Current Psychology

, Volume 31, Issue 4, pp 365–380

Assessing Vindictiveness: Psychological Aspects by a Reliability and Validity Study of the Vengeance Scale in the Italian Context


  • Simona Ruggi
    • e-Campus University
  • Gabriella Gilli
    • Department of PsychologyCatholic University of Sacred Hearth
  • Noreen Stuckless
    • Department of PsychologyYork University
    • Department of PsychologyCatholic University of Sacred Hearth
    • Department of PsychologyCatholic University of Sacred Hearth

DOI: 10.1007/s12144-012-9153-2

Cite this article as:
Ruggi, S., Gilli, G., Stuckless, N. et al. Curr Psychol (2012) 31: 365. doi:10.1007/s12144-012-9153-2


Vengeance can be commonly defined as the disposition towards the infliction of harm in return for perceived injury or insult or as simply getting back at another person. This paper describes a contribution to the Italian validation of the Vengeance Scale (Stuckless and Goranson, Journal of social Behavior and Personality 7: 25–42, 1992) following the same steps of the original authors and shows psychological implications of vindictive behavior. 377 under-graduate students responded to the Big Five Questionnaire, State Trait Anger Expression Inventory and a back-translated Italian version of the Vengeance Scale (IVS). The IVS shows good psychometric properties. Convergent validity is shown by correlations with crucially connected variables (anger, empathy, social desirability). Factorial analysis suggested that the IVS is basically a one-dimensional measure. Regression analysis reveals that empathy, anger and emotional stability are significant predictors of vengeance. General results show that the IVS is a good instrument of evaluation of the tendency to be vindictive. Statistic analysis highlights that specific personality traits are involved in vindictive behavior; furthermore the interactions between some features of subject and the environment appear determinant. The implications and utility of the IVS in future research are discussed.


VindictivenessAngerPersonalityVengeance Scale


The focus of this contribution is the study of vengeance, or the attitude to perform vengeful acts. Our starting point was the operationalization of the vengeance construct in an Italian context, through analysis of a questionnaire that was originally created in Canada (Stuckless and Goranson 1992). We also intend to confirm the idea of the measurability of the vengeance construct and to provide a psychological interpretation about vindictive behavior.

Generally speaking, a dispositional and trait-based interpretation of personality would define vengeance as an individual phenomenon that is quite stable and dependent on the individual’s psychic and experimented features. Stuckless and Goranson (1992) defined vengeance as “the infliction of harm in return for perceived wrong” (1992, 25): their study and measurement instruments of the concept are used in this paper. Even if the logic of vindictive behaviors is the “montecristo” one – tit for tat –, Stuckless and Goranson warn that they need to be distinguished from other similar behaviors. For example, punishing behavior does not imply the rage that is related to the vindictive act, because it is pedagogical in some way. Vengeance is also motivated by a resentment following the perception of some harm and is therefore different from generalized hostility supported by undifferentiated feelings of aversion towards others. Another differential concept is retaliation: in fact, if retaliation is not used as a form of prevention, generally it stems, as vengeance does, from experienced harm, but is not generally motivated by the desire of obtaining satisfaction through reciprocation.

McCullough et al. (2001) found three main motivations at the origin of vengefulness: getting retribution to re-establish the moral order through a remedial act; giving the perpetrator (Baumeister 1997) what he/she deserves in a moral-pedagogical way; and keeping up appearances and regaining self-esteem and value in front of the offender.

Gouldner (1960), examining analyses conducted by several sociologists and anthropologists – such as Merton, Parsons, Malinowski – explained vindictive attitude through reciprocity. He claimed this norm to be universal, though it can assume peculiar features depending on different cultures. In this approach, the desire for vengeance can be understood as a specific form of reciprocity. In detail, Eisenberger et al. (2004) stated a negative norm of reciprocity which involves an unequivocal set of beliefs about the adequacy of vengeance as the proper answer to an offence (reciprocation to unfavorable treatment). Stuckless and Goranson (1992) criticized the link between an attitude towards vengeance and a negative norm of reciprocity. Individual differences in the form and strength of such a norm can depend on various factors, for example the educational level or responsiveness of the other: rewards and recognitions following vindictive behaviors play a crucial role in the perception of vengeance as being adequate. In other words, the context in which the subject grows up is an extraordinarily relevant variable (Bowlby 1973, 1988; Kohut 1977).

The beliefs about the innate and generalized malevolence of people, and the presence of personality traits concerning impulsiveness and the will to dominate are essential aspects. In his studies Caprara (2004) also addressed some themes pertaining to individual aggression and violence, affections and interpersonal relations, conflicts among groups and nations, altruism and cooperation. Examining the determinants and the motivational components of guilt, he found relations between negative affections, need for reparation and fear of punishment (Caprara et al. 2001).

Vindictiveness, defined as a personality trait, has substantial connections with the dimensions of Neuroticism and Agreeableness. Studies by McCullough et al. (2001) show that persons who are very sensitive to negative events, and therefore have high levels of Neuroticism, are predisposed to suffering offence and behaving vindictively. Similarly, persons showing low levels of pleasantness are characterized by a difficulty in empathically opening towards relationships and, being quite selfish and not very kind, emphasize explicit vindictive intentions. Another determining characteristic in the activation of vindictive actions is the predisposition towards anger, interpreted as an effort to respond to an offence with a willingly aggressive action against the person perceived as the culprit (McCullough et al. 2001).

The vindictiveness concept is also intertwined with concepts of forgiveness and narcissism. McCullough et al. (2003) emphasized that individuals with high levels of narcissism tend towards more aggressive responses to negative feedback from the environment, and to interpret others’ actions as attacks against their own selves, with a blaming finality. They would thus justify their own actions, framing them as defensive behaviors. Even without a determinate link, literature (Stuckless and Goranson 1992; McCullough et al. 2001; Brown 2003) has shown that vindictiveness and the ability to forgive are dispositional factors, connected to the personality organization of the individual. Brown (2004) also tried to answer one question that psychological research left unresolved, namely the relationship between forgiveness and vindictiveness. Brown maintained that the most important personality factor in determining the search or refusal of vengeance is narcissism. Brown’s study has also been confirmed by studies that showed a link between narcissism and low capacity to forgive (Emmons 2000). Specifically, narcissism appears to be positively correlated to hostility and to the tendency to increase the positive consideration of oneself (Kernis and Sun 1994) and negatively correlated to empathy (Watson et al. 1984). The lack of empathy induces the strongly narcissistic individual to react to offence by giving back the damage, putting into action therefore a mechanism of compensation for the Self with all the characteristics of vengeance.

An interpretation very interesting of vengeful acts comes from the psychodynamic framework (Oasi and Massaro 2004). K. Horney, who since 1948 has spoken about patients labeled as “arrogant-vindictive”, imputed their behavior and mood to infantile experiences, whose features were: sheer brutality, humiliations, derision, neglect, and flagrant hypocrisy (Horney 1950). In general, the study of the primal relationship experienced by vindictive subjects is crucial for the majority of psychoanalytic contributions (Socarides 1966; Daniels 1969; Bowlby 1973, 1988; Kohut 1977). The clinical approach suggests that these patients are full of resentment and grudges. Invaded by these emotions they tend to “colonize” and aggressively control others (Kancyper 2003) or to build psychic retreats: an internal situation in which they feel trapped and offended, but are not able to obtain justice in the external world (Steiner 1993).

We took these important points of view and interpretations about the disposition to vindictive acts into account. Nevertheless the present study focused on the reliability and validity of Vengeance Scale in the Italian context, following the version proposed by Stuckless and Goranson (1992). Anyway important psychological aspects are discussed starting from collected data.



A total of 377 undergraduates were recruited from different university courses: Psychology (228), Engineering (57), Law (49) and Sociology (43). The sample included 133 males and 244 females. The age range was from 18 to 56 years (Mean = 20.8, Standard Deviation = 3.6). The make-up of the group of participants was similar to the one suggested in the study by Stuckless and Goranson (1992), both in the Faculties that had been chosen and in the ratio between males and females. Although the group of participants was mainly composed of students from the Faculty of Psychology, since they were easier to contact, the total number of participants was considered wide enough for suitable validity analyses.

In the Italian language there are no instruments to measure the tendency to vindictiveness. In order to contribute to validating the instrument, we therefore chose instruments measuring similar variables (the same emphasized in the original study; Stuckless and Goranson 1992), which might be useful for convergent validity.


The following self-report instruments were used.
  • VS (Vengeance Scale by Stuckless and Goranson 1992), in the Italian Version by Oasi, Ruggi, Gilli in (Appendix). We used the 20-item version by Stuckless and Goranson (1992). The Vengeance Scale was translated into Italian by an English native speaker with an expertise in scientific psychology translation. The Italian version was then translated back into English by another translator, in order to verify its linguistic adequacy. The scale proposes 20 statements regarding the desire/proposal of vengeance or the ethic/moral judgment about vindictiveness. The subject must express his/her degree of agreement with every statement on a 7-step scale (from “I completely disagree” to “I completely agree”). In the Appendix the English and Italian versions of VS are given.

  • STAXI (State-Trait Anger Expression Inventory by Spielberger 1988). Italian version by Comunian (1992), which requires the subject to describe his/her personal inclination to anger, in its different types of expression. There can be anger as a state, as a trait, as temperament, and as expression aimed outside or inside, or as ability to control it. The test is composed of 44 items, to which the subject must answer on a 4-step scale.

  • BFQ (Big Five Questionnaire by Costa and McCrae 1985). Italian Version by Caprara et al. (1993), 132 items which allow the subjects to describe their personality characteristics, regarding Energy (dynamism and dominance), Friendliness (cooperativeness/empathy and politeness), Conscientiousness (scrupulousness and perseverance), Emotional Stability (emotional control and impulse control), Openness (openness to culture and openness to experience). The subject replies to the 132 items on a 5 step scale. Each of the ten sub-factors is described by 12 items; half of the statements are given in a positive way, while the other half are given in a negative way, to check possible set-response events. The remaining 12 items are from the Lie scale.

  • In order to measure Social Desirability, we used the Lie scale within the Big Five Questionnaire (Caprara et al. 1993). According to the authors’ definition, the Lie scale is an expression of Social Desirability, since it measures the tendency of the subject to give a distorted image of him/herself. The subject replies to 12 items on a 5 step scale. We think that, offering the questionnaires in an anonymous format, the subjects will describe themselves sincerely, without the need to show themselves as desirable.


All instruments were administered to the subjects in a single session and in the following order: 1) STAXI; 2) VS; 3) BFQ. The time for the administration of the whole battery was about one hour. The entire administration was attended by the researcher in charge. A re-test was performed on 70 subjects two months later. All 377 research participants met the subject/item ratio criterion from 5:1 to 10:1 advocated by Nunnally (1978): exactly 18.85 subjects per item. Thanks to its anonymity it is felt that the subjects answered sincerely, thus assuring the reliability of data.


The mean of the 20 item scale was 64.48 and standard deviation was 18.91 on a scale from 20 to 140.


The Italian version of the Vengeance Scale had a Cronbach’s alpha of .88. The alpha was .90 for males and .87 for females. The scale also had an acceptable means inter-item correlation of .28 (Briggs and Cheek 1986).


The BFQ and STAXI scores are shown in Table 1. These are the starting point for the analysis.
Table 1

State-Trait Anger Expression Inventory and Big Five Questionnaire: descriptive statistics



Minimum scores

Maximum scores



Anger as a state






Anger as a trait






Anger-prone temperament






Anger reaction






Anger directed inside






Anger directed outside






Anger control






Anger expression




























































Emotion control






Impulse control






Emozional stability






Openness to culture






Openness to experiences












Lie scale






Not all subjects filled in all item of BFQ (see also Table 3)

Convergent validity is demonstrated by correlations with related scales (Cohen and Cohen 1983). We can observe (Table 2) correlations between vindictiveness and a tendency towards anger, especially with the measure of angry feelings at a certain moment (S-Anger, anger as a state), with the measure of disposition towards feeling anger (T-Anger, anger as a trait) – which includes a general disposition to feel or express angry feelings without specific reasons (T-Anger/T, anger-prone temperament) and a tendency to express anger if criticized or threatened (T-Anger/R, angry reactions). Furthermore, we obtained significant positive correlations between vindictiveness and the measure of the expression of anger toward persons and objects (AX/Out, anger directed outside) and, finally, with the measure of the general index of the frequency of the expression of anger (AX/EX, expression of anger). On the other hand, there is a negative correlation between vindictiveness and the measure of the ability to control the anger (AX/Con, anger control) and there is no significant correlation between vindictiveness and the tendency to restrain angry feelings (AX/In, anger directed inside).
Table 2

Correlations between Vengeance Scale and State-Trait Anger Expression Inventory (n = 377)

Anger expression (STAXI)

Vengeance (Vengeance Scale)

Anger as a state


Anger as a trait


Anger-prone temperament


Anger reaction


Anger directed inside


Anger directed outside


Anger control


Anger expression


** Significant correlation at the 0,01 level (2-tail)

* Significant correlation at the 0,05 level (2-tail)

The correlations between vindictiveness and personality factors are presented in Table 3. We can observe a significant positive correlation between vindictiveness and the Energy factor (tendency towards activity, extroversion) on one hand, and the sub-factor dominance (tendency to impose oneself) on the other. However there are negative correlations between vindictiveness and the Friendliness factor (co-operative, altruistic abilities), and the sub-factors of cooperativeness/empathy and politeness (openness toward others), with the Conscientiousness factor (tendency to reflection and order), and with the sub-factor of scrupulousness; with the Emotional Stability factor (ability to control anxiety and emotionality), and the sub-factor of impulse control (ability to control personal reactions, also in situations of distress) and with Openness factor, in particular sub-factor openness to culture (interest towards knowledge). No significant correlations emerge with the Lie scale, understood as being as a measure of social desirability, thus attesting that the Vengeance Scale is not contaminated by social desirability.
Table 3

Correlations between Vengeance Scale (n = 377) and Big Five Questionnaire

Personality factor (BFQ)

Vengeance (Vengeance Scale)

Dynamism (n = 377)


Dominance (n = 377)


Energy (n= 377)


Cooperativeness/Empathy (n = 377)


Politeness (n = 377)


Friendliness (n= 357)


Scrupulousness (n = 366)


Perseverance (n = 368)


Conscientiousness (n= 351)


Emotion control (n = 372)


Impulse control (n = 369)


Emotional stability (n= 349)


Openness to culture (n = 360)


Openness to experiences (n = 366)


Openness (n= 362)


Lie scale (n = 370)


** Significant correlation at the 0,01 level (2-tail).

In bold font the five main factor of personality in BFQ. Not all subjects filled in all item of BFQ (see also Table 1), but the scoring procedure guarantees the reliability of results

A linear regression analysis (Jaccard et al. 1990) shows that the scores of the Italian version of the Vengeance Scale is significantly predicted by the tendency to be in anger in general (R2 change = .21; F = 13.40, p < .001), and specifically by the tendency to be in anger in a specific moment: we could say to be more impulsive. Shortly the Anger as a state or the Angry in reaction to be anger. Whereas it is negatively predicted by the Anger directed inside and the Ability to control anger (Table 4).
Table 4

Regression analysis: State-Trait Anger Expression Inventory and Vengeance Scale


Non-standardised coefficients


Standardised coefficients





Std. Error









Anger as a state






Anger as a trait






Anger-prone temperament






Anger reaction






Anger directed inside






Anger directed outside






Control of anger






Anger expression






Dependent variable: Vengeance Scale scores

The same linear regression analysis also shows that the scores of the Italian version of the Vengeance Scale is positively predicted (R2 change = .30; F = 27.05, p < .001) by some personality factors such as Energy, and negatively by Friendliness, Conscientiousness and Emotional Stability (Table 5). Linear regression evaluating the predictability of Cooperativeness/Empathy on the scores of the Italian version of the Vengeance Scale (as in the study by Stuckless and Goranson 1992) showed significant negative results: (R2 change = .20; F = 91.83, p < .001; beta = −.444, p < .000).
Table 5

Linear regression analysis: Big Five Questionnaire and Vengeance Scale


Non-standardised coefficients


Standardised coefficients





std. error



























Emotional stability












Lie scale






Dependent variable: Vengeance Scale scores

Factor Analysis

The Kaiser-Meyer-Olkin measure of sampling adequacy and the Bartlett Test of Sphericity were conducted on the data from the Italian version of the Vengeance Scale before factor extraction. This ensured that the characteristics of the data set were suitable for the conduction of the factor analysis. The Kaiser-Meyer-Olkin analysis yielded an index of 0.905 in concert with a highly significant Bartlett Test of Sphericity (Chi-square = 2662,918; df = 190; p < .001).

A factor analysis of principal components (Comrey and Lee 1992) with oblique rotation (Varimax rotation yielded essentially the same results) (Comrey 1967) on the Italian version of the Vengeance Scale gives as result a four factor structure, accounting for 54.13 % of the total variance. The four factors account for 23.41 %, 12,62 %, 9,69 % and 8,41 % of the total variance respectively. An application of the scree test (Cattell 1966, 1978) indicated that only the first factor should be retained. This result is confirmed by the eigenvalues analysis: the four eigenvalues amount to 6,690; 1,787; 1,264; and 1,060 respectively; from the five factors the eigenvalue is less than 1. Forcing a single factor solution resulted in loadings for each item of .40 or above. Item n° 1, “It’s not worth my time or effort to pay back someone who has wronged me”, whose factorial weight is .30 above, could be excluded. These results suggest that the Italian version of the Vengeance Scale is basically a uni-dimensional measure (Fabrigar et al. 1999).

Confirmative Factor Analysis

In order to assess the goodness of fit of the model a confirmatory factorial analysis was conducted using Amos Graphic. In structural equation models, results can be interpreted, on the one hand, at a level of the goodness of fit of the entire model, and, on the other, at a level of the significance of single parameters (Corbetta 1992). As far as the global goodness of fit of the model is concerned, the following indices have been considered in the present study: it is expected that the value of chi-square be associated to a non-significant probability. However, such a statistic is sensitive to the dimensions of the sample: numerous samples tend to show significant chi-square, even in cases of a good fit of the model to data (Corbetta 1992); the RMSEA (Root Mean Square Encor of Approximation), acceptable when less than 0.08 (Brown and Cudeck 1993); the CFI (Comparative Fit Index) acceptable when more than 0.9 (Bentler 1990).

The first model (Fig. 1) we tested did not show acceptable fit indices (chi-square = 632,1; df = 152; RMSEA = 0,092; CFI = 0,808).
Fig. 1

Confermative factor analysis - model 1. Chi-square = 581,7; df = 135; RMSEA = 0,094; CFI = 0,817

In a new model we deleted item 5, whose weight in the factorial analysis resulted already low (0,346), but fit indices was not good also in this case (chi-square = 581,7; df = 135; RMSEA = 0,094; CFI = 0,817).

From the output analysis, and particularly by indices modifications, the necessity of correlating the errors referring to the following items emerged: error item 2 - error item 3; error item 3 - error item 7; error item 11 - error item 13; error item 12 - error item 17; error item 2 - error item 20; error item 14 - error item 15; error item 16 - error item 18; error item 18 - error item 19 (Fig. 2). Even though this solution is criticized by some,1 in this case it is allowed by the similarity content of the implied items. For example, both item 2, It is important to me to get back at people who have hurt me and item 3, I try to even the score with anyone who hurts me express a tendency to explicit vengeance, as a sudden reaction to a hurt. Again, both item 11 Revenge is morally wrong and item 13 People who insist on getting revenge are disgusting imply a moral evaluation of the vengeance. In order to evaluate whether this procedure would be adequate, we modified the model correlating the errors of the implied items. This modification contributed to enhancing the fit indices of the model (RMSEA = 0,068; CFI = 0,91), even though the chi-square continues to be significant (chi-square = 349,6; df = 127). This model (Fig. 2) confirmed the explorative factorial analysis deleting item 5.
Fig. 2

Confermative factor analysis - model 2. Chi-square = 349,6; df = 127; RMSEA = 0,068; CFI = 0,91

Test-Retest Reliability

As Stuckless and Goranson (1992) proposed, we checked the reliability of the Vengeance Scale by a test-retest procedure. Two months after first compilation, the Vengeance Scale is resubmitted to 70 students of Psychology. The anonymity is guaranteed. The test - retest correlation was r(70) = .90, p < .01 (Traub 1994).

Conclusion Remarks

The work here reported aimed to develop a reliable and valid measure of vindictiveness using the Scale developmental procedures suggested by Stuckless and Goranson (1992). The Scale that emerged from the validation study is a highly reliable instrument (as evidenced by high alpha coefficients, inter-item correlations and test-retest reliability) with good psychometric properties.

The good psychometric properties of the instrument, particularly its convergent validity, are demonstrated by the significant values of the correlations with the measures of other instruments (Caprara et al. 1993; Spielberger 1988); with the sub-dimensions of the tendency to anger2; with the personality characteristics, and specifically, with the tendency to empathy, as emerges in Stuckless and Goranson (1992).

The results of the regression analysis show that the State Anger, Anger Reaction and Anger Expression have a positive influence on the scores of Italian version of Vengeance Scale. However, Inward Anger has a negative influence. Moreover it appears that contextual, situational aspects of anger are connected to the attitude towards vengeance, and to a stable personality trait (anger as a trait, anger-prone temperament) in a lesser degree. The State Anger weight in predicting the scores of the Vengeance Scale allows to considerer the great influence of the context featuring trigger events. This result agrees with much evidence in the literature, particularly the psychoanalytic ones (Socarides 1966; Daniels 1969; Kohut 1977; Bowlby 1973; Lane 1995). Many of these authors underline that childhood experiences can be the starting point of a specific aggressive reaction towards the object, and of a particular personality traits (Horney 1948).

Moreover, in accordance to Goranson and Stuckless’ contribution (1992), the linear regression analysis shows the scores of the Italian version of the Vengeance Scale are negatively predicted by Friendliness, Conscientiousness and Emotional Stability and positively predicted by Energy of BFQ (McCullough et al. 2001; McCullough et al. 2003). High scores in the disposition towards vindictiveness are predicted by specific sub-factor of the Energy factor like dominance and the taking part to a context. On the contrary, low scores in the disposition towards vindictiveness are predicted 1) by specific sub-factor of the Friendliness factor like generosity, cordiality, cooperativeness, and empathy; 2) by specific sub-factor of the Conscientiousness factor like self-control and self-regulation; 3) by specific sub-factor of the Emotional Stability factor like impulsivity, irritability, and vulnerability. From the psychoanalytic standpoint it is possible to consider that childhood experiences of mirroring (Kohut 1977) or of secure attachment (Bowlby 1973) may enhance the development of specific personality traits able to prevent vindictive behavior.

The factor analysis confirms the mono-factorial structure (Stuckless and Goranson 1992). The factorial weight of item 1 and 5 (item1: “It’s not worth my time or effort to pay back someone who has wronged me”; item5: I live by the motto “Let bygones be bygones”) is low, and therefore not very representative of the scale. A confirmatory factorial analysis shows the goodness of the model on a single factor (tested on 18 items; item 1 and item 5 was excluded). The mono-factorial structure is confirmed through acceptable indices of fit. So, the Italian version of Vengeance Scale is composed of 18 items.

Thanks to anonymity the subjects are felt to be sincere and the scores of the Lie Scale of BFQ were adopted to monitor the disposition of the subject to give a not true profile while he filled in the Italian Vengeance Scale.

The general applicability of the present results is limited. The evidence for concurrent validity is based entirely on correlations with data based on hypothetical and self-reported situations. However, since vengeance, by its very nature, typically involves serious acts of aggression, there are implicit difficulties. A second consideration regarding the general applicability of the present results arises from the fact that the subjects were all university undergraduates. Further research could examine the relationship between the vengeance construct and the temperament construct (Cloninger 1994) or mental functioning in particular clinical condition like Depression (Calati et al. 2010); also it could be interesting to explore the link between the vengeance construct and the fear of punishment and the need for reparation, extending studies by Caprara et al. (2001) on the motivational components of guilt.

Finally, the present Italian version of the Vengeance Scale could give rise to new very interesting cross cultural studies (Anolli 2011) or researches in specific setting (Castelnuovo et al. 2003), examining the influence of cultural, educational and clinical aspects as compared to personality factors in the towards vengeance tendency.


In the Corbetta’s opinion (1992) a relationship among the errors means admitting the inadequacy of the model.


Anger as a state, anger as a trait, anger-prone temperament, anger reaction, anger directed outside, expression of anger, and, negatively, with anger control.


Copyright information

© Springer Science+Business Media New York 2012