Australian Journal of Psychology 2011; 63: 243251

doi:10.1111/j.1742-9536.2011.00026.x

Measuring romantic jealousy: Validation of the multidimensional

jealousy scale in Australian samples ajpy_26 243..251

Rachel A. Elphinston, Judith A. Feeney, and Patricia Noller

School of Psychology, The University of Queensland, Brisbane, Queensland, Australia

Abstract

Although a number of questionnaires have been developed to measure romantic jealousy, there is a lack of independent research that
has substantiated their psychometric properties and minimal research that has tested and applied these measures in a consistent and
reliable way. We aimed to address these gaps by providing the first validation study of an existing measure of romantic jealousythe
Multidimensional Jealousy Scale (MJS). The MJS was subjected to exploratory and confirmatory factor analyses in two separate
samples of individuals in romantic relationships of at least 2 months duration. A 17-item Short-Form MJS (SF-MJS) was supported,
with three factors representing cognitive, emotional, and behavioural jealousy. Internal reliabilities were high. Consistent evidence
supported the discriminant and concurrent validity of the SF-MJS. Overall, the current research provides a brief, validated measure
of romantic jealousy for use in the general population. Potential clinical uses are discussed.

Key words: attachment, jealousy, rumination

Over almost three decades, a considerable body of research involving cognitive, emotional, and behavioural compo-
has accumulated investigating jealousy in the context of nents (Guerrero & Andersen, 1998).
romantic relationships. The majority of this research has Despite agreement about the multidimensional nature of
focused on predictors and consequences of romantic jeal- romantic jealousy, there is a paucity of brief, validated, and
ousy (e.g., White & Mullen, 1989). For example, jealousy widely used measures of romantic jealousy. In the current
has been associated with attachment anxiety (e.g., Guerrero, study, we provide the first psychometric evaluation of the
1998), rumination (e.g., Carson & Cupach, 2000), relational Multidimensional Jealousy Scale (MJS; Pfeiffer & Wong,
dissatisfaction (e.g., Guerrero & Eloy, 1992), and verbal and 1989) since its development, in Australian samples of indi-
physical abuse (e.g., Barnett, Martinez, & Bluestein, 1995). viduals currently in romantic relationships of 2 months or
With increasing evidence that romantic jealousy plays a key more.
role in relationship experience and relationship problems,
it becomes important to ensure that assessment instruments
are validated and mapped onto acceptable conceptualisa- SELF-REPORT MEASUREMENT OF ROMANTIC JEALOUSY
tions of romantic jealousy.
As such, the way in which romantic jealousy is defined In a series of studies, Pfeiffer and Wong (1989) developed
has crucial implications for how this construct is measured. and tested the MJS as a measure of romantic jealousy in a
Among jealousy researchers, there is a general consensus Canadian sample. The MJS consisted of 24 items, each with
that romantic jealousy involves a complex mix of emotions, a 7-point Likert-scale response format. Three factors com-
thoughts, and behaviours that are experienced and prised the MJS, labelled cognitive, emotional, and behav-
expressed by those who perceive their relationship as threat- ioural jealousy. Cognitive jealousy referred to the frequency
ened by a potential third party (e.g., Guerrero, Spitzberg, & of a persons suspicions and worries regarding partner inter-
Yoshimura, 2004; Pfeiffer & Wong, 1989). A broad theoreti- est in a rival, and interest received from a rival. Emotional
cal model of jealousy experience and expression (the com- jealousy was defined as the degree of upset a person expe-
ponential model) also supports the view of jealousy as riences when exposed to jealousy-evoking situations, while
behavioural jealousy referred to the frequency with which a
Correspondence: Rachel Elphinston, Ph.D. (Clin Psych) School of person engages in detective and protective behaviours such
Psychology, The University of Queensland, Brisbane, QLD 4072,
as questioning and surveillance of their partner.
Australia. Email: [email protected]
Received 21 November 2010. Accepted for publication 16 The researchers theorised that these three dimensions
January 2011. of romantic jealousy interact together and occur simulta-
2011 The Australian Psychological Society neously, rather than developing from a sequential process of
244 R.A. Elphinston et al.

cognitions to emotions to behaviours, as previously sug- A number of other questionnaires have been developed to
gested by White (1981). Further, they reported that each measure romantic jealousy, based on different theoretical
of these dimensions can be described on a continuum from conceptualisations. To date, however, minimal research has
normal to pathological. Specifically, normal jealousy may tested and applied these measures in any consistent and
encompass some degree of jealousy-related cognitions, emo- reliable way. Early research examined unidimensional scales
tions, and behaviours, while pathological jealousy would measuring romantic jealousy, such as the Relationship Jeal-
involve a high degree of suspicion and emotional upset and ousy Scale (White, 1981). Despite the parsimonious concep-
extensive detective behaviours. In their original study, low tualisation of jealousy as a single dimension, these measures
scores indicated normal levels of jealousy and higher scores did not assist in clarifying the associations between jealousy
indicated more pathological levels of jealousy (Pfeiffer & and individual-difference variables (Pfeiffer & Wong, 1989).
Wong, 1989). Other jealousy measures incorporate some aspects of roman-
Based on discussions among six psychology students and tic jealousy; however, these questionnaires were generally
the two authors (Pfeiffer & Wong, 1989), eight items per designed to measure more dispositional characteristics
subscale were generated. Scale validation involved subject- across all past relationships, such as the Chronic Jealousy
ing the MJS scores of 178 individuals (with a mean age of Scale (CJS; White, 1981), the SRJS, and the Revised SRJS
approximately 45 years) recruited from an educational (Bringle, 1982; Bringle et al., 1979). These measures are
setting to a series of exploratory factor analyses (EFAs) using limited in their applicability for two reasons: First, they focus
the Principal Components Analysis (with Varimax rotation). on jealousy that occurs across many different relationships
A stable three-factor solution was indicated in all studies, (romantic and non-romantic), and second, they do not fit
with cognitive, emotional, and behavioural factors specified. the theoretical conceptualisation of romantic jealousy as
The scales were shown to have adequate internal consis- multidimensional.
tency (alphas all above 0.82). Further, results provided evi- Forced-choice and continuous items based on hypotheti-
dence of convergent and discriminant validity: Emotional cal scenarios have also been used to measure jealousy,
and behavioural jealousy were linked negatively to happi- particularly in relation to the study of sex differences (e.g.,
ness, cognitive jealousy was negatively related to love, emo- Buss, Larsen, Western, & Semmelroth, 1992; Wiederman &
tional jealousy was positively associated with love, and all Allgeier, 1993). In these studies, hypothetical scenarios
three dimensions were negatively related to liking. All sub- describing partner behaviour highlight either sexual or emo-
scales of the MJS were also related to the romantic subscale tional infidelity. Based on evolutionary theory, males should
of the Self-Report Jealousy Scale (Bringle, Roach, Andler, & experience more distress in relation to a sexual infidelity
Evenbeck, 1979), thus demonstrating concurrent validity and females should experience more upset in relation to
(Pfeiffer & Wong, 1989). an emotional infidelity (e.g., Buunk, Angleitner, Oubaid, &
In addition, Pfeiffer and Wong (1989) suggested that emo- Buss, 1996).
tional jealousy may be a relatively common experience, while Despite the contribution of evolutionary theory and asso-
cognitive and behavioural jealousy may imply more severe ciated studies of responses to hypothetical scenarios involv-
levels of (pathological) jealousy. They concluded that the MJS ing infidelity, multi-item scales such as the MJS offer several
may provide researchers with a starting point for investigat- advantages. First, the use of multi-item scales lends itself to
ing more severe levels of jealousy in the population. multidimensional conceptualisations of romantic jealousy,
One limitation of the studies was the sample of partici- thus providing support for its complex structure. Second,
pants, who were asked to respond in relation to either their there are problems with the reliability of forced-choice
current relationship partner or a past partner (if they were methods to assess sex differences (e.g., Harmon-Jones, Peter-
not currently in a romantic relationship); measures of son, & Harris, 2009). While continuous response-types
romantic jealousy may have limited relevance and validity to have attempted to address these measurement concerns,
those who are not in a relationship, even if they have been inconsistencies in findings regarding sex differences in jeal-
in one in the past. Although the authors tried to address this ousy remain (e.g., DeSteno, Bartlett, Braverman, & Salovey,
limitation by conducting correlational analyses on the two 2002).
groups (current and past relationship), and showing similar Another scale examining specific jealousy-related behav-
patterns across groups, the amount of time these individuals iours has been developed, termed the Communicative
were in a romantic relationship, together with the results of Responses to Jealousy Scale (CRJS; Guerrero & Andersen,
these analyses, were not reported. Further, in Pfeiffer and 1998; Guerrero, Andersen, Jorgensen, Spitzberg, & Eloy,
Wongs (1989) study, only an EFA approach was used. 1995). The CRJS consists of 70 items, with 14 subscales
According to Tabachnick and Fidell (2007), scale validation measuring various responses to jealousy. Although this scale
should involve both EFA and confirmatory factor analysis provides a detailed assessment of ways in which jealous
(CFA) (using structural equation modeling). feelings are expressed, it contains a large number of
2011 The Australian Psychological Society
 Measuring romantic jealousy 245

items, involves a complex factor structure, and is still in the METHOD

process of validation.
Participants

Individuals currently in a romantic relationship of at least

THE PRESENT STUDY
2 months duration completed questionnaires and provided
demographic data. According to social penetration theory
Overall, as compared with other measures of romantic
(e.g., Altman & Taylor, 1973), communication exchanges
jealousy, the MJS offers many advantages; these include its
become broader and deeper over time. Similarly, intimacy
multidimensionality, brevity, possible application to examin-
and commitment generally increase over time, and the
ing more pathological levels of jealousy, and parsimonious
initial stages of relationships are characterised by relatively
factor structure. Since its introduction, however, the MJS
superficial exchanges. Therefore, a minimum 2-month
has not been subjected to a complete scale-validation inves-
inclusion criterion was chosen in an attempt to ensure rela-
tigation. Further, the MJS has not been applied consistently
tionships were sufficiently well-established for issues per-
in studies examining romantic jealousy. Although a substan-
taining to jealousy to become relevant. Two independent
tial number of researchers have utilised this measure, they
samples were obtained. All participants were recruited from
often report modifications to its use.
the community by media advertisement and the first-year
Specifically, researchers have variously used only one or
participant pool at the University of Queensland.
two subscales, removed and modified items, collapsed items
The EFA sample consisted of 199 individuals currently
into a single jealousy dimension, incorporated items from
in a romantic relationship from 2 months to 33.5 years in
other measures, and altered the factor structure (e.g.,
length. The mean length of the current relationship was
Afifi & Reichert, 1996; Guerrero, 1998; Guerrero et al.,
4.11 years (SD = 5.81). Ages ranged from 17.5 to 60 years,
1995; Knobloch, Solomon, & Cruz, 2001; Theiss &
with a mean of 27.6 years (SD = 9.87). Approximately 61%
Solomon, 2006). For instance, Rydell and Bringle (2007)
of the sample were females and 39% were males. Almost
used a two-factor structure: reactive jealousy (similar to
half of the participants had completed some university
emotional jealousy) and suspicious (cognitive and behav-
study (45.2%), and approximately 80% of participants
ioural) jealousy. In another example, Afifi and Reichert
were employed full-time, part-time, or on a casual basis.
(1996) developed a one-factor jealousy measure based on
The majority of participants were born in Australia
statistical and conceptual similarities between cognitive and
(73.4%).
emotional dimensions.
The CFA sample comprised 127 participants. Approxi-
Given the use of different versions of the MJS, insufficient
mately 68.5% were females and 31.5% were males. Ages
previous validation of the MJS (no previous CFA has been
ranged from 17.5 years to 49.83 years, with a mean of
conducted), and lack of research supporting the applicability
20.55 years (SD = 4.30). The mean length of the relationship
of other self-report romantic jealousy measures, there is a
was approximately 2 years (SD = 3.45), and ranged from
need for a valid tool assessing romantic jealousy, for use in
2 months to 30 years. The majority of respondents had
the general population. Using both exploratory and CFA, we
undertaken some university studies (70.1%), and approxi-
aimed to address this research gap. First, as no study has
mately 71% were employed on a part-time, full-time, or
investigated the psychometric properties of the MJS since its
casual basis. Australia was the most common country of
development, we tested the factor structure and internal-
birth (74.8%).
consistency reliability of the MJS in Australian samples.
Second, we aimed to provide evidence of the concurrent and
discriminant validity of the MJS. Consistent findings docu-
ment positive associations between neuroticism and jealousy
Measures
(e.g., Buunk, 1997; Melamed, 1991) and between romantic
and chronic/dispositional jealousy (e.g., Pfeiffer & Wong,
Negative emotionality
1989); hence, these constructs were employed to assess the
validity of the MJS. Given the similarities between attach- Braithwaites (1987) Scale of Emotional Arousability (SEA)
ment anxiety and measures of jealousy (e.g., both may be consists of 15 items designed to measure neuroticism. The
activated in times of threat and involve concerns about six-item general emotionality subscale was incorporated as
the relationship), attachment anxiety was also included this subscale is a brief and reliable measure of neuroticism.
to provide evidence for concurrent validity (e.g., Guerrero, Responses were measured on a 5-point Likert scale, ranging
1998). Finally, we provide descriptive statistics on the vali- from 1 (very unlike me) to 5 (very like me). A sample item
dated measure to inform the use of this questionnaire in the is, I frequently get upset. The scale was scored such that
general population and to guide future research. high scores indicated higher levels of neuroticism.
2011 The Australian Psychological Society
246 R.A. Elphinston et al.

Attachment anxiety ing EFA, the suitability of the data for factor analysis was
assessed. The KaiserMeyerOklin value for the MJS was
The 18-item anxiety subscale from the Experiences in
0.86, exceeding the recommended minimum of 0.6 (Kaiser,
Close Relationships Questionnaire (ECR; Brennan, Clark, &
1970, 1974). Bartletts Test of Sphericity (Bartlett, 1950)
Shaver, 1998) provides a continuous measure of attachment
was statistically significant, supporting the factorability of
anxiety. Participants were asked to specify how much they
the correlation matrix. Missing values analyses indicated
agreed or disagreed with each statement, from 1 (strongly
that no variable in the principal components analysis
disagree) to 7 (strongly agree). A sample item is, I worry
(PCA) or CFA analyses had more than 5% of cases missing.
about being abandoned, with items scored such that higher
Accordingly, the pairwise correlation matrix was analysed
scores indicated higher attachment anxiety.
(Tabachnick & Fidell, 2007). Four univariate outliers were
detected. The following analyses were then run with and
Chronic jealousy
without these cases and there was no substantive change
Whites (1981) CJS provides a global measure of the fre- in results; therefore, the data reported include all cases
quency and intensity of chronic jealousy across past and (N = 199). No multivariate outliers were found.
current relationships, and consists of six items. Responses Consistent with Pfeiffer and Wongs (1989) approach,
were recorded on a 5-point Likert-scale format. The scale an initial exploratory factor analysis via PCA with
anchors varied for each item, ranging from 1 (not at all) or oblique (direct oblimin) rotation was used to investigate
1 (not usually) to 5 (extremely) or 5 (almost always). A the structure of the MJS. Five components with eigenval-
sample item is, How jealous a person are you generally? ues greater than 1 were identified. However, examination
Scores were summed, with higher scores indicating more of the scree plot suggested the presence of four principal
chronic jealousy. components. Therefore, PCAs specifying two, three, and
four components were conducted. Of these PCAs, the
MJS three-component analysis provided the optimal solution
in terms of percentage of variance explained, number
The MJS contains three 8-item subscales, and asks partici-
of items per factor, and absence of cross-loadings.
pants to respond with their current partner in mind. For the
Extracted components were moderately correlated
cognitive subscale, participants indicated how often certain
(r = 0.23, r = 0.31, r = 0.37); therefore, an oblique rotation
thoughts about their partner occurred, with responses
was retained.
ranging from 1 (never) to 7 (all the time). A sample item is,
In total, the three extracted components accounted for
I suspect that X (my partner) may be attracted to someone
53.2% of the variance, with component 1 contributing
else. The emotional subscale asks participants to consider
32.32%, component 2 contributing 11.91%, and compo-
their emotional reactions to various situations, such as X
nent 3 contributing 8.92%. Significant cross-loadings were
hugs and kisses someone of the opposite sex. The response
observed for cognitive items 2 (I am worried that some
format ranged from 1 (very pleased) to 7 (very upset). For
member of the opposite sex may be chasing after X), 5 (I
the behavioural subscale, participants were asked to rate
think that some members of the opposite sex may be
how often they engaged in particular behaviours such as
romantically interested in X), and 6 (I am worried that
I question X about his or her telephone calls. Responses
someone of the opposite sex is trying to seduce X), emo-
ranged from 1 (never) to 7 (all the time). Items for each
tional item 3 (X smiles in a very friendly manner to
subscale were summed.
someone of the opposite sex), and behavioural items 4 (I
say something nasty about someone of the opposite sex if X
Procedure shows an interest in that person) and 8 (I pay X a surprise
visit just to see who is with him or her); these items were
Participants completed questionnaires as individuals. Some excluded.
participants completed the measures alone and others in The PCA was re-run on the remaining 18 items. The
small groups. A reply-paid envelope was provided to those results of this analysis again suggested that a three-factor
who completed the mail-out questionnaires. solution with oblique rotation was optimal. The three com-
ponents accounted for 56.4% of the total variance. Item
loadings are presented in Table 1. Based on the item load-
RESULTS
ings, the first component was labelled emotional jealousy,
the second component was labelled cognitive jealousy, and
EFA of the MJS
the third component was labelled behavioural jealousy. As
All analyses were performed in SPSS (Version 15.0, IBM Table 1 shows, the three components had moderate to high
Australia, St Leonards, NSW, Australia). Prior to perform- internal consistency.
2011 The Australian Psychological Society
 Measuring romantic jealousy 247

Table 1 Item loadings from principal components analysis of the multidimensional jealousy scale
Item Component 1 Component 2 Component 3
Cognitive jealousy
1. I suspect that X is secretly seeing someone of the opposite sex. 0.14 -0.88 0.15
3. I suspect that X may be attracted to someone else. 0.22 -0.68 0.37
4. I suspect that X may be physically intimate with another member of the opposite sex 0.17 -0.87 0.19
behind my back.
7. I think that X is secretly developing an intimate relationship with someone of the 0.16 -0.87 0.20
opposite sex.
8. I suspect that X is crazy about members of the opposite sex. 0.15 -0.71 0.29
Emotional jealousy
1. X comments to you on how great looking a particular member of the opposite sex is. 0.73 -0.29 0.27
2. X shows a great deal of interest or excitement in talking to someone of the opposite 0.78 -0.32 0.37
sex.
4. A member of the opposite sex is trying to get close to X all the time. 0.78 -0.21 0.37
5. X is flirting with someone of the opposite sex. 0.83 -0.08 0.25
6. Someone of the opposite sex is dating X. 0.59 -0.15 0.18
7. X hugs and kisses someone of the opposite sex. 0.73 -0.06 0.22
8. X works very closely with a member of the opposite sex (in school or office).* 0.63 -0.34 0.35
Behavioural jealousy
1. I look through Xs drawers, handbag, or pockets. 0.29 -0.16 0.56
2. I call X unexpectedly, just to see if he or she is there. 0.14 -0.22 0.68
3. I question X about previous or present romantic relationships. 0.32 -0.11 0.61
5. I question X about his or her telephone calls. 0.30 -0.25 0.79
6. I question X about his or her whereabouts. 0.23 -0.18 0.82
7. I join in whenever I see X talking to a member of the opposite sex. 0.18 -0.13 0.54
Cronbachs alpha 0.85 0.84 0.76
% Variance 30.80 15.05 10.53
The 18 items in this table refer to the numbers in the original scale (as outlined by Pfeiffer & Wong, 1989).
*The asterisk refers to the item that was removed following the CFA.

CFA of the MJS the measurement error and resulted in a better fit on
all indices, providing an adequate fit to the data
A CFA was conducted in AMOS (Version 6.0, IBM Australia,
(c2(116, N = 127) = 182.29, p < .001, c2/df = 1.57; CFI =
St Leonards, NSW, Australia) on the hypothesised three-
0.91; RMSEA = 0.06; SRMR = 0.07).
factor model. The hypothesised model included cognitive
jealousy with five indicators, emotional jealousy with seven
indicators, and behavioural jealousy with six indicators (see
Table 1 for item listing). Reliability
Maximum likelihood estimation was used to estimate the Internal consistency reliabilities (Cronbachs alpha) were
parameters of the model. Model fit was evaluated in several calculated for each of the jealousy dimensions. Cronbachs
ways, including the chi-square test, chi-square/degree of alpha coefficients suggested that cognitive (a = 0.77),
freedom ratio, comparative fit index (CFI), root mean square emotional (a = 0.81), and behavioural jealousy (a = 0.70)
error of approximation (RMSEA), and standardised root- dimensions had adequate internal consistency.
mean-square residual (SRMR). The following cut-offs were
used to analyse goodness of model fit: a chi-square/degrees
of freedom ratio 23, a CFI 3 0.90, and values of RMSEA and
Concurrent and discriminant validity
SRMR less than 0.060.08 and less than 0.08, respectively
(Kline, 2005). A non-significant chi-square statistic also indi- The CJS and the Anxiety Subscale from the Experiences
cates a satisfactory fit; however, the chi-square fit measure is in Close Relationships Questionnaire were completed by
highly sensitive to sample size, and therefore, goodness of fit participants to provide evidence of concurrent validity,
indices are typically considered the most robust indicators of whereas a measure of emotionality was employed to dem-
model fit (Ullman, 2001). onstrate discriminant validity. See Table 2 for a summary
Overall, the model did not fit the data adequately, of the results. Both concurrent and discriminant validity
c (132, N = 127) = 224.01, p < .001, c2/df = 1.69; CFI = 0.88;
2
were evident. Specifically, in line with previous research
RMSEA = 0.07; SRMR = 0.07. However, removal of one (e.g., Buunk, 1997), cognitive, emotional, and behavi-
item (emotional jealousy, item 8) with substantial covari- oural dimensions of jealousy were associated positively
ance with four or more other items significantly reduced with chronic jealousy and attachment anxiety; further,
2011 The Australian Psychological Society
248 R.A. Elphinston et al.

Table 2 Correlations between the multidimensional jealousy scale and chronic jealousy, attachment anxiety, and emotionality
Cognitive Emotional Behavioural Chronic Anxiety Emotionality
Cognitive 1
Emotional 0.20* 1
Behavioural 0.26** 0.42*** 1
Chronic 0.37*** 0.50*** 0.45*** 1
Anxiety 0.34*** 0.40*** 0.49*** 0.49*** 1
Emotionality 0.13 0.33*** 0.22* 0.34** 0.39*** 1
Note. *p < .05, **p < .01, ***p < .001.

Table 3 Descriptive statistics for the MJS subscales across the combined sample
Subscale Coefficient alpha Mean (SD) Range Skewness Kurtosis
Cognitive 0.85 8.35 (4.17) 527 14.54 16.19
Males 8.40 (4.50) 527
Females 8.32 (3.98) 526
Emotional 0.83 29.06 (6.37) 842 2.82 -0.69
Males 26.89 (6.38) 839
Females 30.30 (6.03) 1342
Behavioural 0.74 13.84 (5.56) 632 4.83 -0.44
Males 12.52 (4.94) 628
Females 14.59 (5.76) 632

emotionality was associated only with emotional and behav- original 24-item MJS; instead, support was found for a reli-
ioural jealousy, providing support for discriminant validity. able and valid SF-MJS, consisting of 17 items.
The findings confirmed a three-factor structure, consistent
with the structure identified in Pfeiffer and Wongs (1989)
Descriptive statistics on the SF-MJS across the original study, representing cognitive, emotional, and
combined sample behavioural aspects of jealousy. Initially, the three factors
accounted for a total of 53.2% of the variance. However,
Descriptive statistics were generated on the new Short-Form
given the cross-loadings, six items were removed from the
MJS (SF-MJS) on the combined sample of 326 participants.
three subscales, showing an increase in the percentage of
Overall, there were 118 male participants (36.2%) and 208
total variance (56.4%), and improving the interpretability of
female participants (63.8%). See Table 3 for a summary of
the factors. A subsequent CFA conducted on the remaining
the mean scores for dimensions of jealousy and descriptive
18 items using an independent sample provided further
statistics. There were significant gender differences for emo-
support for the three-factor model. Removal of one item due
tional jealousy (F(1, 324) = 23.05, p < .001) and behavioural
to substantial covariance with other items reduced the mea-
jealousy (F(1, 324) = 10.68, p < .01), with female partici-
surement error and enhanced the fit indices, resulting in
pants obtaining higher mean scores than male participants
a 17-item SF-MJS. The three factors were statistically and
on both dimensions; there was no significant gender differ-
conceptually related to each other.
ence for cognitive jealousy (F(1, 324) = 0.03, ns). The cog-
Previous research has found some evidence for one-
nitive subscale had moderate to severe positive skewness
and two-factor models of romantic jealousy (e.g., Afifi
and kurtosis, the behavioural subscale had moderate positive
& Reichert, 1996; Buunk et al., 1996). In these studies,
skewness, and the emotional subscale approximated a
however, these measures were not subjected to thorough
normal distribution.
validation analyses. We also believe that unidimensional
structures do not capture the complexity of the experience
DISCUSSION of jealousy in romantic relationships. Our results support
this assertion, together with differences found between the
The purpose of the present research was to examine the dimensions of romantic jealousy in relation to both gender
psychometric characteristics of an existing measure of and general emotionality (see below for a discussion of
romantic jealousy (the MJS) using sample of individuals findings).
currently in romantic relationships. The results did not Each dimension of the SF-MJS had sufficient internal con-
provide specific validation of Pfeiffer and Wongs (1989) sistency reliability. The results of the present study further
2011 The Australian Psychological Society
 Measuring romantic jealousy 249

verified that the SF-MJS demonstrated adequate concurrent possible explanations for these results and potential media-
and discriminant validity; specifically, all dimensions of tors that may be at play. Overall, our finding that cognitive
romantic jealousy were associated with chronic jealousy and jealousy is relatively less common than emotional jealousy is
with attachment anxiety, while emotionality was associated relevant for clinicians working with couples, as the presence
only with emotional and behavioural jealousy. These results of high levels of jealous thoughts may signal that therapeutic
add to the many studies documenting links between attach- intervention is warranted.
ment anxiety and romantic jealousy (e.g., Buunk, 1997; Future studies that focus on establishing normative data
Guerrero, 1998). As noted earlier, attachment anxiety and from a large, non-clinical sample of young adults may help
jealousy share some common characteristics: Both aim to compare individual scores on dimensions of the SF-MJS to
maintain close relationships and are often activated in normative values. Percentile tables may aid the interpreta-
times of relationship threat (Sharpsteen & Kirkpatrick, tion of scores in clinical work with individuals, and could
1997). Hence, these results provide compelling evidence of also be used to set inclusion and exclusion criteria for
concurrent validity. research purposes. In addition, the means and standard
The results showed significant gender differences in mean deviations could provide comparative standards for the
scores for emotional and behavioural jealousy, with female scores obtained in studies of clinical groups. More research is
participants reporting higher levels of jealousy than male necessary to establish whether the SF-MJS is indeed suitable
participants for both dimensions. These findings are consis- for these purposes.
tent with previous research, which has found that female Additional research could also examine the validity of
participants generally report more jealousy than male par- the SF-MJS in clinical samples. Specifically, studies could
ticipants (e.g., Mathes & Severa, 1981). Studies focusing on assess differences between those diagnosed with delusional
distress related to types of infidelity (sexual or emotional) disorderjealous typeaccording to the Diagnostic and Sta-
have provided alternative explanations for sex differences tistical Manual of Mental Disorders-IV-TR (DSM-IV-TR;
based on evolutionary theory (e.g., Buss et al., 1992). American Psychological Association, 2000) and those not
However, inconsistent findings and small effect sizes make it diagnosed with this disorder. This approach is suggested
difficult to draw firm conclusions regarding sex differences in because delusional disorderjealous typeis the only diag-
jealousy (DeSteno et al., 2002; Edlund & Sagarin, 2009). It nostic category with standardised DSM-IV-TR criteria related
is also important to consider that two thirds of the overall to jealousy. This disorder involves jealousy associated with
sample were females. Future research involving a larger delusions of partner infidelity. It is acknowledged, however,
sample of male participants is needed to further investigate that this research would be limited by the small percentage
gender differences in the dimensions of romantic jealousy. of the population (estimated between 1% and 4%) who are
Again, the present results regarding sex differences support diagnosed with this disorder. However, it is possible that
our contention that it is important to adopt a multidimen- future revisions of the DSM-IV-TR may incorporate broader
sional conceptualisation of romantic jealousy. categories that take into account jealousy-specific disorders
Scores on cognitive jealousy were highly skewed; (Easton, Shackelford, & Schipper, 2008). Similarly, the
however, exploratory and CFAs appear to be relatively SF-MJS could be applied and tested in comparisons of
robust against violations of normality (Gorsuch, 1983). treatment-seeking samples of individuals and/or couples
Further, removal of outliers did not contribute to substantive with jealousy-related concerns and those who are not
changes in the pattern of results. It is possible that this seeking treatment.
dimension is not normally distributed in the general popu- Overall, the SF-MJS has the potential to be used as a
lation. Specifically, high levels of cognitive jealousy were not preliminary screen to identify individuals who score high
as common as high levels of emotional jealousy. In line with in terms of cognitive, emotional, and behavioural levels of
Pfeiffer and Wongs (1989) conclusions, high levels of cog- jealousy; particularly high levels of cognitive jealousy may
nitive jealousy may indicate more severe levels of jealousy, warrant further assessment. Questionnaires are especially
whereas emotional jealousy may signify more functional relevant for clinical practice in that they represent a time-
jealousy. efficient way to gather information about jealousy in the
The potentially adaptive function of jealousy is also high- current relationship. In this way, the SF-MJS is one tool to
lighted by other researchers (e.g., Barelds & Barelds-Dijkstra, guide assessment and treatment of individuals and couples
2007) who have measured types of jealousy that closely that may buffer against ongoing psychological illness and
resemble the present dimensions. In Barelds and Barelds- negative relational outcomes.
Dijkstras (2007) study, the experience of reactive jealousy At present, no published measure of romantic jealousy has
(similar to emotional jealousy) was proposed as a positive been validated and consistently used in the research litera-
relationship phenomenon, based on its positive links to rela- ture. This study was designed to overcome these limitations.
tionship satisfaction. Future research is needed to investigate Hence, we provide the first validated measure of romantic
2011 The Australian Psychological Society
250 R.A. Elphinston et al.

jealousy (SF-MJS) that is brief and easy to administer and Barnett, O. W., Martinez, T. E., & Bluestein, B. W. (1995). Jealousy
interpret. Additional strengths of the present study include and romantic attachment in maritally violent and nonviolent
men. Journal of Interpersonal Violence, 10, 473486.
a reasonable sample size (N = 326) with a broad age range, Bartlett, M. S. (1950). Tests of significance in factor analysis. British
and the direct application of a theoretically-derived measure Journal of Psychology, Statistical Section, 3, 7785.
to a sample from the general population. Given that an Braithwaite, V. A. (1987). The scale of emotional arousability: Bridg-
additional item was removed in the CFA, we recommend ing the gap between the neuroticism construct and its measure-
ment. Psychological Medicine, 17, 217225.
that cross-validation with another independent sample Brennan, K., Clark, C., & Shaver, P. (1998). Self-report measure-
would be needed to provide further evidence for the 17-item ment of adult attachment: An integrative overview. In J. A.
SF-MJS (Tabachnick & Fidell, 2007). A larger sample would Simpson & W. S. Rholes (Eds.), Attachment theory and close relation-
have been desirable for the independent EFA and CFA ships (pp. 4676). New York: Guilford Press.
Bringle, R. G. (1982). Preliminary report on the revised self-report
analyses; however, using Guadagnoli and Velicers (1988, as jealousy scale. Unpublished manuscript.
cited in Thompson, 2004) factor saturation criteria, the Bringle, R. G., Roach, S., Andler, C., & Evenbeck, S. (1979).
sample size was appropriate for the EFA, while using Strein- Measuring the intensity of jealous reactions. Catalog of Selected
ers (1994) criteria with five participants per variable, the Documents in Psychology, 9, 2324.
Buss, D. M., Larsen, R., Western, D., & Semmelroth, J. (1992). Sex
sample size for the CFA was also adequate. In addition, differences in jealousy: Evolution, physiology, and psychology.
approximately two thirds of the sample were females. Given Psychological Science, 3, 251255.
the sample size in this study, it was not possible to assess the Buunk, B. P. (1997). Personality, birth order and attachment styles
validity of the SF-MJS on the basis of gender. It could be as related to various types of jealousy. Personality and Individual
Differences, 23, 9971006.
argued that our inclusion criterion of 2 months for relation- Buunk, B., Angleitner, A., Oubaid, V., & Buss, D. M. (1996). Sexual
ship duration may have resulted in a loss of data from and cultural differences in jealousy: Tests from the Netherlands,
individuals highest in attachment anxiety (whose relation- Germany, and the United States. Psychological Science, 7, 359
ships often break up early). However, the sample covered a 363.
Carson, C. L., & Cupach, W. R. (2000). Fueling the flames of the
broad range of scores on attachment anxiety. green-eyed monster: The role of ruminative thought in reaction
In summary, the SF-MJS is the first measure of romantic to romantic jealousy. Western Journal of Communication, 64, 308
jealousy to be validated in samples of individuals currently in 329.
a romantic relationship, using both EFA and CFA statistical DeSteno, D., Bartlett, M., Braverman, J., & Salovey, P. (2002). Sex
difference in jealousy: Evolutionary mechanism or artifact of
analyses. Further, the current research provides descriptive measurement? Journal of Personality and Social Psychology, 83,
data on the dimensions of the SF-MJS. The psychometric 11031116.
characteristics of the scale are strong, with demonstrated Easton, J. A., Shackelford, T. K., & Schipper, L. D. (2008). Delusional
construct validity and internal reliability across the three disorderjealous type: How inclusive are the DSM-IV diagnostic
criteria. Journal of Clinical Psychology, 64, 264275.
dimensions. The next phase of validation of the SF-MJS Edlund, J. E., & Sagarin, B. J. (2009). Sex differences in jealousy:
requires a replication of the current research in a new, larger Misinterpretation of nonsignificant results as refuting the theory.
sample, with a focus on sampling a greater number of male Personal Relationships, 16, 6778.
participants. Future research may also aim to establish nor- Gorsuch, R. L. (1983). Factor analysis (2nd ed.). Hillsdale, NJ:
Erlbaum.
mative data and to investigate the application and validation Guerrero, L. K. (1998). Attachment-style differences in the experi-
of the SF-MJS in a clinical sample. Overall, the current study ence and expression of romantic jealousy. Personal Relationships,
supports the utility of the SF-MJS as a reliable and valid 5, 273291.
measure of romantic jealousy. Guerrero, L. K., & Andersen, P. A. (1998). Jealousy experience and
expression in romantic relationships. In P. A. Andersen & L. K.
Guerrero (Eds.), Handbook of communication and emotion: Research,
theory, applications, and contexts (pp. 155188). San Diego, CA:
Academic Press.
Guerrero, L. K., & Eloy, S. V. (1992). Relational satisfaction
REFERENCES
and jealousy across marital types. Communication Reports, 5,
2331.
Afifi, W. A., & Reichert, T. (1996). Understanding the role of Guerrero, L. K., Andersen, P. A., Jorgensen, P. F., Spitzberg, B. H., &
uncertainty in jealousy experience and expression. Communica- Eloy, S. V. (1995). Coping with the green-eyed monster: Concep-
tion Reports, 9, 93103. tualising and measuring communicative responses to romantic
Altman, I., & Taylor, D. A. (1973). Social penetration: The development jealousy. Western Journal of Communication, 59, 270304.
of interpersonal relationships. Oxford: Holt, Rinehart & Winston. Guerrero, L. K., Spitzberg, B. H., & Yoshimura, S. M. (2004). Sexual
American Psychological Association. (2000). Diagnostic and statistical and emotional jealousy. In J. H. Harvey, A. Wenzel, & S. Sprecher
manual of mental disorders (4th ed., text rev.).Washington, DC: (Eds.), The handbook of sexuality in close relationships (pp. 311344).
Author. Mahwah, NJ: Lawrence Erlbaum.
Barelds, D. P. H., & Barelds-Dijkstra, P. (2007). Relations between Harmon-Jones, E., Peterson, C. K., & Harris, C. R. (2009). Jealousy:
different types of jealousy and self and partner perceptions of Novel methods and neural correlates. Emotion, 9, 113117.
relationship quality. Clinical Psychology and Psychotherapy, 14, 176 Kaiser, H. E. (1970). A second generation little jiffy. Psychometrika,
188. 35, 401416.

2011 The Australian Psychological Society

 Measuring romantic jealousy 251

Kaiser, H. E. (1974). An index of factorial simplicity. Psychometrika, Tabachnick, B. G., & Fidell, L. S. (2007). Using multivariate statistics
39, 3136. (5th ed.). Boston: Pearson/Allyn & Bacon.
Kline, R. B. (2005). Principles and practices of structural equation Theiss, J. A., & Solomon, D. H. (2006). Coupling longitudinal data
modeling. New York: Guilford. and multilevel modeling to examine the antecedents and conse-
Knobloch, L. K., Solomon, D. H., & Cruz, M. G. (2001). The role of quences of jealousy experiences in romantic relationships: A
relationship development and attachment in the experience of test of the relational turbulence model. Human Communication
romantic jealousy. Personal Relationships, 8, 205224. Research, 32, 469503.
Mathes, E. W., & Severa, N. (1981). Jealousy, romantic love, and Thompson, B. (2004). Exploratory and confirmatory factor analysis:
liking: Theoretical considerations and preliminary scale develop- Understanding concepts and applications. Washington, DC: American
ment. Psychological Reports, 49, 2331. Psychological Association.
Melamed, T. (1991). Individual differences in romantic jealousy: The Ullman, J. B. (2001). Structural equation modelling. In B. G.
moderating effect of relationship characteristics. European Journal Tabachnick & L. S. Fidell (Eds.), Using multivariate statistics (4th
of Social Psychology, 21, 455461. ed., pp. 653771). Needham Heights, MA: Allyn & Bacon.
Pfeiffer, S. M., & Wong, P. T. P. (1989). Multidimensional jealousy. White, G. L. (1981). A model of romantic jealousy. Motivation and
Journal of Social and Personal Relationships, 6, 181196. Emotion, 5, 295310.
Rydell, R. J., & Bringle, R. G. (2007). Differentiating reactive and White, G. L., & Mullen, P. E. (1989). Jealousy: Theory, research, and
suspicious jealousy. Social Behavior and Personality, 35, 10991114. clinical strategies. New York: The Guilford Press.
Sharpsteen, D. J., & Kirkpatrick, L. A. (1997). Romantic jealousy Wiederman, M. W., & Allgeier, E. R. (1993). Gender differences
and adult romantic attachment. Journal of Personality and Social in sexual jealousy: Adaptionist or social learning explanation?
Psychology, 72, 627640. Ethology and Sociobiology, 14, 115140.
Streiner, D. L. (1994). Figuring out factors: The use and misuse of
factor analysis. Canadian Journal of Psychiatry, 39, 135140.

2011 The Australian Psychological Society

Copyright of Australian Journal of Psychology is the property of Wiley-Blackwell and its content may not be
copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written
permission. However, users may print, download, or email articles for individual use.