摘要：Authors report a study on psychometric properties of Plutchik's test,called Emotions Profile Index (EPI).A new Slovene translation and adaptation of English version of the test,consisting of combinations (pairs) of 12 words reflecting eight different emotional conditions,was prepared and compared to the old one.Both versions as well as the Big Five Questionnaire (BFQ) were administered on the sample of 239 participants.Different statistical analyses were performed examining psychometric features of both versions of EPI.Discriminative power was tested by cluster analysis and analysis of frequency distributions,reliability was studied via internal consistency index and correlation between the two versions,and validity was examined by correlating PIE dimensions with BFQ dimensions and subdimensions,by comparing profiles of groups on both versions of EPI and BFQ and by fitting the theoretical model proposed by Plutchik to the data.Discriminative power of EPI seems to be affected by avoiding (not choosing) the socially desirable expressions in the test,parallel reliability seems to be susceptible to the use of different words (expressions) in the new version of EPI having the same meaning as words in the old version.Dimensions expected to reflect similar constructs in BFQ and EPI do not correlate satisfactory.Data gathered with EPI cannot be fully explained with the model proposed by Plutchik's theory.
其他摘要：Članek poroča o preverjanju metričnih karakteristik Plutchikovega testa Profil Indeks Emocij (PIE).Na podlagi ponovnega prevoda izvirnih angleških izrazov je bila staremu prevodu vprašalnika izdelana vzporedna verzija.Stara in nova verzija PIE ter BFQ so bili aplicirani na vzorcu 239 posameznikov.Izvedene so bile statistične analize za preverjanje merskih značilnosti obeh verzij.Občutljivost je bila izračunana s pomočjo klastrske analize in frekvenčnih distribucij,zanesljivost s koeficientom notranje zanesljivosti ter vzporedna zanesljivost s korelacijami med verzijama,veljavnost pa s pomočjo korelacij dimenzij PIE in BFQ,s primerjavo profilov skupin iz klastrske analize na PIE in njim pripadajočih rezultatov na BFQ ter s preverjanjem prileganja teoretičnega modela empiričnemu s pomočjo korelacij.Na občutljivost PIE močno vpliva izogibanje socialno nezaželenim izrazom,na vzporedno zanesljivost pa sprememba uporabljenih besed (ne da bi se spremenil njihov pomen).Dimenzije,ki naj bi merile podobne konstrukte na vprašalnikih BFQ in PIE,ne korelirajo ustrezno med seboj.Podatkov,ki jih pridobimo s PIE,ne moremo pojasniti z modelom,ki ga obljublja teorija.Empirični model na nobeni verziji podatkov se ne prilega dobro modelu,ki ga pričakujemo v skladu s teorijo.
