7HuysamenSchepers.qxd Leadership and creativity have been studied extensively but separately; it seems that there is a void in the literature with regard to the correlation between the two (Kirkpatrick, 1991). Furthermore, research on the importance of leadership as a key to success in business has been established. Roodt (2001) and Schlechter (2001) showed that creative traits can explain an increase in profits, a decrease in stock losses and also labour turnover. Based on personal experience the researcher has found that one of the biggest challenges in leadership development is to identif y specific developmental needs, and focus training accordingly. Even more problematic is the selection of creative people. Leadership traits According to Ackermann, Schepers, Lessing, and Dannhauser (2000) research has made it clear that successful leaders are not like other people. The evidence indicates that there are certain core traits that significantly contribute to a business leader’s success. While none of the traits reviewed by Stogdill (1948, 1974) were found in all studies to be associated with leadership, the consistency with which some traits were found to be associated with leadership and the magnit udes of these associations are impressive. For example, the traits that most consistently showed high correlations with leadership were intelligence (Mann, 1959); dominance (Dyson, Fleitas and Scioli, 1972; Rychlak, 1963); self-esteem (Bass, 1957); task ability (Marak, 1964, Bass, 1961); and sociability (Kaess, Witryol and Nolan, 1961). Kreitner and Kinicki (2001, p. 555) states that we can no longer afford to ignore the implications of leadership traits because traits play a central role in how we perceive leaders. Kirkpatrick and Locke (1991) argue that there is less clear evidence for traits such as creativity, originality and flexibility. These authors believe that the key leadership traits help the leader to � acquire the necessary skills; � formulate an organisational vision, and an effective plan for pursuing it; � and take the necessary steps to implement the vision. If a positive correlation can be found between creativity and leadership style then the question arises as to whether some leaders are more creative than others. Transformational and Transactional Leadership It appears from the literature that Transformational Leadership as conceptualised by Burns (1978), is currently a major focus area. According to Ackerman et al., (2000, p.58) Transformational Leadership, probably serves as the most appropriate style in managing the changes presently taking place in South African organisations. Transactional Leadership st yle focuses on interpersonal transactions between managers and employees. Leaders are seen as engaging in behaviours that maintain a quality interaction between themselves and followers (Kreitner & Kinicki, 2001, p. 567). According to Gibson (1997, p.313) the leader helps the followers identif y what must be done to accomplish the desired results, i.e. better quality output, more sales or services and reduced cost of production. In helping the followers identif y what must be done, the leader takes into consideration each person’s self-concept and esteem needs. Transformational Leadership is viewed as a special case of Transactional Leadership, and unlike the external rewards promised by Transactional Leaders, the employee’s reward is internal (Gibson, 1997). Transformational Leadership According to Bass and Avolio (1985) Transformat ional Leaders motivate others to do more than they originally intended and often even more that they thought possible. They set more challenging expectat ions and t y pically achieve higher performance. Transformational Leaders do more with colleagues and followers than set up simple exchange agreements. R HUYSAMEN J M SCHEPERS J ZA AIMAN Programme Leadership in Performance and Change Department of Human Resources Rand Afrikaans University ABSTRACT The article argues that there is a statistically significant positive relationship between leadership style and creativity. It was found that there is a higher statistically significant positive relation between transformational leadership style and creativity than between creativity and the other leadership styles. Creativity (fluency of thought and originality) can partially be accounted for by the Multifactor Leadership Questionnaire. To demonstrate this in the study, the researcher uses, firstly the Multifactor Leadership Questionnaire to identif y leadership style, and secondly three measuring instruments that measure creativity. The argument is substantiated by the results of a statistical analysis of leadership style and creativity. OPSOMMING Hierdie artikel argumenteer dat daar ’n statisities-beduidende positiewe verband bestaan tussen leierskapstyl en kreatiwiteit. Daar is bevind dat daar ’n hoër statisties-beduidende positiewe verband tussen transformasionele leierskapstyl en kreatiwiteit is as tussen kreatiwiteit en die ander leierskapstyle. Kreatiwiteit (vlotheid van denke en oorspronklikheid) kan gedeeltelik verklaar word deur die Veelfaktorleierskapvraelys (MLQ). Om dit te illustreer, gebruik die navorsers in hierdie studie die Veelfaktorleierskapvraelys ten einde leierskapstyl te identifiseer en tweedens drie meetinstrumente wat kreatiwiteit meet. Die argument word gesubstansieer deur die resultate van ’n statistiese ontleding van leierskapstyle en kreatiwiteit. THE BEHAVIOURAL AND PERSONALITY CORRELATES OF TRANSACTIONAL AND TRANSFORMATIONAL LEADERSHIP Requests for copies should be addressed to: R Huysamen, Department of Human Resource Management, RAU University, PO Box 524, Auckland Park, 2006 53 SA Journal of Human Resource Management, 2003, 1 (2), 53-59 SA Tydskrif vir Menslikehulpbronbestuur, 2003, 1 (2), 53-59 Creativity Although there is seemingly less clear evidence of creativity as a trait that leaders possess, an abundance of research have been done on creativity as such (Getzels, 1964; Greeno, 1978; Sternberg, 1999). Divergent thinking and ideational fluency of thought According to Guilford (1957) it is in divergent thinking that we find the most obvious indications of creativity. Divergent thinking ability require individuals to produce several responses to a specific prompt. The more responses the individual produces, the higher the ideational fluency ability of the person (Guilford,1962). Associational fluency of thought Creative thinking has been viewed by Mednick (1962) as the recombining of associative elements that either meet specified requirements or are in some way useful. He argues that solutions may occur through mediation (combination via common elements). Mednick (1967) argues that the greater the number of associations that an individual has with the requisite elements of a problem, the greater the probability of his reaching a creative solution. To assess creativity, Mednick (1962) devised the Remote Associates Test (RAT). Originality of thought Originality of thought is defined by Cougar (1995, p.370) as ‘the capacity to produce unusual ideas, solve problems in unusual ways, and use things and situations in an unusual manner’. Originality consists of uniqueness and nonconformity with thought and action. Barron (1955) argues that original must be defined relative to usual, and the degree of originality must be specified in terms of incidence of occurrence. Thus the first criterion of an original response is that it should have a certain specifiable uncommonness. Sternberg (1999, p.450) states that there appears to be consensus that the t wo defining characteristics of creativity are originality and usefulness. Creativity and Problem Solving Guilford (1957) argues that creative steps are necessary in solving new problems. A problem only exists because the situation presents the necessity for new productions of some kind. Factors are abstractions of components from total activities. Some of the components are recognised as being more creative than others, for example, qualities of originality and fluency. Creativity is therefore not a gift possessed by rare and exceptional individuals (though some components are undeniably more creative), but all individuals are creative to some extent. This study has attempted to establish whether a correlation exists between creativity and behavioural leadership style. METHOD This section describes the method that was employed in the study, including a description of the target population, raters and measuring instruments. Participants/respondents All students currently enrolled for their MBA and M.Com. degrees comprised the population of interest. The unit of study (Cooper, 1998, p.215) consisted of students who are busy with their master’s degrees in three different universities in the Gauteng area in South Africa. The sampling method used is non-probability sampling as defined by Cooper (1998, p.237). For the purpose of this research judgement sampling was used as a type of purposive sampling. The researcher selected sample members who conformed to some attributes. The most common attribute of masters’ students is their work experience on managerial level. The researcher believes that although the use of students may imply homogeneous groups, this sample is more heterogeneous due to the fact that the respondents have managerial experience as a common attribute. The sample size (N) was 196. Measuring instruments Transformational Leadership was measured by the Multifactor Leadership Questionnaire (MLQ) developed by Bass and Avolio (1985). The MLQ measures a broad range of leadership style from Laissez-fair to Idealised influence. The MLQ consists of 45 items loading on four factors that represent the Full Range Model. Ideational fluency and originality of thought were measured by two divergent thinking ability tests similar to those developed by Chand and Runco (1993). The first test assessed ideational fluency and originality of thought in respect of a concrete concept, eg. � List the uses of a paper clip, and the second test assessed ideational fluency and originality of thought in respect of an abstract concept, eg. � List the possible solutions for crime in SA. Associational fluency of thought was measured by the Remote Associates Test as developed by Thompson (1993). Complete records were obtained in respect of 178 participants. Procedure The research procedure comprised the following steps: � The researcher administered the questionnaires in person. � Respondents were told that confidentiality would be maintained. � Respondents were asked to complete the Multifactor Leadership Questionnaire answer sheet. It was explained to the respondents that the questionnaire consists of statements about typical leadership behaviour and they were asked to indicate how often they behave in a certain way. The items required that the respondents should indicate how strongly they identif y with the behaviour. Because the questionnaire was self-explanatory the respondents did not have to provide their names. This took approximately 20 minutes. � Respondents then completed the Remote Associates Test. A number of practice examples were given before beginning the test. The amount of time allocated for the Remote Associates Test was 20 minutes. � Respondents then completed the two divergent thinking ability tests. Explicit instructions were given in order to ensure that respondents understood that the alternatives generated should be different from one another. The respondents were asked to be as creative as possible when generating alternatives. The amount of time allocated was three minutes for each test. � All four sets of tests were stapled together for each respondent. � The researcher and two independent raters then evaluated the divergent thinking ability tests in order to establish originality of thought. A five-point scale for individual components was used where 1= very common and 5= very original. � The researcher evaluated the Remote Associates Test. � All four tests were sent to the Statistical Consultation Service at the Rand Afrikaans University for processing. RESULTS The factor analysis of the MLQ Two different types of instruments were used. Firstly the MLQ was used to identif y leadership style and secondly, three creativity measuring instruments, measuring fluency and originality of thought were used. The 45 items of the MLQ were subjected to a principal factor analysis. The analysis was beased on 196 complete records. As a first step, the items were intercorrelated and the eigenvalues of the unreduced intercorrelation matrix were computed. As fifteen of the eigenvalues were greater than unity, fifteen factors were extracted and rotated to simple structure by means of a Varimax rotation. HUYSAMEN, SCHEPERS, ZA AIMAN54 To overcome the effect of differential skewness of items, subscores in respect of each factor were computed by adding the items with high loadings on a factor, together. The subscores were then intercorrelated and subjected to a principal factor analysis. The eigenvalues of the unreduced intercorrelation matrix are presented in Table 1. TABLE 1 EIGENVALUES OF UNREDUCED INTERCORRELATION MATRIX IN RESPECT OF THE MLQ Root Eigenvalue % of Variance Cumulative % 1 3,911 26,071 26,071 2 1,411 9,406 35,477 3 1,233 8,218 43,695 4 1,182 7,881 51,576 5 0,958 6,384 57,960 6 0,910 6,069 64,029 7 0,787 5,247 69,276 8 0,769 5,128 74,405 9 0,730 4,865 79,270 10 0,648 4,319 83,588 11 0,628 4,188 87,776 12 0,575 3,835 91,611 13 0,483 3,217 94,829 14 0,434 2,892 97,721 15 0,342 2,279 100,000 Trace 15,00 Four factors were extracted, as four of the eigenvalues were greater than unity. Factor 1 had substantial loadings on 16 items and Factor 4 on 19 items yielding Cronbach alphas of 0,767 and 0,885 respectively. Factors 2 and 3 were poorly determined, having loadings on four and six items respectively. The reliability of the corresponding scales were 0,55 for Factor 2 and 0,479 for Factor 3. Accordingly a three-factor solution was tried. Factors 3 and 4 combined into a single factor with a reliability of 0,84. No items were lost. The reliability of Factor 4 fell, however, from 0,88 to 0,847. The obtained factor matrix was rotated to simple structure by means of a Direct Oblimin Rotation, as given in Table 2. TABLE 2 ROTATED FACTOR MATRIX OF THE MLQ (DIRECT OBLIMIN ROTATION) VARIABLES ITEMS K FACTOR 1 FACTOR 2 FACTOR 3 h²j Subtest 10 5,29 2 0,635 -0,252 0,178 0,679 Subtest 8 9 1 0,577 0,018 -0,003 0,349 Subtest 5 19,23,6 3 0,526 0,080 0,070 0,023 Subtest 7 2,8,25 3 0,489 0,077 0,030 0,435 Subtest 9 18,40 2 0,450 0,095 0,204 0,312 Subtest 2 13,11,14,33,1 5 0,439 -0,040 0,222 0,428 Subtest 15 7 1 -0,116 0,085 -0,066 0,253 Subtest 4 12,28,20 3 -0,209 0,578 -0,087 0,329 Subtest 14 3 1 0,061 0,511 -0,109 0,339 Subtest 13 17 1 0,034 0,191 0,139 0,424 Subtest 6 41,42 2 0,056 -0,030 0,623 0,416 Subtest 1 31,36,44,16, 13 0,336 -0,050 0,593 0,225 37,35,30,38,32, 34,43,45,39 Subtest 12 15 1 -0,070 -0,112 0,489 0,060 Subtest 11 10,26,21 3 0,240 -0,090 0,486 0,260 Subtest 3 27,22,24,4 4 0,018 0,034 0,141 0,035 Number of items 45 17 5 23 4,547 per factor Note: Factor 3 has been reflected Table 2 shows that three factors are reasonably well determined, with substantial loadings on all three. However, from an inspection of the communalities it is apparent that several of the subscores share only a small proportion of their variance with the other subscores. In order to identif y the obtained factors, the items associated with each factoe were grouped into categories as given by Bass (1999). Bass’s classification (1999) is presented in Table 3. TABLE 3 ITEMS ASSOCIATED WITH THE THREE FACTORS OF THE MLQ ITEMS FACTOR 1 FACTOR 2 FACTOR 3 Transformational 9,10,15,21,25, 13 (with a 2,6,8,14,18,19, leadership 26,30,31,32 negative loading) 23,29,34 and 36 Transactional 16,25 3,4,12,20,22,24 1,11 and 17 leadership and 27 Non- None 5,7 and 33 28 Transactional leadership Leadership 37,38,39,40,41, None None outcome 42,43,44 and 45 Total number 20 11 14 of items Note: Categorisation of items according to Bass (1999) Table 3 shows that Factor 1 is strongly representative of items categorised as Transformational Leadership and Leadership Outcome. There are only two items categorised as Transactional Leadership. Factor 1 can therefore be identified as Transformational Leadership/Leadership outcome. Factor 2 is representat ive of items categorised as Transactional Leadership and Non-Transactional Leadership (Laissez Faire). There is also one item with a negative loading, categorised as Transformational Leadership. The items listed as Transactional Leadership, all deal with failure to take action when required, and keeping track of mistakes, irregularities, failures and complaints. Factor 2 was identified as Non-Transactional Leadership. Factor 3 is representative of items categorised as Transform- ational Leadership. There are three items categorised as Transactional Leadership. Factor 3 is therefore identified as Transformational/Transactional Leadership. For the purpose of this research the items in respect of each factor were compared to the items as given in Table 3. The three factors of this research differ from those of Bass (1999) in the sense that the factor analysis revealed some differences in the items grouped under the different factors. Although this research also postulated three factors, the items that make up the three factors differ in minor ways: � Bass (1999) argues that Factor 1 represents Transformational Leadership/Leadership outcome. The items are similar to Factor 3 of this research. � According to Bass’s research (1999), Factor 2 represents the Laissez Faire leadership style (a finding confirmed by this research). However, the fact that only five items remained in Factor 2 (after the three factor solution was done) decreased the reliability coefficients of both Factors 2 and 3. These items focus more on the leaders’ ability to solve problems and rectif y mistakes. When problems are solved and mistakes rectified, the leader achieves an outcome. These items therefore do not focus on any specific leadership style. Bass (1999) states that his Factor 3 represents the Transformational/Transactional Leadership style. The items are similar to Factor 1 of this research. It is interesting to note that 65% of the items under Factor 1 represent the Transformational Leadership style in its pure form. The researcher came to the BEHAVIOURAL AND PERSONALITY CORRELATES 55 conclusion that the three factors of this research produced essentially the same categories of information as those produced by Bass (1999) on leadership i.e. Transformational/ Transactional, Laissez faire and Transformational/Leadership outcome. The intercorrelations bet ween the factors are presented in Table 4. TABLE 4 FACTOR CORRELATION MATRIX OF THE MLQ FACTOR 1 2 3 1 1,000 -0,134 0,522 2 -0,134 1,000 0,010 3 0,522 0,010 1,000 Table 4 shows that Factors 1 and 3 are positively correlated. Three scales were formed corresponding to the three factors obtained. To determine the reliability of these scales, they were subjected to item analysis, using the NP 50 program. The item statistics in respect of the first scale are presented in Table 5. TABLE 5 ITEM ANALYSIS OF SCALE 1 OF THE MLQ Mean of Standard Reliability Item item deviation index of total of item item correlation ITEM N Xg Sg rgxsg rgx A1 196. 2,526 1,139 0,407 0,345 A2 196. 2,816 0,881 0,352 0,400 A5 196. 3,393 1,064 0,482 0,453 A6 196. 2,740 1,042 0,338 0,324 A7 196. 3,439 1,008 0,290 0,288 A8 196. 3,168 0,938 0,460 0,491 A9 196. 3,184 0,904 0,493 0,545 A11 196. 2,888 1,090 0,571 0,524 A13 196. 3,291 0,896 0,547 0,610 A14 196. 3,265 0,823 0,506 0,615 A18 196. 3,087 0,927 0,365 0,394 A19 196. 3,010 1,146 0,475 0,414 A23 196. 3,143 0,911 0,560 0,615 A25 196. 3,010 0,982 0,339 0,345 A29 196. 3,464 0,774 0,372 0,481 A33 196. 2,801 1,299 0,515 0,396 A40 196. 3,051 0,802 0,464 0,578 Cronbach alpha = 0,752 Mean of test = 52,276 Standard deviation of test = 7,536 Number of items = 17 Table 5 shows that the obtained reliability in respect of Scale 1, is 0,752. The item total-correlations range from 0,288 to 0,615. The scale is internally consistent. No items were rejected. The reliability coefficient is sufficient for the purpose of this research. The item statistics in respect of the second scale are presented in Table 6. Table 6 shows that the obtained reliability, in respect of Scale 2, is 0,494. The item- total correlations range from 0,482 to 0,628. None of the items were rejected, but there are too few items. The item statistics of the third scale are presented in Table 7. TABLE 6 ITEM ANALYSIS OF SCALE 2 OF THE MLQ Mean of Standard Reliability Item item deviation index of total of item item correlation ITEM N Xg Sg rgxsg rgx A3 196. 1,204 1,163 0,672 0,578 A12 196. 0,582 1,032 0,693 0,672 A17 196. 1,628 1,244 0,599 0,482 A20 196. 0,888 1,197 0,752 0,628 A28 196. 0,628 0,960 0,517 0,538 Cronbach alpha = 0,494 Mean of test = 4,929 Standard deviation of test = 3,234 Number of items = 5 TABLE 7 ITEM ANALYSIS OF SCALE 3 OF THE MLQ Mean of Standard Reliability Item item deviation index of total of item item correlation ITEM N Xg Sg rgxsg rgx A4 196. 2,245 1,203 0,264 0,219 A10 196. 2,673 1,098 0,467 0,425 A15 196. 2,867 0,973 0,425 0,437 A16 196. 3,199 0,869 0,532 0,612 A21 196. 3,296 0,919 0,415 0,452 A22 196. 2,209 1,270 0,351 0,277 A24 196. 2,010 1,203 0,359 0,298 A26 196. 3,036 0,936 0,579 0,619 A27 196. 1,898 1,240 0,409 0,330 A30 196. 3,291 0,812 0,442 0,544 A31 196. 3,352 0,697 0,453 0,649 A32 196. 3,082 0,867 0,439 0,506 A34 196. 3,046 0,867 0,402 0,464 A35 196. 3,469 0,825 0,521 0,631 A36 196. 3,423 0,716 0,429 0,600 A37 196. 3,077 0,784 0,468 0,596 A38 196. 3,219 0,743 0,435 0,585 A39 196. 2,719 0,976 0,520 0,533 A41 196. 3,454 0,682 0,320 0,469 A42 196. 3,158 0,945 0,585 0,619 A43 196. 3,393 0,675 0,372 0,552 A44 196. 3,240 0,715 0,400 0,559 A45 196. 3,036 0,914 0,386 0,423 Cronbach alpha = 0,837 Mean of test = 68,393 Standard deviation of test = 9,971 Number of items = 23 Table 7 shows that the obtained reliability of Scale 3 is 0,837. The item-total correlations range form 0,219 to 0,649. None of the items were rejected. The scale consists of 23 items. Factor analysis of creativity measures The three creativity ability tests measure fluency and originality of thought. The two divergent thinking ability tests measure fluency and originality of thought. The fluency of thought scores (Scores 2 and 4) were calculated by adding the number of responses, and these responses were then subjected to an evaluation of originality (Scores 1 and 3) by independent raters. Four scores were derived from these fluency tests. HUYSAMEN, SCHEPERS, ZA AIMAN56 Secondly, the Remote Associate test (RAT) measures associational fluency of thought. This ability test is divided into two sections (Scores 5 and 6) and comprises 15 items per section. Scores are calculated by adding the number of correct responses for each item. Two scores were derived from this test. The six derived scores of the creativity tests were subjected to a factor analysis. The analysis was based on a sample of 178 participants. As a first step, the six scores were intercorrelated and the eigenvalues of the intercorrelation matrix were determined. The obtained eigenvalues are given in Table 8. TABLE 8 EIGENVALUES OF UNREDUCED INTERCORRELATION MATRIX OF THE CREATIVITY MEASURES Root Eigenvalue % of Variance Cumulative % 1 2,764 46,064 46,064 2 0,943 15,723 61,787 3 0,731 12,187 73,974 4 0,609 10,147 84,120 5 0,559 9,309 93,429 6 0,394 6,571 100,000 Trace 6 From Table 8 it is clear that only one eigenvalue is greater than unity. Accordingly only one factor was extracted and is given in Table 9. TABLE 9 FACTOR MATRIX OF THE CREATIVITY MEASURES Factor 1 h²j Score 1: Paperclip (originality) 0,714 0,509 Score 2: Paperclip (fluency) 0,678 0,459 Score 3: Crime (originality) 0,670 0,449 Score 4: Crime (fluency) 0,477 0,228 Score 5: RAT 1 0,476 0,227 Score 6: RAT 2 0,532 0,283 From Table 9 it is evident that all six scores have substantial loadings on the same factor. This factor represents fluency and originalit y of thought. Scores 1 and 3 are comparable (originality of thought). Scores 2 and 4 are comparable (f luency of thought). Scores 5 and 6 are coded. The communalities of the six scores vary from 0,227 to 0,509. Scores 4, 5 and 6 have the lowest communalities, indicating that they have less in common with the other measures, but are not necessarily poor measures of creativity. The means and standard deviations in respect of the creativity measures are presented in Table 10. Table 10 shows that the obtained reliability of the Creativity Scale is 0,735. None of the scores were rejected. The mean of the Remote Associates Test 1 was lower than that of the other measuring instruments. Fewer correct answers were found in respect of this measuring instrument. The mean of the Crime Rater total (originality) is slightly less than that of the other measuring instruments. Table 11 shows the Pearson product moment (Cooper,1998) correlations between the three scales of the MLQ and the seven scores of creativity. Some of the variables are positively related and others are negatively related. TABLE 10 MEANS AND STANDARD DEVIATIONS IN RESPECT OF THE CREATIVITY MEASURES Mean Std Dev N 1. Paperclip (originality) 8,9719 4,4702 178 2. Paperclip (fluency) 8,8371 5,2196 178 3. Crime (originality) 6,1517 3,0018 178 4. Crime (fluency) 8,5225 3,5815 178 5. RAT 1 1,0337 0,9909 178 6. RAT 2 9,1124 3,5668 178 Cronbach alpha = 0,735 Scale 1 (Transformational/transactional) of the MLQ correlated positively with all three creativity tests. Although this scale represents both the Transformational and Transactional Leadership styles, 65% of the items loaded on Transformational Leadership style. The lowest correlation was found between Scale 1 (Transformational/Transactional) and Score 3 (Crime originality). There is a positive relation between Scores 1 and 3 (r=0,481; p=0,01). Scale 1 (Transformational/Transactional) has a positive relation with both scales of the RAT. A positive relation was also found between Scale 3 (Transformational/Leader outcome) and Scale 1 (Transformational/Transactional) of the MLQ (r=0,593; p=0,01). It is the researcher’s contention that the reason for this phenomenon can be found in the fact that both the scales are made up of items that measure Transactional and Transforma- tional Leadership styles. As would be expected Scales 1 (Transformational/Transactional) and 2 (Laissez faire leadership) are inversely related (r= –0.192; p=0,01). Scale 3 (Transformational/Leadership outcome) is inversely related to the six fluency/originality scores, but should be reflected because factor 3 of the MLQ has been reflected. Two stepwise regression analyses were done. In the first analysis, Scale 1 (Transformational/Transactional) of the MLQ served as the independent variable and in the second analysis Scales 1 (Transformational/Transactional) and 3 (Transformational/ Leader outcome) of the MLQ were used. Table 12 provides a model summary of the regression analysis of the independent variables on the dependent variable. The regression analyses indicate that only two predictor variables Scales 1 and 2 hold promise of the MLQ. A multiple correlation of 0,263 was obtained in terms of the first model, which indicates that 6,4% of the variance of fluency/originality can be accounted for by the first predictor variable, i.e. Scale 1. A multiple correlation of 0,332 was obtained in terms of the second model, indicating that 10% of the variance in fluency/originality can be accounted for by Scales 1 and 3. The combination of Scales 1 and 3 therefore accounts for a greater proportion of the variance of creativity. Table 13 gives the analysis of variance. From the analysis of variance it is clear that the linear regression accounts for a statistically significant proportion of the total variance. F(2,175) = 10,855; p(F)<0,01. All the regression coefficients are statistically significant. An estimate of creativity can be obtained from the following regression equation: Y = 26,677 + 13,269 (MLQ 1) – 8,319 (MLQ 3) BEHAVIOURAL AND PERSONALITY CORRELATES 57 ^ TABLE 12 REGERSSION ANALYSIS OF THE INDEPENDENT VARIABLES ON THE DEPENDENT VARIABLE Model Summary Model R R Square Adjusted Std. Error R Square of Estimate 1 0,263(a) 0,069 0,064 14,132 2 0,332(b) 0,110 0,100 13,86 a Predictors: (Constant), MLQ 1 b Predictors: (Constant), MLQ 1, MLQ 3 From the standardised regression coefficients (ßeta coefficients) as given in Table 13 it is clear that Transformational/ Transactional Leadership (MLQ 1) carries a greater weight in the equation than the Transformational/Leader outcome (MLQ 3). TABLE 13 ANALYSIS OF VARIANCE Multiple R: 0,332 R Square: 0,110 Adjusted R Square: 0,100 SOURCE OF VARIATION DF SUM OF MEAN SQUARES SQUARE Regression 2 4171,017 2085,51 Residual 175 33620,51 192,117 F (2,175)=10,885; p (F) < 0,01 Regression coefficients MODEL REGRESSION Std Error STANDARDISED t p(t) COEFFICIENTS of B COEFFICIENTS B Beta (Constant) 26,677 8,075 3,304 0,001 MLQ 1 13,269 2,850 0,407 4,656 0,000 MLQ 3 -8,319 2,932 -0,248 -2,838 0,005 Dependent variable: R TOT Predictors (Constant) MLQ 1, MLQ 3 HUYSAMEN, SCHEPERS, ZA AIMAN58 TABLE 11 PEARSON’S PRODUCT MOMENT CORRELATIONS OF MLQ AND MEASURES OF CREATIVITY Correlations Paperclip Paperclip: Crime in SA Crime in Remote Remote R_TOT MLQ1 MLQ2 MLQ3 Rater Total Total Rater Total SA: Total Associates Associates Test 1 Test 2 Paperclip Pearson 1 0,486 ** 0,481 ** 0,240 ** 0,339 ** 0,425 ** 0,772 ** 0,208 ** 0,006 0,025 Rater Total Correlation N 196 193 196 192 185 196 178 196 196 196 Paperclip: Pearson 0,486 ** 1 0,391 ** 0,405 ** 0,340 ** 0,318 ** 0,792 ** 0,158 * 0,049 -0,094 Total Correlation N 193 193 193 189 182 193 178 193 193 193 Crime in SA Pearson 0,481 ** 0,391 ** 1 0,439 ** 0,295 ** 0,279 ** 0,704 ** 0,144 * 0,040 -0,016 Rater Total Correlation N 196 193 196 192 185 196 178 196 196 196 Crime in Pearson 0,240 ** 0,405 ** 0,439 ** 1 0,175 * 0,152 * 0,610 ** 0,167 * 0,105 0,001 SA: Total Correlation N 192 189 192 192 181 192 178 192 192 192 Remote Pearson 0,339 ** 0,340 ** 0,295 ** 0,175 * 1 0,301 ** 0,472 ** 0,194 ** -0,046 0,092 Associates Correlation Test 1 N 185 182 185 181 185 185 178 185 185 185 Remote Pearson 0,425 ** 0,318 ** 0,279 ** 0,152 * 0,301 ** 1 0,634 ** 0,225 ** -0,069 0,007 Associates Correlation Test 2 N 196 193 196 192 185 196 178 196 196 196 R_TOT Pearson 0,772 ** 0,792 ** 0,704 ** 0,610 ** 0,472 ** 0,634 ** 1 0,263 ** 0,009 -0,013 Correlation N 178 178 178 178 178 178 178 178 178 178 MLQ1 Pearson 0,208 ** .158 * 0,144 * 0,167 * 0,194 ** 0,225 ** 0,263 ** 1 -0,192 ** 0,593 ** Correlation N 196 193 196 192 185 196 178 196 196 196 MLQ2 Pearson 0,006 0,049 0,040 0,105 -0,046 -0,069 0,009 -0,192 ** 1 -0,153 * Correlation N 196 193 196 192 185 196 178 196 196 196 MLQ3 Pearson 0,025 -0,094 -0,016 0,001 0,092 0,007 -0,013 0,593 ** -0,153 * 1 Correlation N 196 193 196 192 185 196 178 196 196 196 **Correlation is significant at the 0.01 level (2-tailed). * Correlation is significant at the 0.05 level (2-tailed). DISCUSSION The researcher concludes that there is a statistically significant positive relationship between leadership style and creativity. There is a higher statistically significant positive relation between Transformational Leadership style and creativity than bet ween creativit y and the other leadership st yles. Approximately 10% of the variance of creativity (fluency and originality of thought) can be accounted for by the MLQ. In the light of the above-mentioned finding the MLQ can be recommended as a test in the selection process. It would also be worthwhile pursuing means of upgrading the MLQ. For the purpose of this investigation respondents were asked to rate themselves in respect of the MLQ. It is recommended that if similar research is to be done in future, the superiors of the respondents should also complete the instrument and rate the respondents from their perspective (360° approach). Despite the positive results obtained with the battery of creativity measures, certain reservations concerning them must be mentioned. Two of the measuring instruments were pertinent to divergent thinking ability. The main drawback from a psychometric point of view, is that the measurement of divergent thinking necessitates an element of subjectivity in the scoring of the tests. Even where the actual scoring procedure is objective, a subjective assumption is made to the effect that the score so obtained provides a measure of whatever the test is supposed to measure. Even more questionable is the assumption that originality can be gauged by weighting responses in terms of the rarit y of their occurrence in the population. The possibility of rater bias must also be recognised. The correlations obtained by intercorrelating the creativity measures are generally so high as to suggest that the measures overlap to a considerable extent. This was also borne out by the general factor of creativity obtained in the factor analysis of the measures. Since the measuring instruments were not designed for factor analysis per se, the finding of a general factor cannot be cited as proof that a general creativity factor exists, but it does suggest that the different instruments used do not measure separate factors beyond fluency of thought and originality. Though there is evidence that creativity is manifested in leadership style, accurate judgements of creative ability can only be made with reference to an individual’s actual output. Ideally the dependent variable should be validated against real life work performance. This can only be achieved if use can be made of experts, who have the knowledge and skills required. But given such an opportunity, one must recognise the possibility of rater bias. The Remote Associates Test has shown promise as a predictor of associational fluency, an element of creativity. The biggest problem with this measuring instrument is the difficulty of establishing an objective and reliable scoring system for the South African context. Problem – solving is a concept that relates to both leadership style and creativity. A problem – solving orientation was identified by the MLQ (Factor 3), but the ability to solve problems as such was not measured in this research as the instruments focussed on f luency of thought and originalit y. It is therefore recommended that a test that measures problem – solving ability should be included in future research. REFERENCES Ackermann, C.P., Schepers, J.M., Lessing, B.C., Dannhauser, Z. (2000). Die faktorstruktuur van Bass se veelfaktor- leierskapvraelys in die Suid-Afrikaanse konteks. Journal of Industrial Psychology, 26 (2), 58-65. Bass, B.M. (1957). Leadership opinions and related characteristics of salesmen and sales managers, in R.M. Stogdill and A. E. Coons (Eds.), Leader behavior its description and measurement, Columbus: Ohio State University, Bureau of Business Research. Bass, M.M. (1985). Leadership and performance beyond expectations. New York: Free Press. Bass, B.M., Waldman, D.A., Avolio, B.J., Bebb, M. (1987). Transformational Leaders: The falling dominoes effect. Group and Organization Studies, 12 pp.73-87. Bass, B.M. (1999). Two decades of research and development in Transformational Leaders. European Journal of Work and Organisational Psychology, 9 (1), 9-23 Burns, J.M. (1978). Leadership. New York: Harper and Row. Chand, I, Runco, M.A. (1993). Problem finding skills as components of the creative process. Journal of Personality and Individual Differences, 14, 155-162. Cooper, D. R. (1998). Business Research Methods. Singapoure: McGraw-Hill. pp. 30-32. Couger, J. D. (1995). Creative problem solving and opportunity finding, USA: Boyd and Fraser Publishing company. Dyson, J.W., Fleitas, D.W., and Scioli, F.P. (1972). The interaction bet ween Leadership Personalit y and Decisional Environments, Journal of Social Psychology, 86 pp. 29-33. Getzels, J.W. (1964). Creative thinking, problem solving and instruction. In Hilgard, E.R. (Ed), Theories of learning and instruction. Chicago, IL: University of Chicago press. Gibson, J.L, Ivancevich, J. M. Donnelly, J. H. (1997). Organisations: behaviour, structure, process. USA: Rob Zwettler. p. 315-319. Gill, J. and Johnson, P. (1991). Research methods for managers, London: Paul Chapman Publishing Ltd. Guilford, J.P. (1957). Creative Abilities in the Arts. Psychological Review, 64, 110-118. Guilford, J.P. (1962). Creativit y: Its measurement and Development, p.156-168. In: A source book for creative thinking. Parnes, S.J. and Harding, H.F., (Eds). New York: Charles Scribner’s Sons. Jankowitz, A.D. (1991). Business Research Projects, London: Chapman and Hall. Kaess, W.A., Witryol, S.L. and Nolan, R.E. (1961). Reliability, sex differences and validity in the leadership discussion group, Journal of Applied Psychology, 45 pp. 345-350. Kirkpatrick, S.A., Locke, E.A. (1991). Leadership: do traits matter? Academy of Management Executive, 5 (2) p. 48. Kreitner, R., Kinicki, A. (2001). Organizational Behavior. (5th ed.) New York: Mc Graw -Hill, pp. 548-571. Mann, R.D. (1959). A review of the relationships between personality and performance in small groups. Psychological Bulletin, 56, 4 pp.241-270. Marak, G.E. (1964). The evolution of leadership structure, Sociometry, 27 pp. 174-182. Mednick, S.A. (1962). The associative basis for the creative process. Psychological Review, 69, pp. 220-232. Roodt, G. (2001). Leadership still the cornerstone of success. Management Today, 17, (3), 8. Schlechter, A. (2001). Achieve performance excellence by practicing leadership effectively. Management Today, 17 (3), pp. 42-45. Sternberg, R. J. (1999). Handbook of creativity. United Kingdom: Cambridge University Press. Stogdill, R.M. (1948). Personal factors associated with leadership: A survey of the literature. Journal of Psychology, 25 pp.35-71. Thompson, T. (1993). Remote associate problem sets in performance feedback paradigms. Journal of Personal and Individual Differences, 14, (1), pp.11-14. Torrance, E.P. (1984). Torrance tests of creative thinking, Chicago: University of Chicago Press, p. 3. BEHAVIOURAL AND PERSONALITY CORRELATES 59