformation is one that standardizes each subject's ratings condition. The proportion of men and women in the two
with respect to their elevation and scatter, preserving only the groups remained closely balanced.
shape of the profile of ratings (Cronbach & Gleser, 1953). If the Adjustment for intensity. As expected, jealousy tended to
same components are salient in the experiences of both envy involve significantly more intense affect than did envy. Jealousy
and jealousy, there should be no differences between the two experiences were rated higher on average than envy experiences
groups on these transformed scores. On the other hand, if dif- on 53 of the 59 scales (see Table 1). The average rating was 5.81
ferences are found, it would suggest that there are qualitative in the jealousy condition, versus 4.96 in the envy condition,
differences between envy and jealousy that are not attributable f(147) = 4.49, p<.0001.
to differences in intensity. We predicted that the following Thus, for the reasons discussed, subjects' ratings were ad-
would be more salient in envy than in jealousy: longing for justed for between-subjects differences in elevation and scatter
what another has, feeling inferior, harboring resentment and ill by subtracting each subject's mean rating from each of his or her
will, and feeling that this ill will is wrong or unsanctioned. individual ratings, then dividing this difference by the stan-
Predicted to be more salient in jealousy than in envy were fear dard deviation of the subject's ratings. The means of these
 ENVY AND JEALOUSY 909

Table 1
Mean Ratings of Questionnaire Items in Envy and Jealousy Conditions of Experiment 1, With
Results of Cluster Analysis and Multivariate Analysis of Variance of Cluster Scores
Untransformed Transformed
scores scores Discriminant
function
Items (by cluster) Envy Jealousy Envy Jealousy loading

Cluster 1: Disapproval of feelings (a = .55) .597

Others would disapprove if they knew
what I was feeling 4.45 3.33 -0.21 -1.08
Embarrassing to admit to 4.67 4.78 -0.11 -0.41
At first denied to myself that I felt this
emotion 3.75 3.86 -0.47 -0.81
Feeling sinful 2.41 2.63 -1.01 -1.37
Guilt over feeling ill will toward someone 4.83 4.63 -0.02 -0.49
This emotion came on unexpectedly 4.57 5.19 -0.17 -0.29
Cluster 2: Longing (a = .53) .439
Longing for what another has 7.75 7.27 1.14 0.61
Feeling wishful 6.66 6.58 0.68 0.32
Feeling unlucky 4.97 5.32 -0.03 -0.20
Cluster 3: Motivation to improve .313
Motivated to improve myself 5.63 5.58 0.32 -0.12
Cluster 4: Degradation (a = .31) .274
Feeling degraded 5.74 5.82 0.32 0.00
Humiliated in front of others 3.78 4.30 -0.48 -0.63
Cluster 5: Irrationality (a = .31) .179
Thinking irrationally 4.42 5.56 -0.23 -0.13
Frustration 7.17 7.36 0.89 0.64
Feeling critical of others 5.39 5.86 0.21 0.01
Cluster 6: Anger (a = .69) .051
Angry 6.32 6.95 0.56 0.50
Bitter 5.93 6.18 0.38 0.17
Hate 4.30 5.52 -0.25 -0.14
Feeling hostile 4.39 5.33 -0.24 -0.24
Resentful 6.13 6.93 0.48 0.47
Intense feeling 6.01 7.37 0.39 0.64
Feeling unfairly treated by life 5.82 5.84 0.31 0.03
Unappreciated 5.21 5.97 0.09 0.06
Cluster 7: 111 will (a = .53) .022
Urge to get even 3.54 4.88 -0.54 -0.43
I would feel some pleasure if the person
who caused the emotion experienced
some failure 4.18 4.71 -0.31 -0.47
I would feel unhappy if the person who
caused this emotion experienced some
good luck 4.03 4.97 -0.38 -0.31
Feelings lasted a long time 5.24 6.08 0.09 0.04
I had a right to feel this way 5.61 6.34 0.27 0.22
Cluster 8: Pettiness -.011
Feeling petty 3.93 4.75 -0.45 -0.44
Cluster 9: Low self-esteem (a = .79) -.026
Lacking confidence 5.45 6.73 0.23 0.38
Self-doubt 5.32 7.01 0.11 0.50
Insecure 5.13 6.53 0.07 0.29
Self-conscious 4.74 6.41 -0.07 0.25
Dissatisfied with myself 5.42 5.90 0.17 0.05
Feeling like a failure 4.62 5.63 -0.16 -0.05
Privately ashamed of myself 4.45 4.19 -0.24 -0.70
Aware of my inferior qualities 5.00 5.55 0.03 -0.08
Feeling inferior 6.03 6.05 0.47 0.11
Cluster 10: Unhappiness (a = .59) -.066
Emotional pain 6.30 7.53 0.54 0.72
Hurt feelings 6.53 7.77 0.64 0.83
Unhappy 7.05 7.52 0.83 0.74
Upset 6.80 7.45 0.73 0.69
Anguish 5.71 6.37 0.31 0.23
(table continues)
 910 W. GERROD PARROTT AND RICHARD H. SMITH

Table 1 {continued)
Untransformed Transformed
scores scores Discriminant
function
Items (by cluster) Envy Jealousy Envy Jealousy loading
Cluster 11: Helplessness (a = .50) -.132
Helpless 5.09 6.21 0.02 0.18
Not in control 4.36 5.26 -0.26 -0.19
Cluster 12: Loneliness (a = .23) -.277
Lonely 4.63 6.81 -0.11 0.42
Left out 6.17 6.84 0.48 0.42
Cluster 13: Uncertainty (a = .48) -.343
Confusion 4.97 6.27 -0.01 0.16
Uncertain 4.32 6.11 -0.25 0.15
Cluster 14: Fear (a = .60) -.420
Fearful 2.58 3.79 -0.97 -0.84
Worried 3.29 5.08 -0.65 -0.29
Afraid of a possible loss 3.68 6.11 -0.51 0.17
Threatened 3.28 5.00 -0.67 -0.33
Anxious 4.11 5.38 -0.33 -0.19
Apprehensive 3.87 4.68 -0.42 -0.47
Cluster 15: Distrust (a = .56) -.636
Betrayed 4.05 6.79 -0.37 0.42
Rejected 5.34 7.36 0.15 0.66
Distrustful 3.22 4.70 -0.69 -0.41
Suspicion 4.14 5.79 -0.33 0.00
Note. Items were rated on a 9-point scale. The discriminant function loading is the correlation between a
cluster score and the multivariate function discriminating envy from jealousy.

transformed scores for the envy and jealousy conditions are The pattern of differences corresponds closely to the tradi-
presented in Table 1. tional distinctions made between envy and jealousy, strongly
Variable reduction. Preliminary principal-componentsanal- suggesting that qualitative differences exist between these af-
ysis suggested the presence of 10 components, but, even after fective experiences. Five of the eight differences clearly corre-
rotation, all 10 remained bipolar, and interpretation was not
straightforward. To avoid these difficulties, data reduction was
accomplished by cluster analysis of variables. The average link-
age method of hierarchical agglomerative clustering was used,
using intervariable correlations as the measure of similarity
(Aldenderfer & Blashfield, 1984; Hartigan, 1990). Stopping the
clustering when the mean similarity between joined elements
ceased to exceed .10 resulted in 15 readily interpretable clusters.
Comparison with results of alternative clustering methods sug-
gested that the clusters were stable. Cluster membership, de-
scriptive labels, and reliabilities are shown in Table 1.
Cluster scores were calculated for each subject by averaging
the items in each cluster. The 15 cluster scores explained 52% of
the total variance of the 59 items. The means for the envy and
jealousy conditions, standardized to facilitate comparison, are
graphed in Figure 1.
Comparing the experiences of envy and jealousy. Figure 1
suggests a clear pattern of differences between envy and jeal-
ousy, which was confirmed by a one-way multivariate analysis
of variance (MANOVA), Hotelling's T2 = 90.29, F(14, 134) = Cluster
5.88, p < .0001, r)2 = .38. Results of subsequent univariate t tests Figure 1. Ratings of envy and jealousy experiences in Experiment 1
for each of the 15 clusters are indicated by the two horizontal expressed as standardized mean scores on clusters of the original vari-
lines in Figure 1; points falling outside the central region de- ables (adjusted for differences in intensity between subjects). (Pairs of
fined by these lines differed significantly, t(l47) > 1.98, p < .05 points falling outside the central region denned by the two horizontal
(two-tailed). Significant differences were found for 8 of the 15 lines differ significantly by a; test, df- 147, p < .05, two-tailed. Mot. =
clusters. motivation.)
 ENVY AND JEALOUSY 911

sponded to features predicted to distinguish envy from jeal- quent co-occurrence of envy and jealousy probably limited the
ousy. "Disapproval of feelings" and "longing" were more salient sensitivity of the present comparison of these emotions. Sec-
for envy than for jealousy. "Distrust," "fear," and "uncertainty" ond, the assymetry of this co-occurrence may help to explain
were more salient for jealousy than for envy. why the term jealousy has come to be more ambiguous than the
The remaining three differences, on inspection, proved to be term envy, a speculation we develop later in this article.
consistent with the traditional distinction. The single-item
cluster "Motivation to improve," more salient in envy than in Discussion
jealousy, correlated most highly with several items in the "dis-
approval of feelings" cluster. It seemed best understood as an- This experiment provided clear evidence of qualitative dif-
other manifestation of the unsanctioned nature of envy. ferences between the experiences of envy and jealousy. Long-
The cluster "loneliness" proved to be more salient in jealousy ing, disapproval of feelings, motivation to improve, and inferior-
than in envy. This difference is not emphasized in writings ity were more salient for envy than they were for jealousy. Jeal-
distinguishing the emotions so had not been explicitly pre- ousy, on the other hand, was characterized by a greater salience
dicted. The finding makes sense, however, in terms of the fear of distrust, fear, uncertainty, and loneliness. Jealousy, on aver-
of loss of a relationship or formative attention that has been age, was more intense than was envy. Once adjusted for inten-
theorized to be necessary for jealousy to occur. sity, the differences between the two emotions appeared quite
Finally, the prediction that envy would be distinguished pronounced.
from jealousy by a greater salience of inferiority was con- To a remarkable extent, these differences correspond to
firmed, but the sole failing of the cluster analysis was to obscure those proposed by theorists. Of the eight clusters yielding signif-
this finding. The item "feeling inferior" itself was significantly icant differences, none was discrepant with theory. Of the seven
more salient in envy than in jealousy. Its second-highest inter- clusters that did not differ significantly between envy and jeal-
item correlation was with the item "degraded," and the cluster- ousy, five would not have been predicted to differ: Envy and
ing algorithm would have combined them had "feeling infe- jealousy have not usually been considered to differ with respect
rior" not correlated marginally more highly with "aware of my to irrationality, pettiness, unhappiness, or helplessness. The
inferior qualities." Thus, there was support for the prediction of nonsignificance of the cluster "low self-esteem" suggested that
greater salience of inferiority in envy and for interpreting the envy and jealousy do not differ overall in terms of self-esteem,
cluster "degradation" as capturing that salience. even though one aspect of it, inferiority, did appear to be more
In sum, the present experiment yielded support for all but salient in envy. The other two clusters can be viewed as failing
one aspect of the traditional distinction between envy and jeal- to find predicted differences. The clusters "angry" and "re-
ousy. The sole failure was in distinguishing forms of hostility venge" contained items dealing with animosity. Envy was theo-
typical of envy (resentment and ill will) from those typical of rized to lead to resentment and ill will, whereas jealousy was
jealousy (justified anger). Hostility was characteristic of both theorized to lead to anger; the distinction rests on the existence
emotions, but the present data provided no evidence that the of perceived transgression in the latter but not in the former.
hostility took different forms. Neither of these animosity clusters captured this distinction,
Assessment of co-occurrence of envy and jealousy: Content however.
analysis. Many authors have noted that jealousy can often be It is noteworthy that the differences between envy and jeal-
accompanied by envy, and such co-occurrence is one factor ousy were apparent only after the data were transformed to
increasing the difficulty of differentiating these emotions. The compensate for differences in overall intensity. The greater
co-occurrence of envy and jealousy has never been empirically average intensity of the episodes of jealousy masked differences
assessed, however, so a content analysis of the 149 written ac- in the salience of certain affects relative to others. The raw
counts in the present data set was undertaken. scores replicated Salovey and Rodin's (1986) finding that jeal-
Two judges (the authors) independently read and rated each ousy's greater average intensity makes it difficult to observe the
written account for the presence or absence of envy and jeal- differences between jealousy and envy. Compensating for be-
ousy as a salient part of the emotional response. Jealousy was tween-subjects differences in intensity was useful in detecting
operationally defined as "a focus on a threat to or a loss of a predicted patterns of differences between the emotions. This
relationship," envy as "a focus on someone who has something finding suggests that the issue of intensity may be important
the author wanted." The judges agreed on 88.6% of the accounts more generally in comparing other emotions.
when rating envy and on 97.3% of the accounts when rating The retrospective method used in this experiment ensured
jealousy. Disagreements were resolved by discussion. that subjects would report on actual emotional experiences, but
The results showed that jealousy did indeed include envy this method had its limitations as well. Not only were the ac-
quite frequently—58.9% of the accounts of jealousy were counts subject to distortions of memory, but also each subject
judged to include envy. The reverse was not true, however. Only recalled an experience that was unique in some respects, com-
10.5% of the accounts of envy also involved a threatened rela- posed of different subvarieties of envy and jealousy, and mixed
tionship, a significantly smaller proportion, x 2 0, N = 149) = with other emotions as well. As a result, there was necessarily a
38.71, p < .0001. Thus, a threat to a relationship frequently also good deal of variability within the envy and jealousy condi-
evokes a focus on someone who seems superior, but not vice tions, with a resulting loss of sensitivity to the differences be-
versa. tween these emotions. Further weakening the sensitivity of this
These findings are significant for two reasons. First, the fre- method is the fact that over half of the episodes of jealousy also
 912 W GERROD PARROTT AND RICHARD H. SMITH

involved envy. It would be desirable to isolate these two affective by each. Although Experiment 1 found that jealousy is on aver-
experiences more effectively. age more intense than envy, there certainly were some cases of
A further limitation of the present experiment is its reliance envy that were extremely powerful. If elicitors of envy could be
on the subjects' understanding of the words envy and jealousy. selected to be as strong as those of jealousy in these stories, it
Although this experiment was certainly not a study of word would be possible to compare the emotions for differences
meanings per se—subjects rated an emotional personal experi- without resorting to statistical adjustments.
ence, not the words envy or jealousy—it is nevertheless the case
that the situations that were recalled depended on subjects' un- Method
derstanding of the meaning of these terms, and some re-
searchers may prefer that the stimulus be unmediated by such Design. Envy and jealousy were manipulated independently. Envy
semantics. was manipulated by adjusting both the success of a rival and the suc-
The effect of overall affective intensity was controlled statis- cess of the protagonist on domains self-relevant to the protagonist.
tically in the present experiment. If our approach to distinguish- Previous research suggests that each component of this manipulation
is necessary for producing envy (Salovey & Rodin, 1984). Jealousy was
ing emotions is correct, there should be circumstances in which manipulated by having the protagonist see the rival interacting roman-
the predicted differences between the two emotions are appar- tically either with the protagonist's boyfriend or girlfriend or with an
ent without statistical adjustment, namely, when the overall unspecified person.
intensity of envy and jealousy is equivalent. Finally, there is an Two parallel sets of stories were created to ensure that findings
overall need for replication of thesefindings,preferably using a would generalize to more than one situation. Separate versions of each
different methodology. The following experiment was per- story were prepared for male and female subjects by appropriately
formed to address all these issues. adjusting the gender (and, occasionally, the descriptions) of the protag-
onists. Thus, this experiment used a 2 (high and low envy) x 2 (high
and low jealousy) X 2 (stories) X 2 (sexes) between-subjects factorial
Experiment 2: Ratings of Situations Eliciting Envy and design.
Jealousy Subjects. Volunteers were recruited from an undergraduate intro-
ductory psychology class at Georgetown University. All 97 students (57
In this experiment envy and jealousy were manipulated inde- women) present agreed to participate in the experiment. Subjects re-
pendently in a set of detailed stories that were rated by subjects. ceived no compensation for their participation.
This method ensured that subjects within a condition all rated Materials. Situations typically evocative of envy and jealousy in an
the same situation and that only the elements of experimental undergraduate population were suggested by the accounts collected in
interest differed between conditions. Subjects were presented Experiment 1 and by the research of Salovey & Rodin (1986, Experi-
with specific situations, not with the general concepts of envy or ment 1). Two situations with potential for eliciting strong envy and two
jealousy, so this method eliminated any issue of semantic inter- for eliciting strong jealousy were selected and combined to create two
pretation present in the previous experiment. This experiment sets of story contexts. One story involved a college freshman trying out
examined whether the differences found in Experiment 1 repli- for the varsity tennis team, and the other involved a college senior
applying to medical school. In both stories, in the high-envy versions a
cate in a new paradigm. Because the vignette paradigm was
rival excels while the protagonist fails, whereas these fates are reversed
used by Salovey and Rodin (1986, Experiment 3), the present in the low-envy versions. In the high-jealousy versions, the protagonist
experiment examined whether differences between envy and later sees the rival flirting with the protagonist's boyfriend or girl-
jealousy can be found using a paradigm that previously failed friend. In one story they kiss goodbye; in the other story they plan a
to find them. dinner date. In the low-jealousy versions, the protagonist observes a
As demonstrated in Experiment 1, envy frequently occurs in similar scene, but the identity of the rival's friend is not specified. The
jealousy situations, and this confound has decreased the sensi- factorial combination of these plots produced four vignettes that
tivity of all previous comparisons between these emotions. In ranged in length from 564 to 590 words for the tennis story and from
654 to 687 words for the medical school story.
the present experiment we tried to reduce this confound by
manipulating envy and jealousy in a factorial design. Stories The dependent variables consisted of 34 Likert items. Most of these
items were among those that successfully distinguished envy from jeal-
were constructed such that envy and jealousy could each be ousy in Experiment I2; 20 were taken verbatim from that experiment,
manipulated separately. Thus, even though it was not entirely and another 7 were reworded to improve clarity. Five items were added
possible to eliminate envy from jealousy, it was at least possible to investigate aspects neglected in the previous experiment. Three of
to have one condition in which jealousy was accompanied by these supplemented items measuring the unfairness and inferiority
relatively little envy and another condition in which jealousy characteristics of envy ("resentment," "depressed," and "feeling me-
was accompanied by relatively much envy. It was expected that diocre"). Two items supplemented those measuring the justifiable
the resulting patterns of variation between conditions would be anger characteristic of jealousy ("justified in having these feelings"
superior to previous research in identifying which dependent and "vengeful"). The 2 final additions were the items "envious" and
variables respond to envy, which to jealousy, and which to an
interaction between the two. Dependent variables were mostly 2
When this experiment was designed, the results of Experiment 1
derived from those in the previous experiment, but some im-
had been analyzed using a different method of data reduction, one that
provements were undertaken, in part to attempt again to cap- obscured the significance of the items "lonely" and "feeling degraded"
ture the distinction between envious and jealous hostility. in distinguishing envy from jealousy. These items were therefore not
In writing the stories, care was taken to manipulate envy and included among the dependent variables of Experiment 2 even though
jealousy so that a strong emotional reaction would be suggested subsequent analysis suggested they should have been.
 ENVY AND JEALOUSY 913

"jealous," which were added to see how subjects would rate the emo- All 3 of the resentment items from the envy component also
tion names themselves. In all, 16 items measured constructs related to loaded on jealousy, and 5 of the anger items from the jealousy
envy, and 18 items measured jealousy. Each item was accompanied by a component also loaded on envy. What seems to distinguish
9-point rating scale in which the low point was labeled not at all and the these latter 5 items from those anger items that did not correlate
high point was labeled extremely. The 34 items were arranged in a with the envy component is ambiguity over the justifiability of
random order. A booklet that had a cover sheet providing instructions, the anger; items such as "hostility toward rival," "hurt," and
followed by one of the stories and then the response items, was pre- "angry at a particular person" loaded on both components,
pared for each subject.
whereas the items "betrayed," "vengeful," and "justified in hav-
Procedure. The experiment was conducted in a large classroom.
Subjects were assigned randomly to an envy, jealousy, and story condi-
ing these feelings" loaded only on the jealousy component.
tion and were asked to read instructions on the cover sheet, which This pattern supported the predicted distinction between envi-
requested they read the story carefully, take the point of view of the ous and jealous hostility—angry feelings loaded onto the envy
protagonist, and imagine how that person would feel and what the component only to the extent that they did not specify an objec-
events would mean to him or her. After finishing the story subjects tive justification, whereas all types of anger loaded onto the
were told to turn the page and answer the questionnaire items to indi- jealousy component, especially those that connoted an objec-
cate how the protagonist would feel at the end of the story. Subjects tive justification for the anger.
took about 15 min to complete the experiment. Also loading onto both the jealousy and envy components
were four items related to low self-esteem. As in thefirstexperi-
ment, and as predicted, items related to feelings of inferiority
Results
could be distinguished from low self-esteem in general and
Overview. Data analysis consisted of three stages. Principal- were much more highly correlated with envy than with jeal-
components analysis was used to summarize patterns of covar- ousy.4
iation among 32 of the questionnaire items and to reduce the Analysis of the factorial design. Standardized principal-
dimensionality of the data.3 (The items "jealous" and "envious" components scores were computed for each subject. Initial com-
were excluded from this analysis.) In the second stage the facto- parisons between the high-envy-low-jealousy and the low-
rial design was analyzed by submitting the orthogonal princi- envy-high-jealousy conditions were made to see whether there
pal-components scores to an analysis of variance (ANOVA). In was any tendency for jealousy or envy to be more intense over-
the third stage, the items "jealous" and "envious" were analyzed all. In contrast with the previous experiment, no significant
in a separate MANOVA and correlated with the principal-com- difference was found using either the component scores or the
ponents scores. ratings of the original items.5 It was concluded that the stories
Principal-components analysis. Four of the 97 subjects
failed to answer one or more of the questionnaire items, so data
analysis was based on the 93 subjects having complete data. 3
Principal-components analysis was used for these data but not for
Numerous high correlations between questionnaire items, those of Experiment 1, because the factorial manipulation of envy and
good sampling adequacy (Kaiser's measure equaled 0.91), and jealousy in the present experiment was expected to result in a small
the small number ofexpected components made principal com- number of unipolar components, whereas the unconstrained accounts
ponents analysis appropriate despite the relatively small num- collected in Experiment 1 produced many more dimensionsof variabil-
ber of subjects. Inspection of the scree plot suggested that three ity and more complex patterns of covariation. This principal-compo-
components should be extracted. The first six eigenvalues were nents analysis was performed on the raw data, which conflated within-
14.9, 4.5, 2.0,1.2,1.1, and 0.9. The first three components ac- cell and between-cells variance. This approach was chosen as the best
counted for 67% of the total variance. Following varimax rota- compromise between entering only within-cell variance (which would
tion, the three components accounted for 33.2%, 22.5%, and not reflect the effects of the manipulation) and using only between-
11.3% of the variance, respectively. Orthogonal rotation was cells variance (which would result in unstable correlations because of
the small number of cells). A further advantage of this procedure was
selected so that component scores would be uncorrelated. to increase comparability with Salovey and Rodin's (1986) analyses,
The three components proved easily interpretable and conr which also analyzed the raw data.
sistent with the hypothesis that envy and jealousy are distinct 4
The explanation for why the item "rejected" loaded on the envy as
emotional experiences. The loadings are reported in Table 2. well as on the jealousy component in this but not the previous experi-
Items loading highly on the first component pertained to four ment is surely that the present manipulation of envy involved rejection
aspects of jealousy: distrust, anxiety, anger, and fear of rejec- from medical school or the tennis team. The item "wishful" loaded
tion or loss. Items loading highly on the second component negatively on the jealousy component because of strong negative
pertained to three aspects of envy: inferiority, resentment, and correlations with suspiciousness, anxiety, and betrayal.
5
longing. The third component may be characterized as a di- Comparisons were made between the high-envy-low-jealousy cell
mension of disapproval: Most of the items pertain to disappro- and the low-envy-high-jealousy cell using three-way ANOVAs in
val of the person's feelings, but some items expressed a disap- which story and subject's sex were included as independent variables.
Using the mean of the three component scores as the dependent vari-
proval of the person. able, "pure envy" (M= 0.12) and "pure jealousy" (M= 0.17) were not
Of the 26 items loading most strongly on the jealousy or envy found to differ in overall emotional intensity, F(l, 40) = 0.12, ns. Like-
components, half also shared 10% or more variance with the wise, using the mean of the 32 questionnaire items as the dependent
other of these two components, and this overlap invites exami- variable, the difference between envy (M = 5.8) and jealousy (M = 6.2)
nation. Eight of these items pertained to anger or resentment. was not significant, F(\, 40)= 2.02, p> .15.
 914 W GERROD PARROTT AND RICHARD H. SMITH

Table 2
Means and Standard Deviations of Cells of the Envy X Jealousy Interaction and Loadings
on the First Three Principal Components: Experiment 2

Cells in the
Envy X Jealousy
interaction Loadings on the
first 3 principal
High envy- High envy- Low envy- Low envy- components
nigh low high low
Item jealousy 3 jealousyb jealousy 6 jealousy0 "Jealousy" "Envy" "Disapproval"

Betrayed
M 8.4 3.9 8.1 2.5 .92
SD 0.8 2.3 1.6 2.0
Distrustful
M 7.2 3.8 7.2 3.1 .88
SD 1.5 2.2 1.8 2.3
Suspicious
M 8.4 4.0 8.3 4.1 .87
SD 1.0 2.8 1.3 2.3
Worried
M 6.9 4.0 7.5 2.8 .86
SD 1.7 2.3 1.4 2.0
Enraged
M 7.6 4.5 7.1 2.2 .84 .33
SD 2.0 2.6 1.9 1.6
Anxious
M 8.1 5.0 7.9 3.7 .81
SD 0.8 2.1 1.8 2.2
Afraid of a possible loss
M 7.6 4.5 7.8 3.0 .81
SD 2.0 2.5 1.5 2.1
Vengeful
M 7.0 4.6 6.5 2.4 .78
SD 1.9 2.5 2.3 1.6
Angry at a particular person
M 8.2 5.8 7.8 2.4 .74 .52
SD 1.2 2.5 2.1 1.6
Hurt
M 8.5 6.5 8.0 2.6 .73 .58 —
SD 0.7 2.3 1.7 1.6
Threatened
M 7.4 5.1 6.9 3.8 .70 — .44
SD 1.9 2.4 1.9 2.8
Justified in having these feelings
M 7.7 5.5 7.5 3.2 .68 — —
SD 1.4 2.0 2.2 2.5
Hostility toward (rival)
M 8.6 6.8 7.5 3.1 .64 .61 —
SD 0.7 2.3 1.9 2.4
Rejected
M 8.4 6.2 6.5 2.7 .62 .56 —
SD 1.2 2.7 2.8 1.9
Uncertain
M 6.7 5.4 7.2 4.4 .61 — .58
SD 1.9 2.3 2.0 2.7
Depressed
M 7.3 7.5 6.1 2.9 .36 .81 —
SD 1.5 1.5 2.3 1.9
Longing for what another has
M 6.5 7.7 5.1 3.7 — .76 .32
SD 2.4 1.3 2.7 2.4
Feeling inferior
M 7.5 7.1 6.0 3.0 .34 .72 —
SD 1.0 1.7 2.5 2.0
Feeling like a failure
M 6.5 6.0 4.5 2.5 — .70 —
SD 1.9 2.3 2.7 1.6
Feeling mediocre
M 5.9 6.7 5.2 3.6 — .67 .38
SD 2.2 1.4 2.6 2.2
 ENVY AND JEALOUSY 915

Table 2 (continued)
Cells in the
Envy X Jealousy
interaction Loadings on the
first 3 principal
High envy- High envy- Low envy- Low envy- components
high low high low
Item jealousy" jealousy11 jealousy1" jealousy0 "Jealousy" "Envy" "Disapproval"

Angry about the unfairness of life

M 7.3 6.6 6.1 2.7 .53 .66 —
SD 1.5 2.1 2.6 1.8
Resentment
M 7.6 7.3 7.1 3.2 .52 .66 —
SD 1.6 1.9 2.0 2.0
Feeling unfairly treated by life
M 7.5 6.8 5.8 2.5 .50 .65 —
SD 1.7 2.3 2.6 2.0
Self-doubt
M 6.8 7.0 6.5 3.8 .33 .57 .37
SD 2.2 1.8 2.4 2.3
Insecure
M 6.6 6.3 7.4 3.3 .49 .52 —
SD 1.8 1.9 2.2 2.2
Feeling wishful
M 3.6 7.0 2.9 3.7 -.40 .43 —
SD 2.6 2.5 2.2 2.5
Guilt over some feelings of ill will
M 3.9 5.1 3.9 4.9 — — .74
SD 2.2 2.3 2.5 2.6
Other people would disapprove
of some of his or her feelings
M 4.0 5.8 4.4 4.0 — — .74
SD 2.4 1.9 2.6 2.5
Dissatisfied with himself or herself
M 5.7 6.8 4.8 4.2 — .57 .61
SD 2.4 2.2 2.7 3.0
Some of his or her feelings would be
embarrassing to admit to
M 6.4 6.8 6.0 5.3 — .45 .60
SD 2.2 2.2 2.2 2.8
Feeling sinful
M 2.6 3.4 2.4 3.0 — — .55
SD 2.0 2.0 2.0 2.4
Motivated to improve
himself or herself
M 4.4 5.0 4.6 3.8 .37
SD 2.2 2.2 2.5 2.6

Note. Items were rated on a 9-point scale. Loadings indicating less than 10% shared variance are indicated by a dash.
"n=22. "n = 24. c «=23.

were successful in creating envy that was as intense as jealousy 0.81) than following the low-jealousy vignettes (M = —0.81),
and that statistical adjustments for intensity were unnecessary. F(\, 11) = 197.65, p < .0001, i)ai, = -72. Neither the main effect
Each of the three orthogonal principal-components scores of envy nor the interaction reached significance.
were then entered as the dependent variable in an ANOVA The jealousy component was also significantly affected by
having four independent variables (envy, jealousy, story, and story context. The tennis story elicited more powerful jealousy
subject's sex). Bonferroni adjustment was used to keep overall {M = 0.18) than did the medical school story (M = -0.18), F(\,
Type I error rate at .05, setting alpha at .016 for each ANOVA. 11) = 9.80, p < .01, ijait = • 11 • This difference was surely due to
The analysis of the jealousy component showed that the the fact that the tennis story involved kissing, whereas the medi-
measure of jealous affect was powerfully affected by the iden- cal school story involved only plans for dinner.
tity of the rival's romantic partner, yet was unaffected by the The analysis of the envy component, in contrast, showed that
manipulation of social comparison with the rival. There was a the measure of envious affect was powerfully affected by the
large main effect for the manipulation of jealousy, the mean manipulation of social comparison with the rival. The mean
being much greater following the high-jealousy vignettes (M = envy component score was higher in the high-envy condition
 916 W GERROD PARROTT AND RICHARD H. SMITH

(M = 0.65) than in the low-envy condition (M= -0.73), F(l, 77) greater in the high-envy condition than in the low-envy condi-
= 98.23, p < .0001, *4 = .56. There was no main effect of tion—even in the high-jealousy condition the mean of the envy
jealousy, but there was significant interaction between the envy component was significantly higher in the high-envy condition
and jealousy manipulations, F(l, 77) = 21.68, p < .0001, ^ It = than in the low-envy condition, F(l, 77) = 13.70, p < .001.
.22. The means are graphed in Figure 2. There were no significant main effects or interactions for the
The interaction of the envy component was examined using third principal component, disapproval.
an analysis of simple effects. In the low-envy conditions, the The items "envious" and "jealous" The analyses described
mean of the envy component was appreciably higher with high thus far have demonstrated the distinctions between the emo-
jealousy than with low jealousy, F(l, 77) = 12.99, p < .001. tions of envy and jealousy. We can now examine subjects' rat-
Thus, in the low-envy conditions the jealousy manipulation ings of the items "envious" and "jealous" themselves to see how
increased rated envy in addition to rated jealousy. This finding subjects' use of these terms compared with the emotions experi-
suggests that receiving the romantic interest of one's own enced. Table 3 reports the correlations between these items and
partner is itself enviable. the three principal-components scores. Only the item "jealous"
In the high-envy conditions, however, the addition of jeal- correlated with the jealousy component. Both of the items
ousy had the opposite effect on the envy component, although correlated with the envy component. Only the item "envious"
to a lesser extent, actually lowering the mean of the envy compo- correlated with the disapproval component. This pattern
nent, F(l, 77) = 8.90, p < .01. Inspection of the 18 items loading agrees with previous findings that the term jealousy can refer
on the envy component revealed that only 4 actually exhibited to either jealousy or envy, whereas the term envy can only refer
this pattern to a meaningful extent ("feeling wishful," "longing to envy and that envy more than jealousy is considered to be
for what another has," "dissatisfied with self," and "feeling me- socially disapproved (R. H. Smith et al., 1988).
diocre"). Possibly the presence of anger from high jealousy Subjects' ratings of these two items were analyzed by a MAN-
caused these aspects of envy to seem less salient. It should be OVA. Multivariate significance was achieved for three effects:
noted that the envy component never failed to be significantly envy, F(2, 76) = 18.23, p < .0001, r,2 = .32; jealousy, F(2, 76) =
31.52, p < .0001, rj1 = .45; and the interaction between envy and
jealousy, F(2, 76) = 8.09, p < .001, n2 = • 18. The interpretation
First Component: "Jealousy" of the two main effects parallels the results of the correlational
1.5 analysis, as shown by the means and discriminant function
loadings in Table 4. The high-envy condition elicited higher
o 1.0 ratings than the low-envy condition both for "envious" and for
u
(A "jealous," Fs(l, 77) > 21.12, p < .0001. A similar difference
0.5 occurred between the low- and high-jealousy conditions for
o ratings of "jealous," F(l, 77) = 41.95, p < .0001, but not for
0.0 ratings of "envious," F(\, 77) = 0.03, ns. The interaction be-
a.
E tween envy and jealousy was significant for both items, how-
o
o -0.5 ever, Fs(l, 77) > 12.12, ps < .001. The pattern of means suggests
c ^ High Envy that the term jealous can refer either to jealousy or to envy,
<0
0)
| [Low Envy whereas the term envious can refer only to envy.
-1.5
High Low
Jealousy Discussion
Overall, the results supported the distinction between envy
Second Component: "Envy"
and jealousy. If these two emotions differed only with respect to
1.5 the situations eliciting them and not with respect to the quality
1.0 of their emotional content, separate principal components
o
u
V)
•^
c
0.5
0) Table 3
o 0.0 Correlations of the Items "Envious" and "Jealous" With the
a.
E -0.5 Principal-Components Scores
o
o
c
ffl High Jealousy
Item
<o -1.0
a
• Low Jealousy Principal component Envious Jealous
-1.5
High Low Jealousy 0.10 0.54*
Envy Envy 0.61* 0.67*
Disapproval 0.40* 0.16
Figure 2. Mean scores of thefirst(jealousy) and second (envy) princi-
pal components for the cells of the Envy X Jealousy interaction in Note. N=93.
Experiment 2. (Error bars indicate standard errors of the mean.) * p<.000\.
 ENVY AND JEALOUSY 917

Table 4 the independent manipulation of romantic competition and

Mean Scores and Discriminant Function Loadings social comparison: It implies that it is unusual to have jealousy
without some envy as well.
for Each Significant Multivariate Effect of the Items
"Envious" and "Jealous" The second aspect of this interaction was that the presence of
romantic threat from a superior rival elicited somewhat lower
Item envy ratings than this rival elicited without romantic threat. An
attributional or interpretative process could plausibly cause this
Effect n Envious Jealous
effect: When circumstances of jealousy are present, the feel-
Envy ings of dissatisfaction caused by the envy manipulation may be
High 46 6.9, 7.8, interpreted as symptoms of unfairness rather than as symptoms
Low 47 4.6, 5.4, of inferiority (Clore & Parrott, 1991). Such a phenomenon may
Discriminant function result from a motivation to fault others rather than oneself
loading 0.76 0.98
(Parrott, 1993). Alternatively, this decrease in envy may be ex-
Jealousy
High 46 5.8, 7.9, plained as the result of inattention when subjects are distracted
Low 47 5.7, 5.3, by jealousy.
Discriminant function It is important to note that the pattern of effects exhibited by
loading 0.18 0.87
the words envious and jealous was not the same as that exhib-
Envy X Jealousy interaction
High envy ited by the component scores measuring the emotions envy and
High jealousy 22 6.0b 8.3, jealousy. The word jealousy could refer either to jealousy or to
Low jealousy 24 7.7, 7.2, envy and was not associated with disapproval. The word envy
Low envy could refer only to the emotion envy and was associated with
High jealousy 24 5.5b 7.4,
Low jealousy 23 3.6, 3.3, disapproval. This finding underscores the importance of distin-
Discriminant function loading 0.85 0.95 guishing emotions from emotion words, a point emphasized by
other authors as well (e.g., Ortony et al, 1988).
Note. Main effect means not sharing subscripts indicate a significant
analysis of variance effect. Interaction means in the same column not
sharing subscripts indicate a significant Tukey pairwise comparison.
Subscripts that are consecutive indicate p < .05, whereas those that are General Discussion
inconsecutive indicate p < .0001. The discriminant function loading is
the correlation between an original item and the multivariate discrimi-
nant function for that effect. Summary and Interpretation of Findings
These two experiments found qualitative differences be-
tween the experiences of envy and jealousy that clearly justify
should not have emerged as they did. Instead, the independent the distinction between these emotions. In both experiments,
manipulation of envy and jealousy resulted in three distinguish- envy was characterized by feelings of inferiority and longing;
able dimensions of covariation. The first two principal compo- envy was also associated with disapproval. Jealousy, in both
nents represented composite measures of jealousy and envy, experiments, was characterized by anxiety, distrust, and fear of
respectively. The emergence of a third component suggested rejection or loss. In the second experiment, improved measures
that disapproval of these feelings can be distinguished from of hostility suggested that jealousy tends to produce justified
envy and jealousy per se. The pattern of loadings generally anger, whereas envy tends to produce unsanctioned ill will.
replicated the results of Experiment 1. In addition, the present These findings are remarkably consonant with the tradi-
experiment produced some support for the predicted distinc- tional distinction between envy and jealousy. They also are
tion between envious and jealous hostility. consonant with the idea, prevalent in cognitive theories of emo-
These principal-components scores, when entered into AN- tion, that the appraisal that another person possesses some-
OVAs, demonstrated the extent to which the independent ma- thing that one desires but lacks ought to lead to a different set of
nipulation of envy and jealousy could result in independent emotions than should the appraisal that one may lose an impor-
effects on measures of these emotions. The jealousy measure tant relationship with a person who now prefers a different
was powerfully affected by the presence or absence of threat to partner.
a romantic relationship yet was not at all affected by the manip- One conclusion suggested by these experiments is that situa-
ulation of envy. The envy measure, on the other hand, was tions creating jealousy inherently also create envy to some de-
strongly affected by the superiority or inferiority of social com- gree. This proposition is widely acknowledged in the literature
parison in personally important domains. (Farber, 1966; Salovey & Rodin, 1986) and was shown to be true
The effect of the envy manipulation on measured envy inter- of the naturalistic sample of Experiment 1. The results of the
acted with the jealousy manipulation, however. The presence factorial design of Experiment 2 provided the first experimen-
of romantic threat from an inferior rival elicited higher ratings tal evidence that this tendency is virtually unavoidable. Even
on the envy measure than occurred without romantic threat, when the rival was described as unattractive, incompetent, and
suggesting that the romantic attention of a desirable person is unsuccessful, the attention the rival received from the protago-
itself an enviable quality, just as is prestige, good looks, or nist's romantic partner was sufficient to elicit a moderate
wealth. It is noteworthy that this interaction occurred despite amount of envy. It may be that it is nearly impossible to describe
 918 W GERROD PARROTT AND RICHARD H. SMITH

a case of jealousy in which at least the possibility of envy is not tionnaire items, the statistical techniques used, and the use of a
present.6 factorial design in Experiment 2.
Why might this be? It could be that "romantic attention from The issue of emotional intensity proved to be important in
a person whose attention is valued" is merely one type of envi- distinguishing these emotions. In Experiment 1, statistical ad-
able quality, on par with success, talent, beauty, and the others. justment for individual differences in overall emotional inten-
Or it could instead be that this enviable quality is different in an sity was needed to discern the differences predicted by theory.
important way, namely, that the judgment of superiority is In Experiment 2, selecting stimuli that produced equal intensi-
made not by the envious person but by another person, one ties of envy and jealousy revealed similar differences without
whose opinion matters a great deal. This distinction is reminis- such adjustments. Thesefindingssupport the general proposi-
cent of the classic conceptual distinction between comparative tion that the similarity or dissimilarity between two emotions
appraisal (Festinger, 1954; Kelley, 1952) and reflected appraisal has to do with the relative salience of the components that char-
(Cooley, 1902; Felson, 1985; Mead, 1934). Envy typically results acterize them. It is the shape of the profile of the components,
from a social comparison that is used for comparative ap- not their absolute value, that identifies an emotion.
praisal, the outcome of this appraisal indicating inferiority. In The methodology devised for Experiment 2 may prove useful
contrast, the envy caused by jealousy results from inferring in studying other emotions. Factorial manipulation of emo-
another person's negative reflected appraisal. A jealous man tions and use of vignettes were combined with principal-com-
may envy his rival despite being unable to answer the question ponents analysis in a way that allowed the affective states of
"What does she see in him?" envy and jealousy to be distinguished despite their co-occur-
Whatever the cause of jealous envy, its frequency suggests a rence and without relying on their verbal labels. Factorial ma-
likely source of the semantic ambiguity of the term jealousy. As nipulation created distinct patterns of item covariation for envy
shown in Experiment 1, the conditional probability of envy and jealousy that principal-components analysis identified.
given jealousy is quite high, whereas the conditional probabil- Modifications of the vignettes manipulated envy and jealousy
ity of jealousy given envy is rather low. Perhaps because of these without naming them. Envy and jealousy are not the only emo-
contingencies, the term jealousy came to refer either to jeal- tions that frequently co-occur, nor are they unique in having
ousy or to envy, whereas the term envy remained unambiguous ambiguous or imprecise verbal labels. The techniques that cir-
(as shown in Experiment 2). cumvented these problems in Experiment 2 may therefore be
One apparent inconsistency between Experiments 1 and 2 helpful in future research on other emotions.
needs to be addressed. In both experiments envy was associated One characteristic of all research comparing envy and jeal-
with disapproval; in Experiment 1 the clusters "disapproval of ousy to date is that it treats these emotions as if each occurred in
feelings" and "motivation to improve" distinguished envy from only one typical form. Such was the nature of Salovey and Ro-
jealousy, and in Experiment 2 the item "envious" correlated din's (1986) experiments and such was the nature of the present
positively with the third principal component ("disapproval"). experiments. Yet it is surely the case that there are subvarieties
However, this third component was not significantly affected of envy and jealousy (Parrott, 1991). It remains for future re-
by either the envy or jealousy manipulation. This apparent in- search to conduct afinergrained analysis of the different types
consistency may be resolved by distinguishing between public of envy- and jealousy-producing situations.
and private disapproval. Examination of individual items sug-
gests that awareness of public unacceptability is indeed charac- The Importance of Distinguishing Envy and Jealousy
teristic of envy but that private unacceptability is not. In Exper-
iment 2 the item "embarrassing to admit to" showed a main The importance of recognizing the distinction between envy
effect for envy and the item "others would disapprove" exhib- and jealousy may be appreciated by considering some recent
ited a similar pattern. In contrast, "guilt over feeling ill will" research in which it was not made. An experiment by Bush,
showed a significant main effect for jealousy, being greatest in Bush, and Jennings (1988) investigated the effects of a romantic
the low-jealousy condition. It therefore seems likely that in rival on emotions and on perceptions of relationships. These
both experiments guilt discriminated envy from jealousy more researchers chose not to include a measure of envy in a battery
by its absence in jealousy than by any association with envy. of measures that included jealousy and other emotions. This
Overall, one might say that envy tended to elicit concern about omission appears to have caused them to miss an interesting
public disapproval, whereas jealousy tended to elicit self- finding, because, contrary to their prediction, subjects rated
righteousness. themselves as having more jealousy when exposed to a low level
of threat to the relationship than when exposed to a high level of
threat. The cause of the unexpected result was that the low
threat (as intended) created only a modest degree of jealousy
Methodological Issues
The two experiments used quite different methods. That the 6
One counterexample, noted by Ulanov and Ulanov (1983), is that
results were fairly consistent in these two experiments may help of the Christian notion of God. God, being all-powerful, is not consid-
mitigate any reservations that derive from the methodologies of ered to be susceptible to envy. Being wishful that people love Him,
either. The control of emotional intensity probably contributed however, God can be considered to be susceptible to jealousy. Among
to the present experiments' success in finding differences that mortals, however, envy always seems possible in situations that cause
had eluded previous investigations, as did the choice of ques- jealousy.
 ENVY AND JEALOUSY 919

but (not as intended) created a greater degree of envy, and be- Festinger, L. (1954). A theory of social comparison processes. Human
cause the word jealousy can mean either jealousy or envy, rat- Relations, 7, 117-140.
ings for this item were significantly greater in the low-threat Foster, G. M. (1972). The anatomy of envy: A study in symbolic behav-
(but high social comparison) condition than in the high-threat ior. Current Anthropology, 13, 165-202.
(but low social comparison) condition. Had the investigators Frijda, N. H. (1986). The emotions. Cambridge, England: Cambridge
University Press.
measured envy and jealousy separately, the confound would
Hartigan, J. (1990). Cluster analysis of variables. In W J. Dixon (Ed.),
have been avoided, and one of the major predictions of this BMDP statistical software manual: To accompany the 1990 software
study might have been borne out. release (Vol. 2, pp. 807-816). Berkeley, CA: University of California
In scientific research, abandoning an important distinction Press.
is generally more costly than is maintaining an unimportant Hupka, R. B. (1984). Jealousy: Compound emotion or label for a partic-
one, and this principle seems to hold in the case of envy and ular situation? Motivation and Emotion, 8,141-155.
jealousy (White & Mullen, 1989). Given the results reported in Hupka, R. B. (1991). The motive for the arousal of romantic jealousy:
this article, there seems ample reason to maintain a distinction Its cultural origin. In P. Salovey (Ed.), The psychology of jealousy and
between the emotional experiences of envy and jealousy. The envy (pp. 252-270). New York: Guilford Press.
study of Bush et al. (1988) is only one example of recent studies Hupka, R. B., Buunk, B., Falus, G., Fulgosi, A., Ortega, E., Swain, R., &
that might have been enhanced by distinguishing envy from Tarabrina, N. V (1985). Romantic jealousy and romantic envy: A
seven-nation study. Journal of Cross-Cultural Psychology, 16, 423-
jealousy, either by manipulating or by measuring envy and jeal-
446.
ousy separately (cf. studies by Schmitt, 1988, on social compari-
Kelley, H. H. (1952). Two functions of reference groups. In G. Swanson,
son in romantic jealousy; by Bringle, Renner, Terry, & Davis, T. Newcomb, & E. Hartley (Eds.), Readings in social psychology (2nd
1983, on situational and dispositional aspects of jealousy; and ed., pp. 410-414). New York: Holt, Rinehart and Winston.
by Salovey & Rodin, 1988, on coping with envy and jealousy). It Mathes, E. W (1986). Jealousy and romantic love: A longitudinal study.
is our view that future research should take care to treat envy Psychological Reports, 58, 885-886.
and jealousy as distinct emotions, not only in terms of the Mathes, E. W, Adams, H. E., & Davies, R. M. (1985). Jealousy: Loss of
situations giving rise to them but also in terms of the quality of relationship rewards, loss of self-esteem, depression, anxiety, and
the emotional experiences themselves. anger. Journal of Personality and Social Psychology, 48, 1552-1561.
Mead, G. H. (1934). Mind, self, and society. Chicago: University of
Chicago Press.
Neu, J. (1980). Jealous thoughts. In A. O. Rorty (Ed.), Explaining emo-
References tions (pp. 425-463). Berkeley, CA: University of California Press.
Nozick, R. (1974). Anarchy, state and Utopia. New York: Basic Books.
Aldenderfer, M. S., & Blashfield, R. K. (1984). Cluster analysis. Beverly Ortony, A., Clore, G. L., & Collins, A. (1988). Cognitive structure of
Hills, CA: Sage. emotions. Cambridge, England: Cambridge University Press.
Barker, R. L. (1987). The green-eyed marriage: Surviving jealous rela-
Parrott, W G. (1988). The role of cognition in emotional experience. In
tionships. New York: Free Press.
W J. Baker, L. P. Mos, H. V Rappard, & H. J. Stam (Eds.), Recent
Baumgart, H. (1990). Jealousy: Experiences and solutions (M. Jacob- trends in theoretical psychology (pp. 327-337). New York: Springer-
son & E. Jacobson, Trans.). Chicago: University of Chicago Press. Verlag.
Bringle, R. G., & Buunk, B. (1985). Jealousy and social behavior: A
Parrott, W G. (1991). The emotional experiences of envy and jealousy.
review of person, relationship and situational determinants. In P.
In P. Salovey (Ed.), The psychology of jealousy and envy (pp. 3-30).
Shaver (Ed.), Review ofpersonality and social psychology: Vol. 6. Self,
New York: Guilford Press.
situations, and social behavior (pp. 241-264). Beverly Hills, CA:
Parrott, W G. (1993). On the scientific study of angry organisms. In
Sage.
R. S. Wyer & T. Srull (Eds.), Perspectives on anger and emotion: Ad-
Bringle, R. G., Renner, P., Terry, R. L., & Davis, S. (1983). An analysisof
vances in social cognition (Vol. 6, pp. 167-177). Hillsdale, NJ: Erl-
situation and person components of jealousy. Journal ofResearch in
baum.
Personality, 17, 354-368.
Rawls, J. (1971). A theory of justice. Cambridge, MA: Belknap Press.
Bush, C. R., Bush, J. P., & Jennings, J. (1988). Effects of jealousy threats Rorty, A. O. (1988). Mind in action: Essays in the philosophy of mind.
on relationship perceptions and emotions. Journal ofSocial and Per- Boston: Beacon Press.
sonal Relationships, 5, 285-303. Roseman, I. J. (1984). Cognitive determinants of emotions: A struc-
Buunk, B. P. (1991). Jealousy in close relationships: An exchange- tural theory. In P. Shaver (Ed.), Review ofpersonality and social psy-
theoretical perspective. In P. Salovey (Ed.), The psychology of jeal- chology, Vol. 5. Emotions, relationships, and health (pp. 11-36). Bev-
ousy and envy (pp. 148-177). New 'York: Guilford Press. erly Hills, CA: Sage.
Cicero. (1927). Tusculan disputations (J. E. King, Trans.). London: Wil- Russell, B. (1930). The conquest of Happiness. New York: Liverlight.
liam Heinemann. (Original work written 45 B.C.) Salovey, P., & Rodin, J. (1984). Some antecedents and consequences of
Clore, G. L., & Parrott, W. G. (1991). Moods and their vicissitudes: social-comparison jealousy. Journal of Personality and Social Psy-
Thoughts and feelings as information (pp. 107-123). In J. Forgas chology, 47, 780-792.
(Ed.), Emotion and social judgment. Elmsford, NY: Pergamon Press. Salovey, P., & Rodin, J. (1986). The differentiation of social-compari-
Cooley, C. H. (1902). Human nature and the social order. New York: son jealousy and romantic jealousy. Journal ofPersonality and Social
Scribner's. Psychology, 50,1100-1112.
Cronbach, L. J., & Gleser, G. (1953). Assessing similarity between pro- Salovey, P., & Rodin, J. (1988). Coping with envy and jealousy. Journal
files. Psychological Bulletin, 50, 456-473. of Social and Clinical Psychology, 7, 15-33.
Farber, L. (1966). Ways ofthe will. New York: Basic Books. Salovey, P., & Rothman, A. J. (1991). Envy and jealousy: Self and soci-
Felson, R. B. (1985). Reflected appraisal and the development of self. ety. In P. Salovey (Ed.), The psychology of jealousy and envy (pp.
Social Psychology Quarterly, 48, 71-78. 271-286). New York: Guilford Press.
 920 W GERROD PARROTT AND RICHARD H. SMITH

Schmitt, B. H. (1988). Social comparison in romantic jealousy. Person- works of Spinoza (Vol. 1, pp. 408-617). Princeton, NJ: Princeton
ality and Social Psychology Bulletin, 14, 374-387. University Press. (Original work published 1678)
Schoeck, H. (1969). Envy: A theory of social behavior. New York: Har- Stearns, P. N. (1989). Jealousy: The evolution of an emotion in American
court, Brace & World. history. New York: New York University Press.
Silver, M., & Sabini, J. (1978a). The perception of envy. Social Psychol- Sullivan, H. S. (1953). The interpersonal theory of psychiatry. New York:
ogy Quarterly, 41, 105-117. Norton.
Silver, M., & Sabini, J. (1978b). The social construction of envy. Journal Taylor, G. (1988). Envy and jealousy: Emotions and vices. Midwest
for the Theory of Social Behaviour, 8, 313-332. Studies in Philosophy, 13, 233-249.
Smith, C. A., & Ellsworth, P. C. (1985). Patterns of cognitive appraisal Tesser, A. (1988). Toward a self-evaluation maintenance model of so-
in emotion. Journal of Personality and Social Psychology, 48, 813- cial behavior. In L. Berkowitz (Ed.), Advances in experimental social
838. psychology(Vol. 21, pp. 181-227). San Diego, CA: Academic Press.
Smith, R. H. (1991). Envy and the sense of injustice. In P. Salovey (Ed.), Tesser, A., Millar, M., & Moore, J. (1988). Some affective consequences
The psychology of jealousy and envy (pp. 79-99). New York: Guilford of social comparison and reflection processes: The pain and plea-
Press. sure of being close. Journal of Personality and Social Psychology, 54
Smith, R. H., Diener, E., & Garonzik, R. (1990). The roles of outcome 49-61.
satisfaction and comparison alternatives in envy. British Journal of Ulanov, A., & Ulanov, B. (1983). Cinderella and her sisters: The envied
Social Psychology, 29, 247-255. and the envying. Philadelphia: Westminster Press.
Smith, R. H., Kim, S. H., & Parrott, W G. (1988). Envy and jealousy: Wegner, D. M., & Vallacher, R. R. (1977). Implicit psychology: An intro-
Semantic problems and experiential distinctions. Personality and duction to social cognition. New York: Oxford University Press.
Social Psychology Bulletin, 14, 401-409. White, G. L. (1981a). A model of romantic jealousy. Motivation and
Smith, R. H., Parrott, W. G., & Diener, E. (1991). Development and Emotion, 5, 295-310.
validation of a scale for measuring enviousness. Manuscript submit- White, G. L. (1981b). Some correlates of romantic jealousy. Journal of
ted for publication. Personality, 49, 129-147.
Smith, R. H., Parrott, W. G., Ozer, D, & Moniz, A. (1992). Subjective White, G. L., & Mullen, P. E. (1989). Jealousy: Theory, research, and
unfairness and inferiority as predictors of hostile and depressive feel- clinical strategies. New York: Guilford Press.
ings in envy. Manuscript submitted for publication.
Solomon, R. C. (1976). The passions: The myth and nature of human Received June 11,1991
emotion. Notre Dame, IN: University of Notre Dame Press. Revision received November 18,1992
Spinoza, B. (1985). Ethics. In E. Curley (Ed. and Trans.), The collected Accepted December 1,1992 •