Download - Maree Pedi(2008)

Transcript
  • 7/29/2019 Maree Pedi(2008)

    1/24

    161International Journal of Adolescence and Youth, 2008, Volume 14, pp. 16118402673843/08 $10 2008 A B Academic PublishersPrinted in Great Britain

    A Pedi Translation o the Bar-OnEmotional Quotient Inventory:Youth Version

    Jacobus G. Maree

    Faculty of Education, University of Pretoria, 0001 Pretoria, South Africa

    Jacobus J. Pietersen

    Department of Statistics, Nelson Mandela Metropolitan University, 6000 Port Elizabeth,South Africa

    AbstrAct

    The Bar-On Emotional Quotient Inventory was translated into Pedi, and thensubmitted to an expert in Pedi to check the language beore being administered to30 Grade 9 Pedi rst-language students to identiy potentially vague instructionsand/or items. 800 students in Grades 9 and 11 in schools in the Sekhukhuneland,Nebo and Apel regions o the Limpopo Province o South Arica completed theEnglish and Pedi versions o the inventory in 2004. The mean age in Grade 11was 18.3 yrs (SD = 1.84), and Grade 9 16.2 yrs (SD = 1.64). Whereas unsatisactoryreliabilities o the anticipated actors were obtained, exploratory actor analysis withoblique (direct oblimin) rotation yielded a actor structure that did not correspondsatisactorily with the actor structure yielded by a North American sample. Moreresearch is needed beore the Bar-On Emotional Quotient Inventory: Youth Version(English and Pedi version) can be used in South Arica.

    Numerous variables have been shown to be related to achievementat school and the tertiary level, including motivation, teachersexpectations, cultural background and parental attitudes,and they contribute to the dierence between achievers andnonachievers (Maree and Ebershn, 2002; Maree and Eiselen,2004). This article addresses a vital aspect o the challenge to

    *Please address correspondence to Pro. Jacobus G. Maree, Education Faculty,University o Pretoria, 0001 Pretoria, South Arica or email ([email protected]).

  • 7/29/2019 Maree Pedi(2008)

    2/24

    162

    optimise achievement in lie: the root o problems also, perhapsparticularly, alls outside the cognitive eld. A stable aect shouldsupport cognitive achievement (Maree and Steyn, 2004). Studentseelings, the way in which they experience their signicant others,and their circumstances at home, play a signicant role in theireventual academic achievement and achievement in lie (Maree,Pretorius and Eiselen, 2003).

    Clearly, a number o interrelated actors play a role in success.Intelligence or IQ on its own only predicts 20% o the variancein achievement (Goleman, 1996). Goleman (1996) maintains thatemotional intelligence (including abilities such as sel-motivation,

    persistence in the ace o ailure, impulse control and graticationdelay, mood regulation, empathy, and the ability to hope and toprevent sadness or distress rom interering with ones thinkingprocess) is a more powerul predictor o achievement and successthan IQ. Furthermore, positive eedback and encouragement,others who have high expectations o one and who inspire one,all appear to make contributions towards helping a person toachieve as she or he could achieve (Elo, Maree and Miller,2006).

    Mehrabian (2000) states that the term emotional intelligence isused widely to explain unique dierences linked to lie success,not expressly assessed with conventional intelligence measures.The concept o emotional intelligence is dened as the ability

    to motivate onesel and persist in the ace o rustrations; to controlimpulse and delay gratication; to regulate ones moods and keepdistress rom swamping the ability to think; to empathize and hope(Goleman, 1996, p. 34).

    Kapp (2000) denes emotional intelligence as that actor whichmotivates one to perorm, and to display behaviours such asones will, determination, creativity, impulse control, social skill,empathy, insight and integrity.

    Mayer and Salovey (1993), Goleman (1996) and Bar-On (1996)supplied adequate proo o the positive correlation betweenemotional intelligence and achievement. Bar-On (De Beer,2002) elaborated on these views and conrmed that emotionalintelligence addresses the less cognitive part o intelligence,which is associated with understanding onesel and otherpersons, relating to others, and adapting to and coping withones direct environment. Furthermore emotional intelligencereveals a persons common sense and capacity to relate well tothe world.

    Hui and Triandis (1985) assert that it is essential to establish

  • 7/29/2019 Maree Pedi(2008)

    3/24

    163

    whether a construct has a similar meaning in dierent cultures(conceptual and unctional equivalence). Secondly, the concepthas to be measured by the same instrument to make meaningulcomparisons across cultures. Lastly, the instruments used indierent cultures have to be identical (items should mean the samething to subjects in dierent cultures) (Van der Vijver and Leung,1997; Petrides and Furnham, 2000; Weems and Onwuegbuzie,2001). Otherwise direct comparison o test results is meaningless.Here, we describe the translation o a questionnaire to measurePedi-speaking students emotional intelligence. In a pioneeringproject, the rst o its kind in Southern Arica, the Bar-On

    Emotional Quotient Inventory: Youth Version was administeredto a group o Grade 9 and 11 students in South Arica in boththeir rst language (Pedi) and their second language (English).Since we believe that it is extremely important that translationso requently used instruments should be developed to ensurethe availability o South Arican versions o these instrumentsas well as representative norm groups, our main aim was todetermine the underlying theoretical and empirical dimensions oemotional intelligence as measured by the Pedi translation o theBar-On Emotional Quotient Inventory: Youth Version. Secondly,we intended to question the use o an internationally developedquestionnaire which does not provide or the potential infuenceso the context (South Arican) within which it was going to beused. For this reason, the data were analyzed and attention wasocused on the actor structure o both the translated questionnaireand the original version to investigate the underlying theoreticaland empirical dimensions o emotional intelligence as measuredby the two inventories.

    MEtHOD

    Respondents

    The population was dened as all students in secondary schoolswith an enrolment o at least 400 in Grades 9 and 11 in theSekhukhuneland, Nebo and Apel regions o the Limpopo Province.

    The native language was Pedi. Whereas these adolescents speakPedi as a home language, they are schooled in English. However,code switching (English to Pedi and vice versa) occurs routinely,and these adolescents read and write in both languages.

    Students (N = 800) were chosen rom 10 randomly selectedschools. 80 students per school were selected. Girls and boys in

  • 7/29/2019 Maree Pedi(2008)

    4/24

    164

    each grade were selected randomly and proportionally to obtain atleast ten students or each sex per grade per school. O learners,693 (87%) completed all items o the English and Pedi versions othe Bar-On Emotional Quotient Inventory. O these, 46.6% weremale and 49.5% were in Grade 9. The mean age o the learnerswas 16.8 years (SD = 1.97).

    Assessment instrument

    The Bar-On Emotional Quotient Inventory: Youth Version (Bar-On

    and Parker, 2000a; 2000b).This inventory is a sel-report measure o emotionally andsocially intelligent behaviour, which provides an estimate o onesunderlying emotional and social intelligence. It was developedover a period o 17 years, normed in North America (N= 9,172)and has 60 items. These items are distributed over seven scales(Intrapersonal Relationships, Interpersonal Relationships, StressManagement, Adaptability, General Mood, Positive Impression,and Total EQ). Based on responses rom 9 172 children andadolescents, the inventory has proved to be suitable or use in thecase o American and Canadian children aged seven to 18 years(Bar-On and Parker, 2000b). Validation o the inventory on NorthAmerican samples suggests that the Bar-On Emotional QuotientInventory: Youth Version has excellent psychometric propertiesand identies core eatures o emotional intelligence in children(Bar-On and Parker, 2000b). The actor structure o the instrumentwas examined using exploratory actor analysis. Four empiricalactors (Intrapersonal, Interpersonal, Stress Management andAdaptability) emerged rom the principal components analysis(varimax rotation) (Bar-On and Parker, 2000b). Reliabilitycoecients or the dierent elds o the questionnaire rangedrom .65 to .90. Sample items rom each o the seven scales areprovided in Table 1.

    The our primary actors (Intrapersonal, Interpersonal, StressManagement and Adaptability, comprising 40 items, whichjointly measure Total EQ) as well as general Mood (14 items,which does not orm part o the total EQ score) were used in

    this investigation. Based on the recommendation o the developero the test (Bar-On), Positive Impression (six items) was notconsidered. Items are anchored by a ve-point scale with atextual response ormat ranging rom 1 = Very seldom or Nottrue o me to 5 = Very oten true o me or True o me. A listo the inventorys items is ound in the instruments technical

  • 7/29/2019 Maree Pedi(2008)

    5/24

    165

    manual (Bar-On and Parker, 2000b). The subjects responsesrender a total emotional intelligence score and seven compositescale scores. The inventory includes a scale that assesses itemresponse consistency (designed to identiy random responses).

    Procedure

    Permission was requested and obtained in writing rom the

    education departments as well as rom the sample o randomlyselected schools in 2003 to conduct the research and publish thendings. Permission was also obtained rom parents or theirchildren to participate in the study. Respondents were given theoption o not participating in the study, and all signed inormedassent orms, while all parents/guardians signed inormedconsent orms. Schools were contacted in mid-February 2004. All800 students completed the inventory during a period o threedays in March 2004. Testing or research purposes was explainedto headmasters and parents. Prior to the administration o theinventory, testees were allowed to ask questions about theitems. Test administrators (id est, Pro. Maree, a psychologistand trained EQ administrator and Dr. Molepo), assured testees

    that their responses would be treated with condentiality andthat their anonymity would be saeguarded at all times. The testadministrators explained to the respondents how to complete theinventory. Completion o the test took about 35 minutes. In viewo the act that the inventory may have aroused questions andanxieties, testees were encouraged to make an appointment with

    TABLE 1

    Sample item rom each actor

    Factor Sample item rom each area

    1. Intrapersonal It is easy to tell people how I eel2. Interpersonal I care what happens to other

    people3. Stress management I can stay calm when I am upset4. Adaptability It is easy or me to understand

    new things5. General mood I am happy6. Positive impression I like everyone I meet

  • 7/29/2019 Maree Pedi(2008)

    6/24

    166

    the test administrators to discuss their eelings and the questions.Testees were assured that general eedback on the results wouldbe provided within two months. Feedback was subsequentlyprovided approximately ve weeks ater the initial assessment.The English version o the questionnaire was administered rst.We waited until all students had completed the English versiono the test, ater which students were given a 15-minute breakbeore the Pedi version was administered.

    The translated version o the inventory was submitted to anexpert in Pedi or assessment. In the assessment o the items,attention was given to clarity, uniqueness, unambiguousness,

    the use o words with the exact meaning and equivalence othe English statements. The inventory was adjusted urtheron the basis o the comments rom the expert. The translatorwas specically requested to take care that the adaptation didnot change the purpose o the questions asked. The translatedinventory was then back-translated, ater which all items werechecked again by the translator and our Pedi expert. It wassubsequently administered to 30 Grade 9 Pedi rst-languagestudents to identiy potentially vague instructions and/oritems. Testees were requested to circle the numbers o, and tounderline phrases and words contained in the items they did notunderstand. On the basis o the testees reactions with respectto the items, the ormulation o a number o items was urtheramended.

    LIMItAtIONs OF tHE stUDY

    This was a limited local study and the ndings have limitedgeneralization. Furthermore, we realize that since the Englishand Pedi versions o the instruments were not administered in asplit hal manner, order eects may be present here.

    rEsULts

    Statistical analysis

    The STATISTICA data analysis sotware system version 7.1(StatSot, 2006) and the CEFA: Comprehensive exploratory factoranalysis (Browne, Cudeck, Tateneni and Mels, 1998) was used inanalyzing the data.

  • 7/29/2019 Maree Pedi(2008)

    7/24

    167

    Reliabilities and confrmatory analysis o theoreticaldimensions

    We rst conducted conrmatory actor analysis (Statsot, 2006). Theproposed ve-actor structure o the Bar-On Emotional QuotientInventory (Interpersonal, Intrapersonal, Stress Management,Adaptability, and General Mood) was tested on the data romboth the English and Pedi versions o the questionnaire. Thiswas done by calculating reliabilities or the actors (measured byCronbachs Alpha) (Hair, Black, Babin, Anderson and Tatham,2006) and by perorming conrmatory actor analysis to assess

    how well the model tted the data. The results appear inTable 2. In the case o the English data, the reliabilities o ouro the actors did not reach the 0.70 threshold while two didnot reach the 0.60 threshold, and in the case o the Pedi datathree did not reach the 0.70 threshold and two did not reachthe 0.60 threshold. In both languages the General Mood actorhad the highest reliability. The results o the conrmatory actoranalysis show that although almost all loadings were statisticallysignicant (mainly due to the large sample) they were generallynot very high. Large portions o the variance o the items couldnot be explained by the actors as was evident rom the sizableunique variances (English: 32 are > 0.80; Pedi: 26 are > 0.80).Correlations between the ve actors appear in Table 3, whereasgoodness-o-t indices appear in Table 4. Correlations were, ingeneral, statistically signicant. The Root Mean Square Error oApproximation (RMSEA) is at an acceptable level to indicate aclose t (p .05) but the Tucker Lewis Index or Non-normed FitIndex (NNFI) and Comparative Fit Index (CFI) indicate that thet is not good (both are substantially smaller than 0.90) (Kenny,2003).

    Exploratory actor analysis

    Since the inventory was implemented or the rst time, exploratoryactor analysis was conducted because the anticipated structuredid not t our data and we were interested in nding out which

    structure would actually emerge rom the data. Twelve o the60 items in the inventory were negatively phrased (items 6, 15,21, 26, 28, 35, 37, 46, 49, 53, 54, and 58). The scores or theseitems were inverted (denoted*) prior to conducting any statisticalanalyses (Bar-On, personal communication, 2006; Schepers, 2004).Secondly, we omitted the six Positive Impression scale items (8,

  • 7/29/2019 Maree Pedi(2008)

    8/24

    168

    TABLE 2

    Conrmatory actor analysis: actor loadings, unique variancesand Cronbach coecients alpha

    ENGLISH PEDI

    Factor Item Factor Unique a Factor Unique ano. loading variance loading variance

    Intrapersonal 7 0.45 0.80 0.30 0.56 0.68 0.44

    (Intra) 17 0.53 0.72 0.59 0.6528* 0.25 0.94 0.22 0.9531 0.56 0.68 0.60 0.6443 0.55 0.70 0.72 0.4953* 0.27 0.93 0.22 0.95

    Interpersonal 2 0.42 0.82 0.64 0.36 0.87 0.69(Inter) 5 0.32 0.90 0.38 0.86

    10 0.04 .00 0.22 0.9514 0.60 0.64 0.47 0.7820 0.48 0.77 0.48 0.7724 0.6 0.98 0.36 0.8736 0.3 0.90 0.37 0.8641 0.45 0.80 0.54 0.745 0.23 0.95 0.36 0.8751 0.62 0.62 0.49 0.7655 0.32 0.90 0.33 0.8959 0.0 0.99 0.3 0.9

    Stress 3 0.9 0.97 0.37 0.09 0.99 0.58management 6* 0.26 0.93 0.25 0.94(SM) 11 0.2 0.96 0.9 0.96

    15* 0.24 0.94 0.29 0.9221* 0.45 0.80 0.39 0.8526* 0.35 0.88 0.43 0.835* 0.37 0.86 0.53 0.7239 0.22 0.95 0.05 .00

    46* 0.28 0.92 0.49 0.7649* 0.27 0.93 0.26 0.9354* 0.44 0.8 0.57 0.6858* 0.3 0.9 0.48 0.77

  • 7/29/2019 Maree Pedi(2008)

    9/24

    169

    Table 2 continued

    ENGLISH PEDI

    Factor Item Factor Unique a Factor Unique ano. loading variance loading variance

    Adaptability 12 0.39 0.85 0.66 0.52 0.73 0.73(Ad) 16 0.38 0.86 0.4 0.84

    22 0.37 0.86 0.50 0.7525 0.39 0.85 0.34 0.88

    30 0.48 0.77 0.57 0.6734 0.38 0.85 0.49 0.7638 0.42 0.83 0.52 0.7344 0.49 0.76 0.47 0.7848 0.39 0.85 0.48 0.7757 0.33 0.89 0.37 0.86

    General 1 0.33 0.89 0.83 0.4 0.83 0.77mood 4 0.56 0.68 0.50 0.75(GM) 9 0.54 0.7 0.54 0.7

    13 0.30 0.9 0.39 0.8519 0.48 0.77 0.42 0.8323 0.55 0.70 0.3 0.929 0.5 0.74 0.46 0.7832 0.47 0.78 0.45 0.8037* 0.3 0.9 0.3 0.9840 0.67 0.56 0.60 0.6447 0.56 0.68 0.50 0.7550 0.6 0.63 0.50 0.7656 0.67 0.56 0.55 0.7060 0.67 0.56 0.54 0.7

    Criterion or minimal actor loadings: loading < .25* Inverted itemsValues in bold: p .05

    18, 27, 33, 42 and 52), since these items are indicators o aking

    good that were articially tacked on to the other items andhave nothing at all to do with Pro. Bar-Ons conceptualizationo the EI construct.

    Acting on the advice o the developer o the questionnaire,Pro. Reuven Bar-On, we rst limited our actor extraction downto ve. Table 5 shows how much variance is explained by the

  • 7/29/2019 Maree Pedi(2008)

    10/24

    170

    TABLE 3

    Intercorrelations (r) o actors or English and Pedi versions:conrmatory actor analysis

    ENGLISH PEDI

    Factor Intra Inter SM Ad GM Intra Inter SM Ad GM

    Intra Inter 0.59 0.56 SM 0.26 0.2 0.02 0.

    Ad 0.7 0.73 0.03 0.54 0.57 0.06 GM 0.49 0.94 0.34 0.68 0.44 0.8 0.28 0.53

    Values in bold: p .05

    TABLE 4

    Goodnessot indices

    NNFI CFI RMSEA

    English 0.775 0.785 0.041Pedi 0.742 0.754 0.043

    actors. Tables 6 and 7 contain the actor loadings and Table 8contains the intercorrelations between the ve actors in Englishand Pedi respectively (Statistica, 2006). The scree plots (Figures1 and 2) did not clearly indicate how many actors existed.They rather suggested a range o possibilities, namely three tove actors. However, the eigenvalues at that point were stilllarger than one and the percentage variance explained less than30%. An oblique rotation, direct oblimin, was used to improveinterpretation (Browne, et al., 1998; Hair et al., 2006). Therationale behind using an oblique rotation and not an orthogonalrotation was the ollowing: Since one expects the dierent actors

    measured by the instrument sub-domains o the overall EQ to be related, allowance should be made or this to eaturein the results. An orthogonal rotation orces the actors to beuncorrelated, and this was undesirable in the current case, sincea total EQ score is, ater all, calculated by adding all o the rst40 items o the questionnaire.

  • 7/29/2019 Maree Pedi(2008)

    11/24

    171

    TABLE 5

    Table o variance explained (rst 20 eigenvalues)

    ENGLISH PEDI

    Eigenvalue % Variance Cum. % Eigenvalue % Variance Cum. %explained variance explained variance

    explained explained

    1 7.96 14.74 14.74 7.40 13.70 13.702 3.20 5.92 20.66 3.22 5.96 19.66

    3 1.80 3.34 24.00 2.19 4.05 23.714 1.57 2.90 26.90 1.70 3.15 26.865 1.41 2.61 29.52 1.58 2.93 29.796 1.31 2.43 31.95 1.42 2.64 32.427 1.30 2.40 34.35 1.33 2.46 34.888 1.26 2.33 36.67 1.29 2.40 37.289 1.25 2.31 38.98 1.24 2.29 39.57

    10 1.19 2.21 41.19 1.15 2.14 41.7111 1.14 2.11 43.30 1.14 2.11 43.8212 1.12 2.08 45.38 1.10 2.04 45.8613 1.07 1.99 47.37 1.05 1.94 47.8014 1.05 1.94 49.30 1.04 1.93 49.7315 1.01 1.87 51.18 1.03 1.91 51.6416 0.99 1.82 53.00 1.00 1.84 53.4817 0.97 1.80 54.80 0.97 1.80 55.2918 0.95 1.77 56.56 0.96 1.78 57.0719 0.92 1.71 58.27 0.94 1.74 58.8220 0.92 1.70 59.97 0.91 1.69 60.51

    Once again, on the advice o Pro. Bar-On (2006), we subse-quently limited our actor extraction (oblique rotation, directoblimin) to our actors (ater eliminating the 14 General Mooditems, viz. 1, 4, 9, 13, 19, 23, 29, 32, 37, 40, 47, 50, 56 and 60). Wedid this to try and see

    what makes the best conceptual sense [this should] hopeullygive you the best theoretical t o a our-actor concept o EI orchildren (Intrapersonal, Interpersonal, Stress Management, and

    Adaptability). General mood does not really orm a part o EQ;it is a acilitator o emotional intelligence (id est, it acilitates theother actorial components o emotional intelligence) (Bar-On, 2006,personal communication).

    Table 9 shows how much variance is explained by the actors.Tables 10 and 11 contain the actor loadings and Table 12 the

  • 7/29/2019 Maree Pedi(2008)

    12/24

    172

    TABLE 6

    Factor loadings (direct quartimin rotation): ve actors: English

    FACTOR

    Factor Item 1 2 3 4 5no.

    Intrapersonal 7 0.21 0.07 0.37 0.05 0.0217 0.11 0.20 0.34 0.04 0.05

    28* 0.00 0.17 0.08 0.19 0.0531 0.21 0.18 0.24 0.01 0.1443 0.02 0.24 0.40 0.03 0.1153* 0.02 0.09 0.24 0.19 0.04

    Interpersonal 2 0.30 0.17 0.10 0.10 0.065 0.19 0.03 0.05 0.01 0.22

    10 0.08 0.04 0.23 0.10 0.114 0.54 0.08 0.04 0.04 0.0420 0.47 0.03 0.07 0.01 0.0624 0.00 0.11 0.05 0.02 0.4936 0.25 0.10 0.11 0.05 0.0241 0.37 0.03 0.21 0.05 0.0445 0.10 0.07 0.02 0.08 0.451 0.66 0.09 0.05 0.03 0.0155 0.19 0.06 0.05 0.14 0.2259 0.07 0.08 0.06 0.14 0.27

    Stress 3 0.11 0.09 0.16 0.00 0.34management 6* 0.02 0.07 0.09 0.36 0.12

    11 0.04 0.02 0.09 0.08 0.3015* 0.09 0.05 0.12 0.27 0.0721* 0.36 0.01 0.30 0.22 0.1126* 0.09 0.05 0.08 0.23 0.1535* 0.01 0.06 0.03 0.47 0.0139 0.06 0.05 0.12 0.07 0.2546* 0.09 0.03 0.20 0.18 0.11

    49* 0.12 0.07 0.03 0.18 0.2354* 0.08 0.05 0.11 0.30 0.1558* 0.08 0.02 0.07 0.49 0.04

  • 7/29/2019 Maree Pedi(2008)

    13/24

    173

    Table 6 continued

    FACTOR

    Factor Item 1 2 3 4 5no.

    Adaptability 12 0.08 0.37 0.12 0.06 0.07#16 0.12 0.37 0.06 0.01 0.0422 0.01 0.44 0.09 0.09 0.0625 0.08 0.13 0.08 0.08 0.3

    30 0.08 0.48 0.13 0.00 0.0834 0.10 0.44 0.00 0.03 0.1038 0.02 0.37 0.07 0.05 0.1444 0.14 0.25 0.00 0.10 0.2648 0.07 0.34 0.15 0.04 0.0557 0.02 0.14 0.13 0.03 0.49

    General mood 1 0.36 0.09 0.09 0.10 0.124 0.62 0.08 0.03 0.14 0.139 0.46 0.15 0.20 0.04 0.07

    13 0.16 0.07 0.02 0.04 0.2419 0.33 0.08 0.01 0.05 0.2623 0.52 0.09 0.02 0.00 0.0329 0.29 0.24 0.12 0.08 0.2032 0.35 0.15 0.14 0.05 0.1037* 0.27 0.02 0.05 0.27 0.0440 0.7 0.04 0.02 0.03 0.0647 0.54 0.05 0.08 0.01 0.1350 0.55 0.04 0.01 0.00 0.0956 0.67 0.03 0.05 0.04 0.0460 0.62 0.03 0.00 0.05 0.03

    Loadings o 0.30 and higher are in bold

    * inverted items

    intercorrelations between the our actors in English and Pedirespectively. Once again the scree plots (Figures 3 and 4) suggest

    a possible range o actors and not a single number.It is clear rom Tables 6, 7, 10 and 11 that a large number

    o items do not load higher than 0.3 on any given actor. Theactor structures obtained in the current study do not correspondsatisactorily with the original actor structure, based on ananalysis o data obtained rom a North American sample. Some

  • 7/29/2019 Maree Pedi(2008)

    14/24

    174

    Plot of Eigenvalues

    Number of Eigenvalues

    0

    1

    2

    3

    4

    5

    6

    7

    8

    9

    Value

    Plot of Eigenvalues

    Number of Eigenvalues

    0

    1

    2

    3

    4

    5

    6

    7

    8

    9

    Value

    Figure 2. Scree plot (rst 20 eigenvalues) Pedi: direct quartimin rotation): veactors:

    Figure 1. Scree plot (rst 20 eigenvalues) English: (direct quartimin rotation):ve actors

  • 7/29/2019 Maree Pedi(2008)

    15/24

    175

    Plot of Eigenvalues

    Number of Eigenvalues

    0.0

    0.5

    1.0

    1.5

    2.0

    2.5

    3.0

    3.5

    4.0

    4.5

    5.0

    5.5

    6.0

    Value

    Figure 4. Scree plot (rst 20 eigenvalues): our actors (without general moodactor): Pedi

    Figure 3. Scree plot (rst 20 eigenvalues): our actors (without general moodactor): English

    Plot of Eigenvalues

    Number of Eigenvalues

    0.0

    0.5

    1.0

    1.5

    2.0

    2.5

    3.0

    3.5

    4.0

    4.5

    5.0

    5.5

    Va

    lue

  • 7/29/2019 Maree Pedi(2008)

    16/24

    176

    TABLE 7

    Factor loadings (direct quartimin rotation): Pedi

    FACTOR

    Factor Item 1 2 3 4 5no.

    1 7 0.06 0.08 0.09 0.6 0.0617 0.01 0.03 0.02 0.6 0.04

    28* 0.06 0.04 0.23 0.13 0.0131 0.10 0.09 0.05 0.50 0.0143 0.01 0.09 0.00 0.64 0.0453* 0.06 0.10 0.30 0.19 0.04

    2 2 0.09 0.04 0.01 0.17 0.35 0.01 0.07 0.03 0.14 0.42

    10 0.15 0.03 0.03 0.13 0.4514 0.49 0.03 0.02 0.04 0.0520 0.43 0.05 0.02 0.05 0.0924 0.26 0.04 0.03 0.06 0.1436 0.11 0.01 0.03 0.11 0.3341 0.34 0.03 0.01 0.19 0.1945 0.24 0.06 0.05 0.02 0.2151 0.57 0.11 0.02 0.01 0.0455 0.19 0.14 0.13 0.02 0.0959 0.04 0.14 0.16 0.03 0.3

    3 3 0.16 0.14 0.01 0.01 0.156* 0.08 0.03 0.22 0.15 0.02

    11 0.39 0.13 0.03 0.00 0.0115* 0.02 0.03 0.30 0.09 0.1721* 0.35 0.01 0.28 0.12 0.0726* 0.23 0.01 0.35 0.06 0.0435* 0.12 0.06 0.59 0.05 0.0239 0.04 0.09 0.08 0.05 0.0946* 0.01 0.00 0.52 0.01 0.01

    49* 0.01 0.01 0.30 0.08 0.0354* 0.00 0.00 0.57 0.02 0.0258* 0.07 0.00 0.45 0.12 0.14

  • 7/29/2019 Maree Pedi(2008)

    17/24

    177

    Table 7 continued

    FACTOR

    Factor Item 1 2 3 4 5no.

    4 12 0.03 0.52 0.06 0.05 0.0616 0.15 0.23 0.01 0.14 0.0522 0.10 0.52 0.00 0.04 0.0525 0.01 0.24 0.03 0.02 0.22

    30 0.05 0.55 0.01 0.04 0.0734 0.07 0.5 0.03 0.01 0.0838 0.15 0.47 0.05 0.00 0.0244 0.10 0.42 0.09 0.02 0.0348 0.07 0.44 0.06 0.09 0.0757 0.14 0.27 0.08 0.11 0.27

    5 1 0.37 0.09 0.06 0.18 0.014 0.52 0.02 0.01 0.04 0.049 0.48 0.02 0.02 0.07 0.03

    13 0.27 0.11 0.02 0.05 0.1219 0.30 0.11 0.04 0.05 0.2123 0.21 0.01 0.03 0.10 0.1529 0.32 0.16 0.08 0.10 0.2832 0.33 0.01 0.00 0.07 0.2137* 0.08 0.03 0.34 0.12 0.1440 0.57 0.00 0.03 0.00 0.0947 0.47 0.07 0.05 0.07 0.0950 0.46 0.15 0.03 0.02 0.0356 0.60 0.07 0.02 0.02 0.1860 0.5 0.04 0.02 0.07 0.02

    Loadings o 0.30 and higher are in bold* inverted items

    actors showed airly strong intercorrelations (larger than 0.25)(Tables 8 and 11), serving as justication or using oblique rather

    than orthogonal rotation.Some o the proposed actors were realized to some extent

    rom the exploratory analyses. In the ve-actor solution (English)the Adaptability and General Mood actors seemed to be presentwhile in the Pedi data the Intrapersonal, Stress Management,Adaptability and General Mood actors were present to some

  • 7/29/2019 Maree Pedi(2008)

    18/24

    178

    TABLE 8

    Intercorrelations (r) o actors or English and Pedi versions:ve actors

    ENGLISH PEDI

    Factor 1 2 3 4 5 1 2 3 4 5

    1 2 0.48 0.32 3 0.03 0.19 0.14 0.00

    4 0.24 0.03 0.17 0.28 0.39 0.03 5 0.27 0.31 0.14 0.02 0.22 0.30 0.04 0.27

    TABLE 9

    Variance explained (rst 20 eigenvalues)

    ENGLISH PEDI

    Eigen- % Variance Cum. % Eigen- % Variance Cum. %value explained variance value explained variance

    explained explained

    1 4.74 11.85 11.85 5.27 13.17 13.172 2.61 6.53 18.39 2.78 6.96 20.123 1.71 4.29 22.67 1.86 4.65 24.784 1.37 3.43 26.11 1.63 4.07 28.855 1.33 3.32 29.43 1.43 3.56 32.416 1.24 3.11 32.54 1.30 3.25 35.677 1.23 3.06 35.60 1.19 2.97 38.638 1.20 3.00 38.60 1.17 2.92 41.569 1.19 2.99 41.59 1.11 2.77 44.33

    10 1.12 2.80 44.39 1.07 2.69 47.0111 1.08 2.71 47.10 1.03 2.56 49.5712 1.08 2.69 49.79 0.99 2.49 52.0613 1.01 2.52 52.31 0.98 2.44 54.50

    14 0.97 2.43 54.74 0.96 2.41 56.9115 0.94 2.36 57.10 0.95 2.36 59.2716 0.92 2.30 59.40 0.91 2.27 61.5417 0.89 2.22 61.63 0.89 2.23 63.7718 0.86 2.16 63.78 0.86 2.16 65.9319 0.85 2.14 65.92 0.85 2.13 68.0620 0.83 2.07 67.99 0.83 2.07 70.12

  • 7/29/2019 Maree Pedi(2008)

    19/24

    179

    TABLE 10

    Factor loadings (direct quartimin rotation): our actors: English

    FACTOR

    Factor Item no. 1 2 3 4

    1 7 0.42 0.02 0.02 0.0817 0.44 0.06 0.04 0.0228* 0.19 0.00 0.06 0.1931 0.35 0.20 0.10 0.0543 0.57 0.13 0.11 0.08

    53* 0.26 0.09 0.07 0.15

    2 2 0.27 0.24 0.04 0.095 0.08 0.21 0.21 0.01

    10 0.21 0.19 0.18 0.0614 0.10 0.55 0.04 0.0320 0.06 0.44 0.04 0.0424 0.05 0.00 0.49 0.0436 0.18 0.21 0.03 0.0741 0.20 0.37 0.01 0.0145 0.05 0.12 0.44 0.0651 0.04 0.60 0.01 0.0455 0.09 0.23 0.21 0.1659 0.07 0.02 0.27 0.13

    3 3 0.06 0.18 0.34 0.04

    6* 0.02 0.02 0.13 0.3311 0.10 0.09 0.27 0.0615* 0.17 0.07 0.11 0.2021* 0.18 0.50 0.07 0.1926* 0.00 0.14 0.16 0.2335* 0.09 0.06 0.02 0.4639 0.15 0.11 0.26 0.0446* 0.16 0.20 0.08 0.1749* 0.03 0.08 0.24 0.1754* 0.07 0.21 0.16 0.2558* 0.05 0.04 0.05 0.5

    412 0.19 0.24 0.03 0.0416 0.29 0.16 0.07 0.0122 0.39 0.08 0.08 0.1025 0.16 0.15 0.29 0.06

    30 0.45 0.17 0.11 0.0434 0.29 0.07 0.07 0.0538 0.30 0.09 0.11 0.0144 0.16 0.29 0.25 0.0548 0.36 0.13 0.05 0.0057 0.00 0.11 0.46 0.02

    Loadings o 0.30 and higher are in bold* inverted items

  • 7/29/2019 Maree Pedi(2008)

    20/24

    180

    TABLE 11

    Factor loadings (direct quartimin rotation): our actors: Pedi

    FACTOR

    Factor Item no. 1 2 3 4

    1 7 0.08 0.09 0.03 0.6417 0.01 0.01 0.01 0.5828* 0.02 0.22 0.06 0.1531 0.09 0.04 0.09 0.4843 0.05 0.01 0.01 0.68

    53* 0.08 0.29 0.10 0.17

    2 2 0.16 0.06 0.13 0.225 0.20 0.05 0.15 0.20

    10 0.16 0.12 0.06 0.2214 0.01 0.04 0.5 0.0320 0.03 0.03 0.54 0.0524 0.10 0.02 0.32 0.0736 0.11 0.09 0.21 0.1841 0.03 0.02 0.42 0.2245 0.14 0.06 0.27 0.0351 0.06 0.01 0.56 0.0055 0.20 0.13 0.20 0.0259 0.25 0.20 0.12 0.07

    3 3 0.20 0.01 0.17 0.04

    6* 0.00 0.23 0.05 0.1611 0.16 0.08 0.35 0.0115* 0.06 0.32 0.07 0.1121* 0.06 0.29 0.36 0.1126* 0.02 0.37 0.23 0.0935* 0.06 0.57 0.08 0.0539 0.11 0.09 0.11 0.0646* 0.00 0.5 0.02 0.0349* 0.02 0.29 0.00 0.0854* 0.02 0.55 0.01 0.0358* 0.03 0.48 0.05 0.08

    4 12 0.50 0.08 0.05 0.0116 0.25 0.03 0.09 0.1622 0.56 0.00 0.13 0.0325 0.29 0.06 0.08 0.06

    30 0.59 0.01 0.06 0.0334 0.49 0.00 0.04 0.0638 0.47 0.07 0.11 0.0144 0.43 0.09 0.11 0.0148 0.45 0.06 0.09 0.1157 0.34 0.11 0.20 0.04

    Loadings o 0.30 and higher are in bold; * inverted items

  • 7/29/2019 Maree Pedi(2008)

    21/24

    181

    extent. In the our-actor solution (English) the Intrapersonal and,to a lesser extent, the Adaptability actors seemed to be present,while in the Pedi data the Intrapersonal and Adaptability actorsseemed to be present and to a lesser extent the Interpersonal andStress Management actors.

    DIscUssION

    Exploratory actor analysis o the 54 items o the Bar-OnEmotional Quotient Inventory: Youth Version did not conrmthe actor structure that emerged when the data obtained romthe North American sample were analysed. The actor structuresidentied in this study were thereore deemed to have limiteduseulness, since no single concept or term encapsulated theessence o most o the actors that could indeed be identied,irrespective o the way one looked at the actor analyses.

    Items that were negatively phrased no doubt contributed tothe problematic situation. They may, in act, have distorted theresults o the actor analysis to a signicant degree (personalinterviews with a number o learners revealed that they oundthese items particularly conusing) (See: Tables 6, 7, 10 and 11).

    Generally speaking, they did not load satisactorily on any o

    the ve actors in the exploratory actor analysis, despite the actthat they were inverted prior to actor analysis. Respondents,negatively aecting the empirical actor structure, probablypoorly grasped negatively phrased items. This is not particularlyuncommon, since research has shown that reverse scoring orwording o items does not translate clearly across cultures

    TABLE 12

    Intercorrelations (r) o actors or English and Pedi versions:our actors

    ENGLISH PEDI

    Factor 1 2 3 4 1 2 3 4

    1 2 0.31 0.03

    3 0.33 0.19 0.27 0.10 4 0.16 0.18 0.10 0.45 0.03 0.28

  • 7/29/2019 Maree Pedi(2008)

    22/24

    182

    (Weems and Onwuegbuzie, 2001). For this reason inverteditems need special attention in uture translations, especiallysince reverse-scored items may have a dierent meaning incomparison with items that are phrased in a positive direction.In addition questionnaires that contain both positively andnegatively keyed items tend to have reduced reliabilities (Weemsand Onwuegbuzie, 2001).

    Clearly it is essential to revise the translation o the Bar-OnEmotional Quotient Inventory: Youth Version beore it can beapplied to Pedi-speaking students in South Arica. Furthermore,some o the realities o the region in which this type o study is

    done, need to be incorporated in uture studies. Assuming thatthe construct o EI (Emotional Intelligence) does not have dierentmeanings in the two cultures under discussion (i.e. North Americaand the rural Pedi-speaking adolescent population o SouthArica), it nonetheless seems plausible to revisit the translatedPedi version, rather than attempting to adapt an English versionor use with Pedi-speaking students. Ater all, the characteristicsthat we reer to as part o the concept Emotional Intelligenceinclude a great many that are only clearly expressible in thenative language, or example eelings, experience o signicantothers, circumstances at home, sel-motivation, impulsecontrol, mood regulation, empathy, hope, prevention osadness, should be done primarily in the mother tongue andnot in a secondary instructional language. Furthermore whereasemotional aspects o any specic culture would be learnedlater in a second or third language, a clients mother tonguewould be the primary instrument or articulating thinking aboutemotionally loaded topics.

    Care needs to be taken not to change the original purpose othe questionnaire during translation. The role o social desirabilitylevels and motivation to respond (Hui and Triandis, 1985) needsto be investigated during re-administration, especially since thenotion o EI is not generally accepted in schools in South Arica,and this actor may have infuenced results negatively.

    From the aoresaid it is sel-evident that the possible infuenceso local contexts on questionnaires developed in other parts o theworld need to be investigated extremely careully beore these

    questionnaires are used in local contexts. We believe that ourresearch emphasizes the need to check, through research, manyo our generally held assumptions with regard to respondentsand their environment.

    It may be possible that the EI concept does not relate to thePedi culture as a unitary concept. Instead, it may be broken up

  • 7/29/2019 Maree Pedi(2008)

    23/24

    183

    into a number o subdimensions, which are all important withinthe Pedi culture and belies. It is thereore suggested that theollowing two issues be investigated thoroughly during utureresearch: a. the degree to which the results contained in thisresearch are an objet dart o the actor-analytic procedures carriedout, and b. we will need to conduct in-depth interviews with anumber o respondents (participants) to investigate the thoughtsexpressed here (id est, whether the ideas contained in the EIquestionnaire tend to be separated by cultural belies).

    It should be clear that the South Arican goal o airness intesting across language groups may only be attained once the

    need or a multistrategy approach is addressed. It is suggestedthat test batteries should include the use o both qualitative andquantitative research techniques, ocusing on inter-individualassessment as well as intra-individual scrutiny: Test results shouldat all times be interpreted with extreme caution, even whenquality assurance criteria such as reliability and validity have beenexplained satisactorily in manuals o psychological tests, and bevalidated by clients (Savickas, 2006). Ideally speaking testingshould be conducted in the language choice o clients. Wordsand expressions should be created in order to be able to expressimportant psychological concepts in it. Ater all, eective testingcan only be brought about when the communication betweenpsychologist and client is adequate.

    AcKNOWLEDGEMENts

    1. We would like to express our sincere thanks to those clientsand teachers whose goodwill and kind cooperation enabled usto conduct this investigation.

    2. We also wish to thank Pro. Reuven Bar-On (University oTexas: Medical Branch) or oering his views and invaluableadvice:

    3. This article is based on the results o research unded by Jopievan Rooyen and Partners. We grateully thank the Companyor its invaluable support.

    rEFErENcEs

    Bar-On, R. (1996). A brief description of Reuven Bar-Ons EQ inventory. Multi-Health Systems; Toronto.

    Bar-On, R. & Parker, D.A. (2000a). Bar-On emotional quotient inventory: YouthVersion. Multi-Health Systems; New York.

  • 7/29/2019 Maree Pedi(2008)

    24/24

    184

    Bar-On, R. & Parker, D.A. (2000b). Bar-On emotional quotient inventory: YouthVersion. Technical manual. New York: Multi-Health Systems; New York.

    Browne, M.W., Cudeck, R., Tateneni, K. & Mels G. (1998). CEFA: Comprehensiveexploratory factor analysis [WWW document and computer program]. AvailableURL: http://quantrm2.psy.ohio-state.edu/browne/). Accessed 1115 March2007.

    De Beer, J. (2002). Training manual: Bar-On Emotional Quotient Inventory. Jopievan Rooyen & Partners; Randburg.

    Elo, I., Maree, J.G. & Miller, L.H. (2006). The role o parents learning acilitationmode in supporting inormal learning in mathematics, Early Child Developmentand Care, 76 (3&4), 313328.

    Goleman, D. (1996). Emotional intelligence: Why it can matter more than IQ.Bantam Books; New York.

    Hair, J.F., Black, W.C., Babin, B.J., Anderson, R.E. & Tatham, R.L. (2006).Multivariate data analysis, 6th edition. Prentice Hall; Upper Saddle River.

    Hui, C.H. & Triandis, H.C. (1985). Measurement in cross-cultural psychology:a review and comparison o strategies. Journal of Cross-cultural Psychology,6, 131152.

    Kapp, C.A. (2000). Emotional intelligence (EQ) and success in post-graduatestudies: A pilot study. SA Journal of Higher Education, 4, 151160.

    Kenny, D.A. (2003). Measuring Model Fit. Available URL: http://davidakenny.net/cm/t.htm. Accessed: 20070410.

    Maree, J.G., & Ebershn, L. (2002). Emotional intelligence and achievement:Redening gitedness? Gifted Education International, 6(3), 261273.

    Maree, J.G., & Eiselen, R. (2004). The emotional intelligence prole o academicsin a merger setting. Education and Urban Society, 37(3), 482504.

    Maree, J.G., Pretorius, A. & Eiselen, R.J. (2003). Predicting success among rst-year engineering students at the Rand Arikaans University. PsychologicalReports, 93, 399409.

    Maree, J.G. & Steyn, T.M. (2004). A study orientation questionnaire in mathematicsor use in a tertiary environment. Psychological Reports, 95, 981987.

    Mayer, J.D., & Salovey, P. (1993). The intelligence o emotional intelligence.Intelligence, 7, 433442.

    Mehrabian, A. (2000). Beyond IQ: Broad-based measurement o individualsuccess potential or emotional intelligence. Genetic, Social, and GeneralPsychology Monographs, 26, 133239.

    Petrides, K.V. & Furnham, A. (2000). On the dimensional structure o emotionalintelligence. Personality and Individual Differences, 29, 313320.

    Schepers, J.M. (2004). Overcoming the eects o dierential skewness o testitems in scale construction. SA Journal of Industrial Psychology, 30(4), 2743.

    StatSot, Inc. (2006). STATISTICA (data analysis software system), version 7.1 .Available URL: www.statsot.com. Accessed: 1217 March 2007.

    Van der Vijver, F.J.R. & Leung, K. (1997). Methods and data-analysis for cross-cultural research. Sage; Newbury Park.

    Weems, G.H. & Onwuegbuzie, A.J. (2001). The impact o midpoint responsesand reverse coding on survey data. Measurement and Evaluation in Counsellingand Development, 34, 166176.

    http://davidakenny.net/cm/fit.htmhttp://davidakenny.net/cm/fit.htmhttp://www.statsoft.com/http://www.statsoft.com/http://davidakenny.net/cm/fit.htmhttp://davidakenny.net/cm/fit.htm