News & Politics

Testing the Psychometric Properties of the Interpersonal Reactivity Index (IRI) in Chile: Empathy in a different cultural context

Testing the Psychometric Properties of the Interpersonal Reactivity Index (IRI) in Chile: Empathy in a different cultural context
of 7
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
  A.M.Fernándezet al.: Psychometric Propertiesofthe IRIin Chile  European Journalof Psychological Assessment   2011; Vol. 27(3):179–185© 2011 Hogrefe Publishing Original Article TestingthePsychometricPropertiesoftheInterpersonalReactivityIndex(IRI)inChile Empathy in a Different Cultural Context AnaMariaFernández 1,2 ,MicheleDufey 2 ,andUweKramp 3 1 UniversidaddeSantiagodeChile,Chile,  2 UniversidadDiegoPortales,Santiago,Chile, 3 UniversidaddeChile,Santiago,Chile Abstract.  The psychometric properties of Davis’ (1980) Interpersonal Reactivity Index (IRI) in Chile were assessed. The IRI was appliedto a sample of 435 college students. Appropriate internal consistencies and test-retest stability resulted. The instrument’s validity wasevidenced by the interrelations among the scales, in addition to its correlations in the predicted direction to other related psychologicalconstructs, and sex differences emerged in three of its dimensions. A confirmatory factor analysis corroborated the theoretical structureof the IRI in Chile, and the suitability of both the four-factor model and a second order factor that integrates three of the dimensions .The implications and comparison of the results with other adaptations of the IRI are discussed. Keywords:  empathy, affectivity, assessment, validity, reliability, measurement An increasing tendency in psychometric research is to es-tablish whether psychological constructs conceived in aparticular cultural context may have a similar conceptual-ization in other countries and cultures (see for example,Matsumoto, Yoo, Hirayama, & Petrova, 2005). According-ly, the present study evaluated the validity and reliabilityof the Interpersonal Reactivity Index (IRI; Davis, 1980,1983) in Chile.Empathy is a versatile construct because it involvessympathy and feeling an emotion appropriate to the affec-tive circumstances of others, thus, evoking a cognitive pre-diction of others intentions and actions, or what Baron-Co-hen and Wheelwright (2004) have denominated  a theory of mind  . All of these aspects can be synthesized in a commoncharacteristic that underlies the components of empathythat “concern responsivity to others” (Cliffordson, 2002a,p.49).When Davis began the development of IRI, the authorwasawareofseveralscalesandquestionnairesthatallowedfor an estimation of people’s ability to empathize with oth-ers, but none of these differentiated the uniqueness of itscognitive and affective components (Davis, 1980, 1996).The IRI incorporates these two domains and allows for aninclusive concept of empathy by appraising cognitive abil-itiesonaFantasyScale(FS),andaPerspectiveTaking(PT)scale. The second of these domains contains the affectivecomponents of empathy assessed through the EmpathicConcern (EC) and Personal Distress (PD) scales. Addition-ally, others have proposed that these dimensions may yielda general empathy and an emotional control second-orderfactor (for example, Cliffordson, 2002a, 2002b; Pulos, Eli-son, & Lennon, 2004).Davis (1983) made several theoretical predictions aboutthe scope of his index and its relationship to other psycho-logical constructs, evidencing the validity of the IRI to as-sess empathy. For example, sex differences in three of thesubscales were confirmed, as well as interrelations amongthe dimensions of the IRI (PT with EC, as well as EC withFS). The convergent validity of each dimension was testedwith other measures of global empathy and interpersonalfunctioning. Particular positive correlations between PT,social functioning, and self-esteem were observed. ECshowed a consistent direct relationship with measures of shyness, anxiety, and emotionality. Likewise, PD was in-verselyrelatedtoself-esteemandinterpersonalfunctioning(emotionality). More recently, Hall, Davis, and Connelly(2000) have shown that the IRI is a good measure of theempathic domains that are related to career vocation, withhigh EC and PT more characteristic of applied psycholo-gists, while high PD appeared more characteristic of scien- DOI: 10.1027/1015-5759/a000065© 2011 Hogrefe Publishing  European Journal of Psychological Assessment   2011; Vol. 27(3):179–185  tists. Parental empathy assessed with the IRI has also beenassociated directly to children’s self-concept and adjustednarcissism, and inversely to depression (Trumpeter, Wat-son, O’Leary, & Weathington, 2008).Cross-cultural research with the IRI has established the-oretical and practical predictions about empathy. In Spain,the adapted scales correlate with other measures of proso-cial behavior, and an increase in PT from puberty to lateadolescence (which is consistent with the normal develop-ment of prosocial behavior) has been confirmed (Mestre,Frías, & Samper, 2004; Pérez-Albéniz, De Paúl, Etxe-berría, Montes, & Torres, 2003). Similarly, Oceda, Lopez-Perez, Ambrona, and Fernandez (2009) have shown an as-sociation between sympathy (assessed with the IRI) andvicarious distress, supporting the robustness of the cogni-tive and affective realms of empathy.Research on the psychometric properties of the IRI inother contexts led Cliffordson (2002a, 2002b) to adapt theinstrument to Sweden, finding that a general empathy fac-tor underlies three of the IRI scales, with the exception of PD. Similarly, Siu and Shek (2005) have arrived at a three-factorChineseversionoftheIRIinwhichthecognitiveandaffective domains are combined. In contrast, in Spain, re-search on the IRI has maintained its srcinal four-factorstructure (Mestre et al., 2004).In the present study, the theoretical and logical hypoth-eses of Davis (1980, 1983) were tested with Chilean col-lege students. The main objective was to adapt the instru-ment replicating Davis’ procedure in the local context, andwe hypothesized convergent relationships between PT andmeasures of self-esteem, and positive affect; EC and anxi-ety. Negative correlations of PD with self-esteem, socialavoidance, negative affect, and sensitivity to punishmentwere expected. Similarly, divergent relationships betweenPT, anxiety, and aggression were hypothesized. It was alsoassessed whether the proposed factor structure of the in-strument complies with current methodological proposalsregarding the configuration of empathy in a second-orderfactor (Cliffordson, 2002a). Method Participants and Procedure The participants in this study were 435 undergraduate stu-dents (201 men and 234 women) from different fields (en-gineering, psychology, journalism, accounting, and adver-tising). The mean age of the participants was 20.07 years( SD  1.9), with a range of 18 to 36 years. All participantsreceived standardized instructions,and gavetheirinformedconsentregardingvoluntaryparticipationandpersonaldatasuch as sex, age, and career. Measures Each respondent completed a self-report test battery consist-ing of the IRI, and some of them completed Rosenberg’sSelf-EsteemScale(RSE),Spielberger’sTraitAnxietyInven-tory (STAI-T),theBussandPerryAggressionQuestionnaire(BPAQ),theSocialAvoidanceandDistressScale(SAD),andthePositiveAffectandNegativeAffectSchedules(PANAS).TheIRIwasadaptedintolocalSpanishbytwobilingualpsy-chologists using the backtranslation method, a judgmentaltechnique for valid cross-cultural comparisons (Triandis &Berry, 1980). First, a bilingual psychologist translated theAmericanquestionnairesintoSpanish.Second,anotherbilin-gual psychologist retranslated (backtranslated) the Spanishadaptationsto English.Finally,the srcinalsourceand back-translated items were compared for equivalence of meaning,and anydiscrepancieswere noted.Thisreiterativeprocessof translation and backtranslation was continued until no se-mantic differences were noticed between the questionnaireforms. The other scales applied in this study had been previ-ously adapted to the Chilean context, and were used to testthe concurrent validity of the dimensions ofthe IRI. Allsub- jectscompletedthetestbatteryinonesessionlastingapprox-imately half an hour. Interpersonal Reactivity Index (IRI) TheIRIwasdevelopedbyDavis(1980,1983,1996)inordertointegratethemultidimensionalityofempathyintofourthe-oretical components: fantasy (the proclivity to identify withfictitious characters), perspective taking (the ability to adoptthe perspective of others in common life), empathic concern(thetendencytoexperiencefeelingsofcompassionandsym-pathy from others’ misfortune), and personal distress (thepronenesstofeeluncomfortable aboutthedistressofothers).Each dimension has seven statements, giving a total of 28items. It is a self-report instrument scored on a Likert-typescale ranging from 0  (doesn’t describe me at all)  to 4  (de-scribes me very well) . Rosenberg’s Self-Esteem Scale (RSE) The RSE is an adapted version of the srcinal instrument(Rosenberg, 1965) tapping into a global concept of Self-es-teem (Fernández, Celis-Atenas, & Vera-Villarroel, 2006). Itpresents10statementsscoredonaLikert-type scale,rangingfrom 1  (doesn’t agree at all)  to 5  (agree strongly) . Spielberger’s Trait Anxiety Inventory (STAI-T) The STAI-T measures Anxiety as a stable general person-ality characteristic, evaluated through 20 items that presentanxiety-producing scenarios in which respondents have toindicate if they identify with the situation  almost never   (1),180 A.M. Fernández et al.: Psychometric Properties of the IRI in Chile  European Journal of Psychological Assessment   2011; Vol. 27(3):179–185 © 2011 Hogrefe Publishing  to  almost always  (3). It was adapted to Chile by Vera-Vil-larroel, Celis-Atenas, Cordova-Rubio, Buela-Casal, andSpielberger (2007). Buss and Perry Aggression Questionnaire (BPAQ) The BPAQ was developed in order to evaluate four dimen-sions of Aggression: (a) physical aggression, (b) verbal ag-gression, (c) anger, and, (d) hostility. It includes 29 itemsthat are scored on a Likert-type scale 1  (extremely unchar-acteristic of me)  to 5  (extremely characteristic of me) 1 . Social Avoidance and Distress Scale (SAD) The SAD is an adapted general measure of the individualtendency to avoid social interactions and experience socialanxiety, which includes 29 true and false items 2 . PositiveAffectandNegativeAffectSchedules(PANAS) The PANAS assesses general emotionality with 20 itemsscored on a Likert-type scale anchored from 1  (very littleor nothing at all)  to 5  (extremely) . It was developed byWatson, Clark, and Tellegen (1988) and adapted to collegesamples in Chile by Dufey and Fernandez (2009). Data Analysis Data analysis is carried out in two steps. The first step (de-scriptive statistics and reliability) examined descriptive sta-tistics (mean and  SD ), internal consistency (Cronbach’s  α ),and test-retest stability (Pearson correlation coefficient). Allanalysiswere carriedoutformen,women,andthetotalsam-ple.Finally,temporalconsistencyoftheIRIscalesweretest-ed 60 days after the first application, with a subsample of 82participants. The second step (validity) examined structuralvalidity(evidencebasedontheinternalstructure)andpredic-tive validity (evidence based on the relationship with othervariables) of the IRI scales and Davis’ model. Structural va-lidity was assessed using Davis’ (Davis, 1980) and Clifford-son’s models (Cliffordson, 2002a).First step and predictive validity were performed usingSPSS v15.0.1.1. (SPSS, 2007). Structural validity was as-sessed by structural equation models (SEM; see Figure 1)using Mplus v.4.1 (Muthén & Muthén, 2006), with a maxi-mum likelihood estimation, with standard errors robust tononnormality, and Satorra-Bentler mean corrections to thegoodness-of-fitteststatistics(seeSatorra&Bentler,1994).Itis known that SEM models show a poor level of global ad- justment,measuredbyChisquareIndex(Bentler,1990;Bol-len, 1989; Bollen & Long,1993; Hair,Anderson,Tatham, &Black, 1999; Marsh, Hau, & Grayson, 2005). Thus, it is ad-visabletoestimatesomekindofapproximategoodness-of-fit(Bollen, 1989; Bollen & Long, 1993; Marsh et al., 2005;Maydeu-Oliveres,Kramp,García-Forero,Gallardo-Pujol,&Coffman, 2009; Raju, Laffitte, & Byrne 2002). In our case,the root mean square error of approximation (RMSEA; Stei-ger, 1990), comparative fit index (CFI; Bentler, 1990), andstandardizedrootmeansquareresidual(SRMR;Kline,2005;Marsh et al., 2005; Schweizer, 2010) were used. Figure 1shows the Davis and Cliffordson models that were tested. Results Descriptive Statistics and Reliability Table 1 provides the means,  SD , Pearson correlation coef-ficients ( r  ), and Cronbach’s  α s of the four scales for men,women, and the complete sample. Inspecting the resultsreveals significant sex differences among all the IRI sub- Table 1.  Descriptive statistics and reliability of the IRI dimensions, by sex and total sample IRIscalesMen ( n  = 201) Women ( n  = 234) Test by sex Total ( n  = 435)  Mean SD r   α  Mean SD r   α  t Mean SD  α FS 14.12 5.41 .82 .72 (.63–.79) 16.48 5.58 .76 .78 (.71–.83) –4.36** 15.41 5.62 .76 (.71–.80)EC 17.24 4.56 .89 .66 (.56–.75) 19.41 4.62 .81 .76 (.69–.81) –4.91** 18.40 4.71 .73 (.67–.78)PT 16.55 5.13 .67 .73 (.65–.80) 17.27 4.76 .67 .72 (.65–.79) –1.52 16.94 4.95 .73 (.68–.78)PD 11.10 4.79 .81 .67 (.57–.75) 13.20 4.97 .78 .70 (.61–.77) –4.46** 12.23 5.00 .70 (.64–.75)  Note . FS = fantasy scale, PT = perspective taking, EC = empathic concern, PD = personal distress.  r   = Test-Retest ( n  = 82);  T  -Test =  t   test forindependent samples;  α  = Cronbach’s  α  coefficient. **  p  < .001. A.M. Fernández et al.: Psychometric Properties of the IRI in Chile 181 © 2011 Hogrefe Publishing  European Journal of Psychological Assessment   2011; Vol. 27(3):179–185 1  The adapted version of the instrument was obtained from Figueroa Rey, P., Ramírez Troncoso, C.G. & Santis Doyhamboure, M. (2005).  Adaptación y validación del Buss-Perry aggression questionnaire a la población adolescente, adulta y adulta mayor del gran Santiago [Adaptation and validation of the Buss-Perry aggression questionnaire to adolescents, adults, and the elderly in Santiago]. Unpublishedundergraduate thesis, Universidad Diego Portales. 2  The local instrument we used is the adaptation of Pérez, A. & Sepúlveda, V. (1991).  Estandarización de las escalas de evitación y ansiedad social (SAD) y temor a la evaluación negativa (FNE), y relación entre la ansiedad social y los comportamientos asertivo y agresivo [Standardization of the avoidance and social anxiety scales (SAD) and fear of negative evaluation (FNE), and the relationship of socialanxiety, assertive, and aggressive behavior]. Unpublished undergraduate thesis, Universidad Católica de Chile.  Figure 1 . Davis’ and Cliffordson’s Models.  Notes:  FS = fantasy, PT = perspective taking, EC = empathic concern, PD =personal distress, EMP = empathy.182 A.M. Fernández et al.: Psychometric Properties of the IRI in Chile  European Journal of Psychological Assessment   2011; Vol. 27(3):179–185 © 2011 Hogrefe Publishing  scalesbutPT,withwomenattaininghigherscoresthanmenin FS, EC, and PD. Similarly, the estimated 99% confi-dence intervalsforthe α coefficients of the subscalesof theIRI are similar to Davis’ (1980) results.The only exceptionis the internal consistency of PD for the male subsample,which is lower than the results reported by Davis (1980).The correlations of the scales are presented in Table 2.The three samples analyzed (men, women, and total sam-ple) show a similar pattern of correlation among the scales.An exception may be seen in the men subsample, whereEC and PT do not show a significant correlation. Structural Validity Considering Davis’ four-factor model, shown in Figure 1,investigating the internal structure of the IRI for the totalsampleshowed χ ²=781.74(344),  p <.001,RMSEA=0.054,CFI = 0.813, and SRMR = 0.070. RMSEA and SRMR sug-gest a good model fit, whereas CFI suggests caution.Regarding the Cliffordson second-order model shown inFigure 1, the investigation of the internal structure showedthefollowingfitindices: χ ²=810.34(346),  p <.001,RMSEA= 0.056, CFI = 0.802, and SRMR = 0.075, for the total sam-ple. RMSEA and SRMR suggest a good model fit, whereasCFI suggests further research is needed.Overall, each one of the models shows an appropriate de-gree of model fit. The results for the two models are rathersimilaralthoughDavis’smodeldoesabitbetterthanClifford-son’s model (for example according to the CFI results). Predictive Validity Evidence based on the convergent and divergent correlationoftheIRIsubscaleswithotherinstrumentsisshownonTable3. As theoretically expected, measures of positive self-per-ception and affect (RSE and PA) show a significant positiverelationship with PT. All the measures that tap into anxiety-related,negativefeelingsandaggressiveconstructs(STAI-T,BPAQ, SAD, and NA) are significantly and inversely corre-lated with PT. Similarly, constructs that represent negativefeelings,suchasSTAI-T,SAD,BPAQ,andNA,arepositive-ly related to PD; while self-oriented positive traits (RSE andPA)showanegativecorrelationwithPD.Finally,ECshowedasignificantpositiverelationshipwithself-esteemandaneg-ative one with social anxiety. Discussion Overall, the Chilean adaptation of the IRI reached accept-able psychometric properties with a college student sam-ple. The reliability and validity of Davis’ and Cliffordson’sIRI models were studied, showing a high internal consis-tency, temporal stability, and the expected associationsamong the IRI scales, and other instruments that tap intoclosely related psychological constructs. Also, there is agood structural adjustment of the analyzed models.The strength of the validity estimation of the Chileaninstrumentisenhanced byitsrelationshipto localmeasuresof self-esteem, anxiety, aggression, and affectivity.Most of these convergent measures had been adapted to the localcontextindependentofthepresentresearch,givingastrongindication that the conceptualization and definition of theIRIto assessempathyisgood.Nevertheless,itwasnotpos-sible to confirm any concurrent correlation of the FS withany of the validation measures, which may be indicative of a current difficulty to assess imagination and fantasy withtheselected instruments(which wasalsonotedintheSpan-ish adaptation of the IRI by Mestre et al., 2004). Finally,similar to the results of Siu and Shek (2005), the lack of correlation between PT and PD for the male sample may Table 2.  Correlation among the IRI scales, by sex and total sample IRI scales Men ( n  = 201) Women ( n  = 234) Total ( n  = 435)FS EC PT FS EC PT FS EC PTEC .35** .38** .37** ,PT –.10 –.10 –.12 .21** –.09 .20**,PD .29** .28** .13 .17** .33** .22** .24** .34** .21**  Notes.  FS = fantasy scale, PT = perspective taking, EC = empathic concern, PD = personal distress. **  p  < .001. Table 3.  Predictive validity of the IRI dimensions with other instruments RSE ( n  = 420) STAI T ( n  = 322) BPAQ ( n  = 105) SAD ( n  = 102) PA ( n  = 210) NA ( n  = 210)FS .049 .070 .156 –.035 .055 .082EC .128** .051 –.007 –.242** .095 .014PT .168** –.120* –.310** –.309** .268** –.164**PD –.288** .578** .305** .227* –.268** .358**  Notes.  FS = fantasy scale, PT = perspective taking, EC = empathic concern, PD = personal distress, RSE = self-esteem, STAI T = trait anxiety,BPAQ = aggression, SAD = social avoidance and distress, PA = positive affect, NA = negative affect. *  p  < .05, **  p  < .01. A.M. Fernández et al.: Psychometric Properties of the IRI in Chile 183 © 2011 Hogrefe Publishing  European Journal of Psychological Assessment   2011; Vol. 27(3):179–185
Similar documents
View more...
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks

We need your sign to support Project to invent "SMART AND CONTROLLABLE REFLECTIVE BALLOONS" to cover the Sun and Save Our Earth.

More details...

Sign Now!

We are very appreciated for your Prompt Action!