THE MOTIVES 0F POLIUCIANS AND BUSENESSMEN E'hesis for the Degree of M. A. MECHIGAN STATE UNWERSITY MARY LOUISE SZARI 1969 J'I'jhfi.)l ‘3 ’Ofifi ‘n}, “0. Mn... :1? {ate Univ-'3! x. .y filo-'- -‘--" a.“ u -. . . .vm I 1__ ',_—-——»‘,_:f-sa——. *5 I: airtime av ’ HMS G SUNS' eayzmeuag I w: It 71 .247 _‘.__‘a, ABSTRACT THE MOTIVES OF POLITICIANS AND BUSINESSMEN BY Mary Louise Szari Interview and TAT data on 94 businessmen classi- fied as leaders, subleaders or inactives in party politics or in civic affairs were available from a previous study (Browning, 1960). The TAT's were scored for need for power (a Pow), need for Influence (g Inf), need for Achieve- ment (E_Ach) and need for Affiliation (E Aff). All motive scores were corrected for story length. Analyses of variance and selected pairwise non- parametric tests were done. The results of the data analyses are as follows: There are no motivational dif- ferences between five of the six groups on the four motives: The Party Leaders, Party Subleaders, Party In- actives, Civic Subleaders, and Civic Inactives. Only the Civic Leaders differ from other groups. They have higher a Pow and lower 2 Aff than the Civic Subleaders and higher a Pow than the Party Leaders. The fact that all men were businessmen was sug- gested as an explanation for the homogeneity of motives. Mary Louise Szari The Civic Leaders are unique in this sample and it was suggested that the position of Leader in Civic affairs allows a freer expression of E_Pow strivings than a Party Leader position. Activity in this sample may simply be the result of a recruitment process by the Parties or Mayor. Methodological procedures which may have elimi- nated useful variance and arguments opposed to the story length correction were considered. Differences between Browning's conclusions and Szari's conclusions were re- solved when the statistical bases were examined. The results of exploratory analyses suggest that the Civic Group has higher 2 Pow than the Party group, that the Party group has higher n_Aff than the Civic group and that a person with low n_Inf and high 2_Aff will tend to be a Party member. An interaction between E_Ach and n Pow suggested that activity is related to two combinations: (a) high E_Pow and high n_Ach and (b) low E Pow and low E_Ach. Other results from the exploratory analyses were regarded as random effects. THE MOTIVES OF POLITICIANS AND BUSINESSMEN BY Mary Louise Szari A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF ARTS Department of Psychology 1969 DEDICATION To Allen, Mom and Dad ACKNOWLEDGMENTS I wish to give special thanks to Dr. Browning, who COOperated so willingly by allowing me the use of his original data. I also wish to thank Dr. Uleman for his aid throughout the duration of this study, and Drs. Barclay and Zucker for their criticisms which improved the qualr ity of the thesis considerably. iii TABLE OF CONTENTS LIST OF TABLES O O O O O C I O O O 0 INTRODUCTION . . . . . . . . . . . . Browning's Study . . . . . . . . Browning' 3 Results . . . . . . Purpose of the Present Study . . METHOD 0 O I O O O O O O O O O O O 0 Subjects . . . . . . . . . . . . Procedures . . . . . . . . . . . Story Length . . . . . . . . . . Correction for Story Length . . . Considerations in Analyses of Variance Data Analyses Used . . . . . . . Additional Analyses . . . . . . . RESULTS 0 O O O O O O O O O O O O 0 Analyses of Variance . . . . . . Non—parametric Tests . . . . . . DISCUSSION 0 O O O O O I O I O O 0 Additional Comments . . . . . . . Data Collection . . . . . . . . . Data Analysis . . . . . . . . Browning-Szari Differences in Conclusions Brief Summary . . . . . . . . . . LIST OF REFERENCES . . . . . . . . . APPENDICES O O O O O C O O I O O O O A COPIES OF THE SIX TAT PICTURES USED IN BROWNING'S STUDY . . . iv Page vi O‘U'IN 11 ll 11 13 15 20 21 23 24 24 24 30 32 33 33 36 38 39 41 41 Page B EXPLORATORY ANALYSES AND RESULTS . . . . 47 C CELL MEANS, F-VALUES AND PROBABILITY LEVELS FOR THE ANALYSES OF VARIANCE COMPUTED WITH UNCORRECTED MOTIVE SCORES AND STORY LENGTH . . . . . . . . 76 LIST OF TABLES Table Page 1. Size and description of the sample used in the Browning study . . . . . . . . . . . . 4 2. Spearman rank order correlations between story length and Browning's motive scores . . . . . . . . . . . . . . . . . . 9 3. Reliability scores for the five under- graduate scorers on n_Pow and n_Inf . . . 12 4. Spearman rank order correlations between story length and Szari uncorrected and corrected motive scores . . . . . . . . . l4 5. Cell means of the motive scores found significant in the two analyses of variance using story length and un- corrected motive scores as dependent variables . . . . . . . . . . . . . . . . l6 6. Spearman rank order correlations among uncorrected Szari motive scores and correlations among corrected Szari motive scores . . . . . . . . . . . . . . l8 7. The coefficients of determination between pairs of motives . . . . . . . . . . . . . 19 8. Analyses of variance designs used in this Study 0 O O O O O O O I O O O O O O O O O 2 2 9. Analyses of Variance with the four cor- rected motives as dependent variables. Cell means, F-values and probability levels are presented . . . . . . . . . . . 25 10. Pairwise comparisons of Party and Civic Leaders and Subleaders . . . . . . . . . . 29 vi Table Page 11. Browning and Szari conclusions compared . . 37 12. Summary of the analyses of variance used in the exploratory phase of this study . . 49 13. Summary of the tests used in the explora- tory phase of this study . . . . . . . . . 51 14. Active vs. Inactive analyses of variance using one motive as an independent variable; analyses are presented ac- cording to the dependent variable used . . 53 15. Summary of ten significant effects and the analyses relating to them, and probability levels of effects if appropriate . . . . . . . . . . . . . . . 63 16. Cell means of an interaction between Party-Civic, n Inf and n Aff when either motive-is the dependent variable . . . . . . . . . . . . . . . . . 65 17. Cell means of an interaction between Active-Inactive, n_Pow, and n Ach when either motive is the dependent variable . 66 18. Cell means of the interaction between Active-Inactive, Party-Civic, n Aff and n_Pow when n_Aff is the dependent variable . . . . . . . . . . . . . . . . . 68 19. Cell means of seven significant results computed by non-parametric tests . . . . . 71 20. Cell means and probability levels of the comparisons of subgroups among themselves . . . . . . . . . . . . . . . . 72 21. Analyses of variance with the four uncor- corrected motives and story length as dependent variables. Cell means, de- grees of freedom (df), F-values and probability levels (p) are given . . . . . 76 vii INTRODUCTION The motivations of persons entering politics has long been under the scrutiny of the general public and certain assumptions have become part of public lore: “politicians are power-hungry, selfish, aggressive and unconcerned with others' welfare. Sampson (1965) feels that power and love are diametrically opposed capabili- ties and that developing one necessarily means weakening the capacity for the other. This of course is congruent with the image of the uncaring, selfish politician. A political scientist, Lasswell (1948) saw power as the cornerstone of politics, involving subordinates' expecta- tions of the authoritative and controlling influence of the one—in-power. Lasswell believed that the political man demands power for himself and accentuates his expecta- tions concerning power, often in order to compensate for low estimates of himself. Milbrath (1965) on the other hand, who studied various personality variables and poli- tical involvement, suggest that it is unlikely for persons with high power motivation to enter politics, or to be very successful in the political area if they do enter it. Lane (1959) agrees in part with both men. He agrees with Lasswell that the need for power is not expressed in politics directly, but is expressed either in a sublimated I or disguised form. He also agrees with Milbrath that it is not likely that a man would attain a high office if he were enamoured with power, but would rather fulfill minor roles in a democracy. One novel comment Lane presents is !that it is just as likely for the man highly interested in power to be active in non-political areas as well as political areas. Many studies have been undertaken in the area of the variables influencing political involvement and poli- tical behavior (Bell, 1967; Lindenfeld, 1962). One study dealing specifically with the motivations of politicians and attempting to answer the question of what kind of per- son becomes a politician is the doctoral dissertation of Rufus Browning (1960). It is the basis for the present paper and will be presented in detail. Browning's Study Browning, a political scientist, interviewed and administered a modified Thematic Apperception Test (TAT) to 94 businessmen in a medium-sized eastern city. These businessmen had been rated as to leadership level in a previous study (Dahl, 1961) and were divided into six categories: Party leaders, Party subleaders, Party in- actives, Civic leaders, Civic subleaders and Civic inac- tives. The Party leaders, Party subleaders, Civic leaders and Civic subleaders were combined into an Active group, while the Party inactives and Civic inactives composed the Inactive sample (see Table l). The Party leaders and subleaders were businessmen active in a major political party in either an elective, appointive or ward capacity. The Civic leaders and subleaders were those businessmen active in urban renewal policy-making or members of the mayor's advisory council. The Inactive sample was com- posed of men who were matched to the Party subleaders and Civic subleaders by Browning on the basis of ethnic and religious background, career level and specific occupa~ tion, type and size of business and residence. The Inac- tive samples were nearly identical in average age and education to the respective matched subleaders. Interviews with the men were preceded by a letter and phone call of introduction. There was a standard in- terview procedure prepared, but in practice, infrequently followed. The modified TAT, a set of six pictures (No. 5, 28, 83, 9, 24 and 53 in that order, listed in Atkinson, 1958, p. 835) was given to the men in their offices or homes. They were given the six pictures one at a time and asked to take a "test of imagination . . . a test of the capacity to think on your feet." Their verbal re- sponses to the standard four TAT questions (typed on a separate sheet) were taken down by Browning in shorthand and transcribed by him later. Table l.--Size and description of the sample used in the Browning study. Party Civic Actives Leaders 3 7 Subleaders: Ward 7 Party Elective ll Appointive _5 Total Subleaders 23 22 Inactives Inactives 18 21 The protocals were scored by Browning alone for the needs for power (3 Pow), achievement (E_Ach) and affiliation (n Aff) according to the scoring manuals developed by McClelland, Atkinson, Clark and Lowell (n Ach), Veroff (anow) and Heyns, Veroff and Atkinson (n Aff). These scoring manuals are given in Atkinson (1958). In addition to these scoring conventions, others were adopted for the three motives as noted in Browning's dissertation. E Ach may be described as the concern over compe- tition with a standard of excellence. E Aff may be de- fined as the concern over establishing, maintaining or restoring a positive affective relationship with another person. §_Pow refers to affect concerned with the con- trol of the means of influencing a person. This includes wanting to avoid weakness. Browningfs Results Browning concludes from his non-parametric analy- ses that there are no differences attributable to the Civic vs. Party dimension; both groups are similar in motive scores. Leaders and subleaders are similar in n Pow and n_Ach, differing only in n Aff: Leaders having less need for warm personal relationships than subleaders. The inactives are similar to the matched subleaders in n Aff, but have lower n Pow and n.Ach. From a further division of the Party subleaders into elective offices (11), ward offices (7) and appointive offices (5) Browning concludes that the ward office- holders are essentially identical to the inactives in motives (low E_Ach, low E Pow, and high E_Aff). The elec— tive politicians have higher n Pow and E_Ach than the ward officeholders, but have similar high n.Aff. The elective group has higher a Pow and n Ach than its matched inac- tives. Browning finds that most of the ward politicians had a relative active in politics and were recruited by the Party for their position. Only four of the eleven elective politicians had relatives active in politics and five out of nine1 self-initiated their entrance into politics. Purpose of the Present Study The present study began as an extension of Brown- ing's study. The original data were to be reanalyzed in terms of the need for influence (3 Inf), a motive scoring system recently developed by Uleman (1966). E Inf may be described as a need for some kind of interpersonal effec- tiveness. This scoring system is thought to measure a motivational complex separate from n Pow. It was hoped 1No Information is available on two men. that analysis in terms of E Inf would aid in defining the boundaries of the two motive scoring systems. However as the Browning study was more carefully investigated, questionable procedures in both collection and analysis of the data were discovered. Since the original interviews and TAT protocals were made avail- able, it was decided to reanalyze the data revising the methodology where necessary (as described below) and then to study the effect of the new variable, a Inf. One of the questionable procedures was the scor- ing of all the TAT protocals by one person. Each person tends to have scoring preferences and idiosyncrasies; this tends to be balanced out when there are more than one scorer. ”The personal preferences of the one scorer in this study may have influenced the direction of the scores and therefore the results. There is a second possibility that the scoring was biased unintentionally since the person who scored the protocals had also col- lected the data and may have remembered the status of the respondents (See Atkinson, 1958, p. 691). The major scoring systems used were those in Atkinson. However additional conventions were adOpted for the three motives (described in the appendix in Browning's dissertation) which make comparisons with other studies using only one system difficult. ‘wAnother factor which may have influenced results was the significant correlations between length of the TAT protocal and each motive score, shown in Table 2. Since the conditions under which the data were gathered were not uniform, there is the possibility that the story length was in part due to factors other than motive strength. Assiories become longer and more detailed, there is a greater possibility that subcategories of the scoring system would be included and thereby raise the motive score for that story. Therefore, higher scores might not be a true reflection of the strength of a motive, but may be the result of uncontrolled factors in the ex— perimental design. Various methods have been used to correct this problem2 (Walker and Atkinson, 1952; Child, Storm and Veroff, both in Atkinson, 1958; Veroff, Atkin- son, Feld and Gurin, 1960), but no corrections were made in this analysis of Browning's data. (These corrections based on Browning's scoring were made in later journal articles.) It is possible that some of the Browning re- sults may be attributable to differences in story length alone. The present study attempted to correct the above problems in the analysis of the data and to investigate the effect of n Inf. 2All experimenters do not agree that the story length correlation is a problem. See the Discussion section. ‘Table 2.--Spearman rank order correlations between story 1' length and Browning's Motive Scores. Motive Correlation ‘ N Pow .27** N Ach .43*** N Aff .l9* *p < .10, two-tailed. **p < .01, two-tailed. - ***p < .001, two-tailed. The analyses were guided by the following ques- tions: (1) Is there a difference in the strength of motives held by those in this sample active in civic af- fairs and those active in a political party? (2) Are the leaders and subleaders similar in the strength of any of the four motives, n Pow, E_Inf, n Ach and 2_Aff? (3) Is there a motive combination which dif- ferentiates the inactives from those who become active in these two areas? In addition to tests related to the above ques- tions, tests were done to verify Browning's findings on the subgroups and to investigate the possible relationships 10 of other variables (e.g., level of aspiration) with the motive scores. These analyses are considered secondary in importance to the goals of this study and are there- fore presented in Appendix B. Method Subjects The subjects are 94 businessmen in an Eastern city described as leaders, subleaders or inactives in the poli- tical and civic affairs of the community. A more complete description is given in the Introduction. Procedures The original interviews and TAT protocals were generously made available for the present study by Pro- fessor Browning. The protocals were scored for n Pow and n Inf according to n_Pow manual in Atkinson (1958) and n Inf manual developed by Uleman (1966). No other scoring conventions were adopted. N Pow was scored by two, and n Inf was scored by three different paid undergraduate experts who had com— pleted training on the respective manuals and had obtained acceptable reliabilities (see Table 3). Each expert scored all protocals for one motive, without knowledge of the purpose of the study or of the status of the respondent. The Spearman rank order correlations between the three a Inf scorers (FP, BG, LA) were .54, .58 and .77. The reliability of their average score as determined by 11 Table 3.--Re1iabi1ity scores for the five undergraduate raters on the practice sets in Atkinson (1958): Rank order correlations between expert and scoring and percentage agreement (%) raters' between rater's and expert scores on the motive imagery category.a N Ach Practice N Aff Practice Sets Sets C D A B C D N Inf Scorers FP rS .94 .94 .87 .92 .92 .96 % 97 100 96 100 100 98 BS rS .94 .85 .59 .77 .77 .81 % 100 94 96 98 98 98 LA rS .86 .89 .84 .80 .79 .83 % 97 97 98 100 96 100 N Pow Practice Sets C D E F G N Pow Scorers LM rs .80 .69 .78 .66 .77 % 89 88 89 85 90 JB rS .59 .57 .21 .72 .61 % 74 78 56 77 80 aPercentage agreement scores equal the ratio of twice the number of times that the expert and rater agree in scoring the presence of motive imagery divided by the number of times the rater has scored motive imagery pres- ent plus the number of times the expert has scored motive imagery present. (Atkinson, p. 688) 13 the Spearman Brown prediction formula (Winer, p. 127) was .84, acceptable for research purposes. The correla- tion between the two n Pow scorers (LM, JB) was rho = .71. Since the correlations between JB and Browning's scoring (rho = .71) and between LM and Browning's scoring (rho = .76) were almost identical, all three scorers were used in the present study. The reliability of their average score was .89 as determined by the above prediction formula. Expert scorers were not available for n_Aff and n_Ach. Therefore, Browning's scoring of these two motives was used with the recognition of its possible research limitations (as described in the Introduction). Story Length Spearman rank order correlations between the motive scores and story length of the TAT protocals (which ranged from 128 to 1,678 words) were significant as shown in Table 4. Consideration was given to the ef— fect story length itself might have had on the motive scores. As a story becomes longer and more detailed, there is a greater probability that a motive scoring sub- category would be applicable and raise the motive score. Since the data had been gathered under non-uniform con- ditions (see Atkinson, p. 837) it was possible that these conditions resulted in the varying story lengths and l4 consequently the higher motive scores. Could results later obtained from this data be an artifact of the length of protocal of these subjects or would results be true motive differences of the sample? Table 4.--Spearman rank order correlations between story length and Szari Uncorrected and Corrected Motive Scores. Correlations Motive Uncorrected Scores Corrected Scores N Pow .37** .12 N Ach .43** -.01 N Aff .19* -.16 N Inf .67** .10 *p < .10, two-tailed. **p < .001, two-tailed. To check if story length itself could differen- tiate between the groups, 2 x 2 analyses of variance (using the least squares solution with proportional cell sizes, see Winer, (1962) section 7.8) were run with the following two designs: (a) Civic vs. Party, Active vs. Inactive; (b) Civic vs. Party, Subleaders vs. Inactives. Story length was used as the dependent variable in both analyses. Both analyses indicated a significant main effect (p < .05) for Civic vs. Party with the Civic group 15 having longer stories than the Party group (see Table 5 for the cell means). There were no other significant ef- fects or interactions. The cell means, F-values and probability levels of these analyses with story length are given in Appendix C. To see if the results of the uncorrected motive scores would be similar to the results obtained from the story length analyses of variance, the same analyses were computed with the four motives as the dependent variables. The Civic vs. Party factor was significant for n_Ach (p < .05) and n Inf (p < .10) in both analyses and for n_Pow in the first (a) analysis (p < .05). In all these cases, the Civic group had higher motive scores than the Party group, the same direction as in the story length results. It was obvious that for these three motives n Ach, n_Inf and anow, the results could just as easily be explained by the story length factor as by the true motive differences of the sample. Cell means are given in Table 5 and the F-values and probability levels are given in Appendix C. Because of the above preliminary results it was deemed advisable to correct the motive scores for story length. Correction for Story Length Linear corrections determined by the regression lines of the four motive scores on story length were 16 Table 5.--Cell means of the motive scores found significant in the two analyses of variance using story length and uncorrected motive scores as dependent variables. Dependent Variable Civic Mean Party Mean The Active vs. Inactive Analysis of Variance Results ’ Story Length 550.0 446.9 N Ach 9.1 6.6 N Inf 17.6 15.5 N Pow 6.5 5.2 The Subleaders vs. Inactive Analysis of Variance Results Story Length 562.6 438.9 N Ach . 8.7 6.5 N Inf 17.7 15.5 17 performed on n Ach and n Aff to the nearest integer and on n Pow and n Inf to the nearest third. The means of the original scores were preserved in these calculations. This procedure successfully eliminated the significant correlations between motive scores and story length (Table 4) and these corrected scores were used in all succeeding data analyses. One consideration concerning the story length corrections was its effect on the variance held in common by pairs of motives. If in eliminating variance attribu- table to story length, the variance common to the motives was increased considerably, we would be measuring in ef- fect only one motivational state rather than four differ- ent motivational states. One way to determine this is by looking at the variance held in common between pairs of motives. The correlations are shown in Table 6. The coefficients of determination, derived from the correlations, indicate the amount of variance accounted for by knowing one of the motives. The coefficients are given in Table 7. Looking at the variance held by each pair of motives, un- corrected and corrected, it is evident that the variance in common was reduced in all pairs excepting those with n Aff. In each pair of motives with n_Aff, the variance in common increased in the corrected scores, although the greatest of these, 10%, is still a negligible amount. 18 Table 6.--Spearman rank order correlations among uncorrected Szari motive scores and correlations among cor- rected Szari motive scores. N Pow N Inf N Ach Uncorrected Scores A N Inf .55*** N Ach .28*** .49*** N Aff -.22* -.Ol -.01 Corrected Scores N Inf .47*** N Ach .16 .32*** N Aff -.31*** -.24** -.18* *p < .10, two-tailed. **p < .05, two-tailed. ***p < .01, two-tailed. 19 Table 7.--The coefficients of determination between pairs of motives indicating the amount of variance accounted for by knowing one of the motives. N Pow N Inf N Ach Uncorrected Scores N Inf .30 N Ach .08 .24 N Aff .05 .00 .00 Corrected Scores N Inf .22 N Ach .03 .10 N Aff' .10 .06 .03 20 In summary, the variance in common between these motives n Pow, n Inf and n Ach, decreased when the cor- rection factor was applied to the motive scores. The variance in common with these motives and n_Aff increased with the corrected scores; but is still low, 10%. It can be stated with confidence then, that since the greatest amount of variance in common to the pairs of motives is 22%, four separate motivational states are being measured. Considerations in Analyses of Variance Unequal variances and small unequal sample sizes of the six groups (Party leaders, subleaders, inactives and Civic leaders, subleaders and inactives) made data analysis by the method of analysis of variance difficult. With these six groups only n_Aff motive scores did not satisfy the homogeneity of variance assumption as deter- mined by the Cochrane and F-Max tests (Winer, section 3.11). If the Party leader and Civic leader groups are entirely dropped or if they are combined with the Party subleaders and Civic subleaders, all four motive scores satisfy the variance assumption computed by the Cochrane and F-Max tests. It would appear that the Party leader and Civic leader groups have the largest variance, an artifact of the small numbers in each sample. It is obvious that the only case in which we should suspect the results of an 21 analysis of variance is if the Party leaders and Civic lead— ers are included in the analysis as intact groups. As a result of the unequal sizes of the groups and the original design of the study, an analysis of variance with a least squares solution and proportional cell sizes was used (Winer, section 7.8). In order to obtain prOpor- tional cell sizes, it was necessary to eliminate some subjects from the analyses. This elimination procedure was carried out by a computer which directed subjects into the prOper cells until the pre-assigned cell totals were reached. The remaining subjects were noted, but dropped from that particular analysis. Data cards were entered according to subject numbers which had been randomly as- signed by Browning.3 No more than eight eligible subjects were dropped in the tests related to the hypotheses. Data Analyses Used The guiding questions of this study were investi- gated by means of analyses of variance and non-parametric tests with the four motives as the dependent variables in both cases. The analyses of variance used are described below and shown in Table 8. Three designs were used: (a) 2 x 3 design using all six groups, n = 86, (b) 2 x 2 3An alternative method for data analysis is the least squares solution with unequal cell sizes (Winer, section 5.23). 22 Table 8.--Analyses of variance designs used in this study, with the four motives as dependent variables. Party Civic gDesign I: Six Groups Leaders 3 3 Subleaders 22 22 .Inactives 18 18 Design II: Four Groups Subleaders 22 22 Inactives 18 18 N = 86, 8 subjects dropped (4 Civic Leaders, 1 Party Sub- leader and 3 Civic Inactives) The 3 Civic Leaders were chosen to represent the typical Civic Leader on the following cri- teria: age, education, salary, job and party affiliation. N = 80, 4 subjects dropped (1 Party Subleader, 3 Civic ’ Inactives) Design III: Leaders and Subleaders Combined into Actives Actives 26 26 Inactives 18 18 N = 88, 6 subjects dropped (3 Civic Subleaders, 3 Civic Inactives) In this analysis, the 3 Party Leaders' and 7 Civic Leaders' data cards were inserted in the computer first so that these small groups would not be eliminated. 23 design with four groups, n = 80, and (c) 2 x 2 design combining the leaders and subleaders into the Active group, n = 88. With these many analyses of variance there are 36 separate F-tests involved (3 designs x 4 motives as de- pendent variables in each design x 3 F-tests (A, B, AB) in each analysis). By chance it can be expected that 3.6 or 4 results will be significant at the 10% level. This should be kept in mind as the results are reported. In the above analyses of variance it was not pos- sible to adequately compare the Civic leaders and Party leaders with each other or with the other groups because of the necessity to eliminate over half of the Civic leader group to attain prOportional cell sizes. Therefore in order to include all the leaders in the comparisons, the non-parametric Mann-Whitney U Test (MWU) was used to com- pare Civic leaders vs. Party leaders, Civic leaders vs. Civic subleaders and Party leaders vs. Party subleaders. With 12 non-parametric tests (3 comparisons x 4 motives) it can be expected that there will be at least 1.2 or 1 significant result at the 10% level due to chance alone. Additional Analyses Exploratory analyses were computed which are not related to any stated hypotheses. Descriptions and re- sults of these analyses are given in Appendix B. RESULTS Analyses of Variance In the three analysis of variance designs (Table 8), there are no significant (p < .10) main effects or interactions. The cell means, F-values, degrees of free- dom and probability levels are given in Table 9. Nonjparametric Tests With 12 separate analyses, one could expect 1.2 results to be significant at the 10% level. In fact, three comparisons are significant, more than double the expected number. In this case there is more confidence that these three results are not random and are true dif- ferences of the sample. The Civic leaders have higher n Pow (p = .03, two-tailed) and lower 2 Aff (p = .10, two— tailed) than the Civic subleaders. The Civic leaders, when compared with the Party leaders, have higher n_Pow (p = .06, two-tailed). Means and probability levels for all comparisons are given in Table 10. 24 25 mm sauce As.m mm.m mm.ha me.w ms oa>flo m~.m mm.s. mo.ma ms.m ms spumm AH.m m~.p me.ma mo.m om m>HuomaH HH.m mo.m wm.>a wo.o es umemmansm AH.H om.m mo.sa ~H.A m “memoq mHMGHmHME mo.m om.» mv.ma m~.m ma m>fluomcfl oH>Ho na.m mm.m mm.na oa.m mm umemwansm 0fl>flo oo.H oo.HH mm.ma so.m m umemma ofl>flo m~.m so.» mh.ma ms.m ma m>fluomcfl sauna mv.m om.» mn.ma ma.m mm “mammansm sauna mm.H oo.m hm.ma am.v m “momma shuns mammz Hamo Lee 2 now 2 mcH z 30m z 2 mm we annoy om HOHHM who.o ~m~.o mmm.o m~.v mha.m m ma m~.v mnm.a omb.o mmm.o eme.o m .uomcH .nsm .emmq . a mam.o mao.H mmm.o HaH.H H on>flo .muumm . m m msam>um m msam>um ‘ m moam>um m msam>um mp condom sea 2 nos 2 mcH z 30m z mmbomw XHm mBHB mmqu¢2¢ uH ZUHme .m mabme ca meHHommU mum mmmwawsm amuse .cm>wm mum Amy mambma mafiaflbmnoum can mosam>|m .Ampv Eocwoum mo momummp .msmmfi Hamo .mmHQMHHm> ucmpsmmmc mm mm>apoa omuomunoo HSOM map spas mosmwnm> mo mmmwamsdnl.m magma 26 mo.m om.> m¢.ma m~.o ma cams m>fluomcH n 0H>Ho >>.~ mm.m mm.>a oa.o mm cam: m>auo¢ u 0H>Ho m~.m mo.> mn.ma mn.m ma cams m>fluoqu u munmm Hm.m om.» mm.ma en.m mm cams m>fluo¢ u muumm mammz Hawo mm« 2 ace 2 mcH z 30m z 2 hr up Hmuoa m h HOHHm nvm.o mva.o Hmm.o mmo.o a ms omo.o eme.o men.o mao.o H m>fluomcHnm>fluo¢ u m m~.v Ham.a omn.o mm~.o mms.o H 0H>Hu .suumm I 4 m msam>um m wsam>um wsam>um m msam>um MU OOHDOm mma z and z mcH z 30m z Aoum .HIQmV quo mmmoamqmsm "manomw maom meHz mmquHu062H 0H>HU n>.~ mm.m mm.>a oa.m mm mumpmmansm oa>wo oo.a mm.oa Hm.hH vm.m h mumcmoa 0H>Ho mm.m 50.5 ms.ma os.m ma mm>fluomafl sunmm vs.m am.» am.mH om.m mm mumwmmansm suumm mm.H oo.m am.ma hm.s m mnmemma shuns mmsouw pomusH "Umpsaosfl msofluw>ummbo commoup auflz names mma 2 sum 2 qu z 30m 2 up U05CHHQOUII o m magma 27 mo.m om.s mv.oH a~.m mH m>HpomcH . oH>Ho mo.m ~4.m mm.hH ma.o om m>Huo¢ . oH>Ho m~.m mo.a ms.mH mn.m mH w>HuomcH u spumm m¢.m No.5 ms.oH mm.m mm m>Huo¢ u mpumm msmoz HHmU HHH 2 nos 2 HaH z 30m z z rm He Hmuoa vm .HOHHM mHo.H avm.o «mm.o mam H ma Hus.o Hmm.H Hon.o mom H m>HuomcHnm>Huo¢ u m m~.v Amm.~ on.H «mm.o Has H 0H>Houspumm u 4 m msHm>um msHm>um msHm>nm msHm>um up condom HHH z noH z Lea z 30m z mm>Hsom oazH omszzoo mmeHMHmom oz¢ mmmaamH BAH: mmmeHZH "HHH szmma o~.m mp.s mo.AH mm.m bm Hmuoa AH.m mm.“ ~m.mH mo.o mm m>HuomaH sm.m ao.m ov.AH sm.m we m>Huo¢ om.~ mH.m AN.AH mH.m ow oH>Ho mo.m om.> sm.mH ma.m as muumm mamswmumz use 2 :04 z _._s qu z 30m z z umscHuaoonn.m mHhma 28 mm.m m>.> mo.>H m¢.m bm Hmuoe 5H.m mm.h mm.wH mo.m mm m>anomcH vm.m mo.m ow.ba «m.m we m>auod om.m mH.m hm.>H mH.m ow 0H>HU mm.m om.> vm.wH mn.m ow muumm mamsfimumz mmfi z :04 z . “SH 2 30m z z possHuGOU|l.m manna 29 Table 10.—-Pairwise comparisons of party and civic leaders and subleaders. Leaders vs. Subleaders Mann-Whitney U-Test, U-Value, p—Value, (Two-Tailed) and Means Party Civic 3 Leaders, 23 Subleaders 7 Leaders, 22 Subleaders Motives U p Means U p Means PL P-SL CL C-SL n_Pow' 28 .60 4.57 5.79 35.5 ‘221 8.54 6.11 g Inf 28.5 .63 15.57 16.89 71.5 .78 17.51 17.93 3.Ach 30 .72 8.00 7.57 57.5 .32 10.29 8.68 n_Aff 20 .24 1.33 3.74 44.5 £323: 1.00 2.77 3 Party Leaders vs. 7 Civic Leaders Mann-Whitney U-Test, U-Value, p-Value, (Two-Tailed) and Means Motives U p Means Party Civic n Pow 2 1222 4.57 8.54 E_Inf 4.5 .23 15.57 17.51 E Ach 9.5 .90 8.00 10.29 n_Aff 8 .67 1.33 1.00 DISCUSSION There were only three significant results in the analysis of this data, all involving the Civic Leader group: The Civic Leaders had higher anow than both the Civic Subleaders and Partereaders and the Civic Leaders had lower n_Aff than the Civic Subleaders. It would ap- pear that the Civic Leaders are unique in this sample, while the other groups are similar. It is surprising that the motives do not differ- entiate the Party sample or the Civic Subleaders and Civic Inactives according to levels of activity. It is possible that in fact there is no difference between these men, but there is another explanation which may be pertinent. The implicit assumption underlying this study is that the motives of a person determine in part the level of activity he will attain and the area of this activity. This assumption finds little support in these results: considering the similarity of motives of five groups (ex- cluding the Civic Leaders) it is obvious that the motives of a person in this sample do not make a substantial dif- ference in.the area or level of activity. 30 C) l‘f} y La AL. V». by V‘. 31 An alternative to the view that the environment is a function of motives is the view that motives are a function of the environment. It must be remembered that all the men in this sample are businessmen-managers. If the "environment of a businessman" is a prime factor in the development of their motives and if this businessman factor overrides the effects of other influences, there should be no differences found in this sample. There are no differences found among five of the six groups, exclud- ing the Civic Leader group. A second possible explanation for the lack of dif- ferences between the five groups is that only men with a certain similar motive structure will enter the business field and become businessmen and managers in the first place. Both of these explanations have as their assump- tion that the effects of activity and differing areas of activity (Party or Civic) will not be major determinants in the development of motives, but will be secondary to other influences, either the "businessmen environment" or the influences which were responsible for the original development of the motive structure. If the four motives do not distinguish active from inactive there still remains the question "why do some men remain inactive while others are involved in party or civic affairs?" In this sample the answer may be simple-—recruitment is a major factor in determining 32 activity. All the Civic Subleaders were recruited by the mayor for their position. One-half of the Party Sub- leaders were recruited or appointed to work for the Party. For these men, it is an external event rather than their motivations which lead them into activity. Only one group appears to differ in motivation from all others--the Civic Leaders. In comparison with both the Party Leaders and Civic Subleaders these Civic Leaders had higher n_Pow scores. A plausible explanation for this effect is that a Civic Leader has reached an elevated position in which demonstrations of anow strivings do not carry the same strictures and harmful repercussions that they might bring to a person in a sub- ordinate position. A Civic Leader may be praised for his strong, forceful leadership and effective management. A Party Leader in an equally elevated position is not in an equally impregnable position: he must be reelected peri- odically by constituents who can discipline him effectively for obvious n Pow strivings. Additional Comments Taking into consideration the common businessman background, it is still surprising that there were not more differences detected between groups which at face value appear to have differing positions of responsibility. Although a few differences were found, there are a few 33 methodological considerations in both data collection and data analysis which may have confounded the results. Data Collection One possibility why so few results were found may be the addition of unwanted variance from the nonstand- ardized data collection procedures. Some of the men had been in a previous extensive political science study and although it is not expected that they would be uncoopera- tive with Dr. Browning, perhaps this previous experience did affect their attitudes toward subsequent political science studies. The men were interviewed in various places, homes, private offices or public places of business. The time of the day varied and it was, of course, impossible to standardize the events preceding the interview. Phone calls, secretaries and customers disturbed the progress of the interviews. The length of interviews varied con- siderably (40 minutes to 3-1/2 hours) and the interview guide was not always followed. While it is understand- able that many of the preceding occurrences were not under the experimenter's control, they may have introduced con— founding variance into the motive scores. Data Analysis There are also some procedures in data analysis that may have resulted in the lack of more results. One 34 of these may have been the story length correction. Two arguments have been presented against correction. The first argument states that if a person has a strong motive, he will see more of those motive cues in TAT pictures and tell longer stories to include all the cues. The second argument refers to the action of telling the story as a manifestation of a motive. That is, those who told longer stories may have been using the situation to influence the investigator. And those who did not wish to influence the investigator, but merely wanted to cooperate, would not find it necessary to tell long stories. This would be reflected in story length correlations, the n Inf correla- tion with story length was .67, the largest of the corre- lations, and the n Aff correlation was .19, the smallest correlation. Both arguments would agree that correcting for story length would eliminate useful variance directly re- lated to the motivational states of the persons. There are valid arguments on both sides of the correction issue, and the decision becomes the preference of the investiga- tor. The reasons justifying the correction factor in this study have already been presented in the Methods Section, but will be briefly summarized here. Atkinson suggests looking at the story length correlation when the data has been collected under non-standardized conditions (p. 837), as described in the previous section. The results of the 35 analyses of variance of these data with uncorrected scores were similar to those received when the story length was used as the dependent variable. One point worth noting is that the significant direction in the analyses with uncorrected scores (Civic is greater than Party on n Pow, n Inf, and n_Ach) is retained in the analyses with cor- rected scores, although the differences are no longer significant. Although the story length correction factor may have eliminated useful variance, this investigator feels that the correction was necessary with these data. A second factor already mentioned is the possi- bility of increased variance in common between the four motive scores caused by the story length correction. If this variance in common was increased greatly between the four motives, the scores could be strictly viewed as measur- ing only one motivational state. However as discussed in the Methods Section, the correlations and coefficients of determination indicate that this is not the case--the four scores do represent separate states since the variance in common between motives is low. Thus it can be reported with a greater degree of assurance that this sample was measured on four variables, and they did not differ in the four motive scores when they were corrected for story length. 36 Browning-Szari Differences in Conclusions The Browning data were the bases for the Szari corrected data and therefore it could be expected that the conclusions of the two analyses would be similar. This is true in the majority of cases but there are five comparisons upon which the two studies disagree on the conclusions. However these discrepancies can be resolved when the analyses are compared (Table 11). One of the five contradictory Browning conclusions (Leaders vs. Sub- leaders, E_Pow) is not based on any formal statistical analyses, but on visual inspection of tables. In three other comparisons (Leaders vs. Subleaders, n_Aff; Sub- leaders vs. Inactives, 2_Pow; and Subleaders vs. Inactives, B.Ach) a liberal interpretation of significance levels re- sultszhldiffering conclusions. The fifth contradictory conclusion (Civic Leaders vs. Party Leaders, n Pow) is based on the same non—parametric test, the Mann-Whitney U Test. However the most reasonable explanation for the differing results takes into account the scores used in the Szari analyses. These scores were the average of three raters scores (including Browning) and they were corrected for story length. The scores on which the tests were computed were slightly different, producing different results. 37 Table ll.--Browning and Szari conclusions compared. Comparisons Browning Conclusions Szari Conclusions Leader— Civic Leader = Party Civic Leader = Party Leader Leader Leader n_Ach n Ach n Aff E Pow E Inf n Aff* Civic Leader > Party Leader — n_Pow Leader— Leader = Subleader Party Leader = Party Subleader Subleader n Ach* all four motives E Pow* Civic Leader = Civic Subleader Leader < Subleader n Ach n Affa H Inf —' Civic Leader ; Civic Subleader n Pow Civic Leader Z Civic Subleader n_Aff Subleader- Subleader = Inactive Subleader = Inactives Inactive n Aff* all four motives Subleader > Inactive n’Powb n Achc *No tests indicated in Browning's dissertation. aBrowning's test results for this comparison, n_Aff. Party Leaders vs. Party Subleaders, ‘ p = .12, one-tailed, MWU Civic Leaders vs. Civic Subleaders p = .05, one-tailed, MWU bBrowning's test results for this comparison, Pow. Party Subleaders vs. Party Inactives — p = .12, one-tailed, Wilcoxon Civic Subleaders vs. Civic Inactives p = .10, one-tailed, Wilcoxon CBrowning's test results for this comparison, Ach. Party Subleaders vs. Party Inactives _ p = .24, one-tailed, Wilcoxon Civic Subleaders vs. Civic Inactives p = .07, one-tailed, Wilcoxon Legend: = is similar to in the following motive > has higher scofEs than __ in the following motive < has lower scores than __ in the following motive 38 Brief Summary There are no differences in the four motives studied, n_Pow, n Ach, g_Inf, and E_Aff, between five groups of men (Party Leaders, Party Subleaders and Party Inactives, Civic Subleaders, and Civic Inactives). The Civic Leaders appear to be composed of men who differ from the other men by having higher n_Pow and lower n_Aff. The fact that all are businessmen may explain the homo- geneity of motives of the five groups; the relatively powerful protected position of Civic Leaders may explain their openness in demonstrating n Pow strivings in the TAT with higher a Pow scores. Activity in this sample may simply be the result of a recruitment process by the Par- ties or Mayor. Methodological procedures which may have eliminated useful variance and arguments opposed to the story length correction were considered. Differences between Browning's conclusions and Szari's conclusions were resolved when the statistical bases were examined. LIST OF REFERENCES REFERENCES Atkinson, J. W., (Ed.) Motives in fantagy, action and society: A method of assessment and study, Princeton: D. Van Nostrand) 1958. Bell, R., Determinants of psychological involvement in politics: A causal analysis. Dissertation Ab- stracts, 1967, 27(10-A), 3493. Browning, R., Businessmen in politics: Motivation and circumstance in the rise to power. Unpublished doctoral dissertation, Yale University, 1960. Browning, R. and Jacob, H., Power motivation and the political personality. Public Opinion Quarterly, 1964, 28, 75-90. Child, 1., Storm, T., and Veroff, J., Achievement themes in folk tales related to socialization practice, In J. Atkinson (Ed.) Motives in fantasy, action and society, Princeton: D. Van Nostrand, 1958. .Dahl, R., Who governs? Democracy and power in an Ameri- can city. New Haven: Yale University Press, 1961. Heyns, R., Veroff, J., and Atkinson, J., A scoring manual for the affiliation motive. In J. Atkinson (Ed.) Motives in fantasy, action and society, Princeton: D. Van Nostrand, 19582 Lane, R., Political life: Whypeople get involved in politics. Glencoe, 111.: Free Press, 1959. ‘ Lasswell, H. D., Power and_personality, (lst. ed.) New York, W. Norton, 1948. Lindenfeld, F., An analysis of political involvement, Dissertation Abstracts, 1962, 22(12), 4426-4427. McClelland, D., Atkinson, J., Clark, R., and Lowell, E., A scoring manual for the achievement motive. In J. Atkinson (Ed.) Motives in fantasy, action and society, Princeton: D. Van Nostrand, 1958. 39 40 Milbrath, L. W., Political participation: How and why do people getginvolved in politics? Chicago: Rand McNally,*I965. Sampson, R. V., The psychology of power, New York, Vintage Books, 1968, c. 1965. Skolnick, A., Motivational imagery and behavior over 20 years. Journal of Consulting Psychology, 1966, 30(6), 463-478. Uleman, J., A new TAT measure of the need for power, Unpublished doctoral dissertation, Harvard, 1966. Veroff, J., Development and validation of a projective measure of power motivation. The Journal of Ab- normal and Social Ps cholo , 1957, 54, 1-8. Re- printediin J. Atkinson (Ed.) Motives in fantasy, action and society, Princeton: D. Van Nostrand, 1958. Veroff, J., A scoring manual for the power motive, In J. Atkinson (Ed.) Motives in fantasy, action and society, Princeton: D. Van Nostrand, I958. Veroff, J., Atkinson, J., Feld, S., and Gurin, G., The use of thematic apperception to assess motivation in a nationwide interview study, Ps cholo ical Monographs: General and Applied, 1960, 7%, 12, (Whole no. 499). Walker, E., and Atkinson, J., The expression of fear re- lated motivation in thematic apperception as a function of proximity to an atomic explosion. In J. Atkinson (Ed.) Motives in fantasy, action and society, Princeton: D. Van Nostrand, 19582 Winer, B. J., Statistical principles in experimental design, New York: McGraw-Hill Book Company, 1962. APPENDIX A COPIES OF THE SIX TAT PICTURES USED IN THE BROWNING STUDY. Just look It the picture briefly (lo-15 seconds), tm'n tho pup and write out the story it suggests. 42 ‘ _-", 4V V . 'u.‘ ‘ ' a "J" t 'J Just look at the picture briefly (lo-15 seconds), turn the page and write out the story it suggests. 43 Just look at the picture briefly (lo-lb seconds), turn the page and write out the story it suggests. 44 Just look at the picture briefly (IO-15 seconds), turn the page and write out the story it suggests. 0v 45 Just look at the picture briefly (IQ-15 seconds), turn the page and write out the story it suggests. 46 ‘ “f. I. LTY.CF . is." ’ fig ‘k'v u. ‘..‘(q- r 1. Just look at the picture briefly (IO-35 seconds), turn the page and write out the story it suggests. APPENDIX B APPENDIX B EXPLORATORY ANALYSES AND RESULTS Additional analyses-not related to the guiding questions of this study were done by both analyses of variance and non-parametric tests. Possible differences in constellations of motives between Actives and Inactives were investigated by using an analysis of variance design with one motive as a dependent variable and a second motive as an independent variable. Also the subgroups in the Party sample and the matched subjects were compared to each other and to the inactive sample by Mann-Whitney U tests and Wilcoxon Matched Signed Ranks Tests. Other vari- ables related to the Party Sample only were also investi- gated. Description of the tests performed and results are reported in the following sections. Analyses of Variance In these analyses, the motive scores were used as both dependent and independent variables. The distribution of one motive was divided as close to the median as pos- sible to separate high and low scores. This division was used as an independent variable in addition to Party vs. 47 48 Civic and Active vs. Inactive in a 2 x 2 x 2 design. Each of the other three motives were then used as dependent variables. In an attempt to refine the results obtained from the 2 x 2 x 2 analyses (given below) the same design was used with three motives as independent variables divided into high, medium and low scores rather than merely high and low. This analysis could not be done with the E_Aff scores as an independent variable because of the impossi- bility of achieving proportional columns in the design. (See Table 12.) With this many analyses of variance, there are 147 separate F-tests computed (7 analyses x 3 motives x 7 F- tests (A, B, C, AB, AC, BC, ABC) in each analysis). It can be expected that 15 results will be significant at the 10% level by chance alone. This must be kept in mind as the results are reported below. Other Tests The Mann-Whitney U Test and the Wilcoxon Matched Signed Ranks Test were used to compare subgroups which were not compared in the analysis of variance and to ex- plore the relationships of motives to three additional variables in the Party sample. Party subgroups compared by the Mann-Whitney U Test were the Elective (11), Ward (7), and Appointive (5) 49 Table 12.--Summary of the analyses of variance used in the exploratory phase of this study, including de- sign, the number of analyses, the factors and dependent variables. Factors Design No. A B C Dependent Variables 2x2x2 3 P-C Ac-I High-Low E Pow Three other motives 2x2x2 3 P-C Ac-I High-Low g Inf Three other motives 2x2x2 3 P-C Ac-I High-Low g Ach Three other motives 2x2x2 3 P-C Ac-I High-Low 2_Aff Three other motives 2x2x3 9 P-C Ac-I High-Medium- Three other motives Low for each motive except 2 Aff. P = Party Civic 0 II Ac = Active I = Inactive 50 Subleaders. The following groups of Subleaders were com- pared with their matched inactives on the four motives by the Wilcoxon: Civic Subleaders (21), Party Subleaders (18), Electives (10), Ward (4) and Appointive (4). The three variables used as independent variables in the analysis of the Party sample alone by the Mann- Whitney U were: (1) Mode of Recruitment (self-initiated vs. recruited), (2) Power potential of the office (high vs. low) and (3) Father active vs. inactive in party poli- tics. A fourth variable, level of aspiration was tested by means of two one-way analyses of variance. This vari- able consists of three 1evels--no aspiration, indefinite aSpirations and definite aspirations. One analysis used all three levels of aspiration (none, indefinite, definite) as the factor. The second combined the none and indefinite categories into one level and used the definite aspiration category as the second level. A summary of these analyses is given in Table 13. These comparisons represent 52 separate tests (13 comparisons x 4 motives as dependent variables in each comparison). By chance alone we can expect 5 results to be significant at the 10% level. This expected number should be kept in mind as these results are reported in a later section. 51 Table 13.--Summary of the tests used in the exploratoryv phase of this study. The dependent variables in all analyses were the four motives, n Pow, g Inf, 2_Ach, and 2_Aff. — Comparisons Ns Mann-Whitney U Tests Elective vs. Ward 11-7 Elective vs. Appointive 11—5 Appointive vs. Ward 5-7 Self-initiated vs. Recruited 8-16 Office with High-Low Power Potential 10-13 Father Active-Inactive in Party Politics 10-34 Wilcoxon Matched Signed Ranks Test Civic Subleader vs. Inactive 21—21 Party Subleader vs. Inactive 18-18 Elective vs. Inactive 10-10 Ward vs. Inactive 4-4 Appointive vs. Inactive 4-4 Simple Analyses of Variance 3 level No aspirations-indefinite 25_10_7 aspirations-definite aspirations 2 level Indefinite and no aspirations 35_7 combined-definite aspirations 52 Results The analyses of variance are presented in Table 14 according to the motives used as the dependent vari- able. The significant results are underlined, and means of the groups shown. It can be seen that the majority of the signifi- cant results occurred when E_Pow and 2_Aff were the de- pendent variables, 11 and 14 results respectively. There are 32 results significant at the 10% level in these analyses, more than twice the number expected (15) if the results were random events. However it must be realized that 32 significant results does not indicate 32 separate main effects or in- teractions. In the design of the analyses, with the de- pendent and independent variables reversed in various analyses, similar interactions can be discovered in three or four analyses (Independent variable X at two- and three-levels (except §_Aff) with dependent variable Y; and Independent variable Y at two- and three-levels with dependent variable X). The same main effect can be dis- covered in five or six analyses in which a motive is a dependent variable: if 2_Aff is the dependent variable, there are six such analyses: the other three motives as independent variables at two- and three-levels. If one of the other motives is the dependent variable, there are oo.m m Hm.m m 0HiuomcHumuumm so.> em.m mm.e m mm.s mm.o s omznuomcHuaunmm m~.e sm.m om.a m mo.s mn.e H HmuuomcHumuumm mv.m o mm.m m oquuosnmuumm me.m mm.e a mm.m OH mH.e m em.e m omznuoeumunmm ms.s me.m s mm.m m Hm.s m ms.m m Hmunoeusuumm memos mass: 2 mass: 2 memo: z memos 2 mates 2 EM mm. .MM .Hb MM mo Hmuoe om mm em we Hm nouns eH.H H e>.o H bhmm mm.~ m eH.o H em.o m one mo.o H No.m H OH.v mn.m m «H.o H sm.o m cm m~.o H Hm.o H wo.H m >>.o H NH.o m oH ~o.o H o~.o H oo.o H ms.o H _|I||I Ho.o H ms nu mo.v em.m H oe.o H 0H.v mo.m m Ho.v «H.0H H Hoo.v He.HH m x m>Huozuo III 85 H I Ho.o H In." om.o H mm.o H mo.o H nomcHuuoHum oH.v NH.m H 0H.v om.m H mo.v mm.m H m~.v sm.H H oH.v sm.m H oH>HonsuummnH oousom m .HM>Im MU m .Hm>lm MU m .Hm>lm MU m .Hm>lm up m .Hm>lm mu Nlmmd NISU< mrfiod NIMCH MIMGH mmHndHHd> udmfidmmeGH .moHQMHum> usmmsommosH mo mHm>mHlo3u mus s can .mHm>mHImmH£u was Iozu and s .mHm>mHlmmH£u was |o3u msH s nuH3 mmmewsm sH .mHQMHHm> usmpsommo may no 30m z .msoum m sH.©mpsmm0Hm ous.pomHum> ucmcsmmmo meow on» nuH3 mommHmsm “wHQMHHm> unmosmmmocH so we w>Huos mso msHms mosmHum> mo mmmMHmsm w>HuomsHIm>Huo¢||.vH loma mN.o mo hH.m em No.m om hm.m Nb mm.m mo Hmuoe 54 mo.w mH mw.w mH onlx wH.h em wm.m mm m~.m om mm.m mm mm.m em omzux wm.m em ow.w mm mH.h mH om.h mm mm.m em Hmux om.m mm o~.o mm hm.m mm mo.m ow mw.m mm o>HuomcH mm.o ow mH.o mm om.m we .mw.m mm hm.m we m>Hu04 hm.m em en.m mm mh.o mm mm.m mm NH.h mm 0H>HO mm.m em we.m mm m~.m mm mm.m em Hm.m mm spams mHmsHmHmE mHmsHmHmz mHmsHmHmz mHmsHmHmz oo.m m oo.m m oqnuomsHI0H>HO om.o n HH.> m 00.x m om.m 0H om.m w omzlpomsHI0H>HU mm.m n wH.m m mm.w m oo.h 0H mo.m w HmluomsHI0H>HU mb.m m om.w w oqluo¢10H>HO mo.m OH mm.m m n>.m OH mw.w m Hm.h m oozluo¢|0H>HO mm.m 0H mH.m m mm.m m mm.n m oH.m m Hmluo¢|0H>HO mass: 2 mass: 2 mass: 2 mass: 2 mass: 2 AUmSGHHGOUV 30m ZII.wH mHQMB hwovH m mm.MH V OAIHUMGHIMflHmm H0.0H A HA.OH H mm.mH m OH.wH O OH.OH O OOSIHomcHumHHOH me.MH a Hm.>H H mw.OH m O0.0H m ON.OH m HmuuomaHusuumm O0.0H O mO.eH w OHIHoHIHHHmm mw.HH OH OO.HH H OO.eH OH NO.mH O OO.OH O Omzupoeusuumm HO.OH OH H0.0H A Ow.nH O wH.OH O w0.0H m Hmuuomusuusm mfimmz unto: 2 (mates 2 mass: 2 memo: 2 mass: 2 hm mm mm Mb hm HO Hsuoa om Om em we we Hounm OH.O H HN.H H mm.v mm.H m O~.O H m~.v ~m.H m one mm.v NH.~ H Om.O H ~0.0 m OH.O H ow.O m cm mO.v Ow.w H m~.v Om.H H mm.v OO.H N m~.v OO.H H O0.0 m OH O0.0 H wH.O H mm.O H m~.O H O0.0 H mH Hm.H H m~.v OH.~ H mm.H m HOO.v m~.~H H mO.v wm.w m x m>Huozno Om.O H H0.0 H OO.O H -.H H mH.O H HomnHuuoeum m~.v OO.~ H OO.O H m~.v OH.~ H m0.0 H O~.O H oH>HoumnummuH monsom m .Hm>uH HO m .Hm>-m HO m .HO>-H HO m .Hm>uH HO H .Hm>uH HO Nlmmd NI£U¢ Mlsod Nl3om ml30m mmHQOHHO> unmosommvsH l .moHanHO> unwosommosH mm mHm>mHlo3D .HH¢.m use mHm>oHlomugw was Io3p .£o¢ s mem>mH Iowan» was I03“ .3om s suH3 mothmsO sH .pomHHm> usopsommo msu mm HsH z HOmscHunoovuu.wH OHQOH 56 ON.>H OO OO.eH wO mm.OH OO O0.0H Ob O0.0H OO Hmuoa om.OH OH OO.eH OH OHIx H0.0H em mm.OH mm N0.0H om Ow.mH mm ow.OH em omzlx O0.0H em OO.eH mm O0.0H OH mw.OH Om N0.0H ON Hmlx O0.0H Om HN.OH mm OO.eH mm Ow.OH Om em.OH Om o>HuomsH mw.hH Ow O0.0H mm O0.0H we Ne.>H Om mm.OH om m>Hu04 w0.0H wm O0.0H Om OO.eH mm O0.0H ow O0.0H om UH>HO Ow.OH em NH.OH mm O~.OH mm mO.>H Om hw.OH om muumm mHmsHmHOz mHmsHmHmz mHmsHmHmS. mesHmHmz O0.0H m ON.OH w OHIUOOGHIOH>HU em.OH O O0.0H m w0.0H m wH.mH OH mm.mH O UmzluomsHIOH>HU Ow.OH O H0.0H m mw.OH m mm.OH OH O0.0H m HmluomsH|0H>HU Nh.wH O OO.eH w oqluo¢10H>HU OH.>H OH HO.wH m wN.OH OH O0.0H OH O0.0H O omzluom oH>HO O0.0H OH en.mH m mw.OH O mm.OH OH Ow.OH m HmluodI0H>HU mass: 2 name: 2 undo: .2 mass: memos z lOmquucooO HcH z--.wH OHan 57 O0.0 O O0.0 w OH-HoOcHuOHHOm O0.0 O O0.0 OH O0.0 e O0.0 O O0.0 O OmZuuomcHuOHHOm O0.0 O O0.0 OH O0.0 w HH.O. O O0.0 O HmuHoOcHuauumm O0.0 O O0.0 e oHnHoHumHHmm O0.0 OH O0.0 O O0.0 O we.O O O0.0 O Omznuumnmunmm O0.0 OH O0.0 O O0.0H O O0.0 O O0.0 O Hmnuomumunmm msmoz mfiflwz Z mcmwz Z mGMOS Z mammz Z mcmwz Z hm Hb mm Oh Oh HO HBOB OO eO em OO Ow Houum HH.O H O0.0 H O0.0 O I H0.0 H E O0.0 m OH.O O0.0 H OH.O H HH.n O0.0 O OO.v O0.0 H OH.v O0.0 O om H0.0 H O0.0 H Om.v O0.0 m O0.0 H OH.O O 0H OH.O H O0.0 H MMLMOOO.O H O0.0 H O0.0 H OH H0.0 H O0.0 H OH v OO.O O O0.0 H O0.0 O x m>HH02no Om.v OO.H H mm.v OO.H H OH.O H OH.v O0.0 H OH.H H HomcHuHoHIO OO.H H O0.0 H O0.0 H OO.H H Ow.O H oH>HosOHHmHuH mousom O .Hm>um HO O .HO>:H HO O .HO>IH HO O .Hm>um HO O .HO>-H HO OIHHH muHcH OuHcH mnsom Ousom moHQmHHO> usmosmmmpsH .mmHQOHHO>-usmOsmswosH map mm mHo>OHlo3p .HH«.m one «mHm>mHuooHsu mam Iosu .HsH s umHo>wHIme£u UGO I03» .3om s suH3 mommHmsO :H .mHQOHHO> unmosmmmw gnu mm and z HOmsaHHcoovnu.wH mHan 58 OH.O OO w0.0 NO Hm.O OO w0.0 OO O0.0 OO Hmuoa O0.0 OH Hm.O OH oqlx O0.0 em O0.0 Om OH.O em O0.0 Om ON.O wm Umzlx O0.0 wm O0.0 Om Ow.m em O0.0 Om mw.O ON Hmlx O0.0 Om O0.0 ow e0.0 mm H~.O Om O0.0 om m>HuomsH O0.0 ow N0.0 mm O0.0 ew O0.0 Om O0.0 OO o>Huo¢ O0.0 em Hm.O Om .N0.0 mm O0.0 ow O0.0 OO uH>HU O0.0 em O0.0 Om O0.0 mm O0.0 Om O0.0 OO muumm mHOsHmumz mHmsHmHOS mHmsHmHmz . O0.0 m mm.~H w oqluomsHIOH>HU Ow.O O O0.0 OH O0.0 w O0.0 OH OH.O O omzluotsHloH>HU O0.0 O O0.0 OH O0.0 w O0.0 OH O0.0 m HmluomsHloH>Ho O0.0 O O0.0 w 0H|u0<10H>HU ow.m OH O0.0 O O0.0 O O0.0 OH OO.HH O oozluodnoH>HU O0.0 OH O0.0H O O0.0H O O0.0H OH O0.0H m Hmluo¢IOH>HO names 2 mass: 2 2 meme: 2 msmoz 2 mass: AOmsaHucooO now z--.OH mHan 9 5 OO.O m O0.0 O OO.O e oHuHum OO.O .O OO.O O O0.0 OH OO.O w O0.0 O OO.O O OmzuHum OO.O O OO.H O OO.H OH OO.H w OO.H O OO.O O HmnHum OO.O O O0.0 O OO.O O oHneum OO.O O O0.0 OH OH.O O OO.O O OO.O O OH.O O Omzueum OO.O O OO.O O OO.O O OO.O O OO.O O OO.O O HOIHIO mGMOZ memo: z memos 2 mass: 2 mass: 2 memo: 2 mass: 2 .mw hm HO .m.» MO .mm HO Hoe OO 3 I HO O I OO Inn... Ow Hounm OH.H H OO.v OO.H O O.v OO.O H OO.H O OO.v OO.O H OO.v OO.O O OOH Hw.O H OO.O O OO v OO.O H OO.v OO.H O OO.O H O0.0 O om OO.v OO.O H O0.0 O dHal OO.O H O0.0 O OO.O H OO.H O 0H OO.v OO.H H OO.O H .mnpw OO.H H OO.H H null OO.H H unnn HO.H H OH OO.O H OH.v OO.O O HO.v OO.O H O0.0 O Hn.v HO.O H OH.v OO.O O xuuozuo OO.v OO.O H III: OO.O H Odqw OO.O H OO.O H OH.v OO.O H On.v OH.O H Hueum OO.v OH.O H OH.v O0.0 H ”daqude H OO.v OO.H H OO.v OO.H H Oqu O0.0 H cumue OOHDOW m >-H HO O >IH HO O >IH HO O >uH HO H >um HO O >IH HO Ounce nurse . OuHcH O'HOH Ousom w: Huang: meanHm> usmvsmmmUGH .mmHhmHHm> usmOswmmOsH mm ImHm>mHIme£u was 103» sod s was .me>oHIme£u was IOBH HsH s .mHo>mHlmmH£u cam :03» .3om s £HH3 mommHOcO sH .oHanum> usoosmmop may we Hus z HOmscHucoovnu.eH OHQOH 60 OO.N wO No.m OO OO.N NO OO.N OO NO.N OO mO.N 0O HMUOB O0.0 OH O0.0 OH O0.0 ON OHIx wO.N NO O0.0 om Om.m Om wo.m wN mw.m Om OO.N ON GOEIX mm.N Nm OO.H OH OH.N Om om.N wN oo.N Om OO.H ON Hmlx OH.O NO Nm.m NN ow.m ow O0.0 NN HN.m Om OH.O om HUMGH NN.N Nm OO.N ww mN.N Nm OO.N ww wN.N Om mm.N om w>Huo¢ Hm.N Om Nw.N Om ON.N Om mm.N mm mw.N ow ON.N om OH>HU OH.O ON HO.m mm O0.0 Om Nw.m mm O0.0 Om ON.m om muumm mHmsHOHma mHOsHmHOE mHmsHmHmz mHmsHOHmz O0.0 m O0.0 m OO.N w OHIHIO O0.0 O O0.0 O ow.m OH ON.m w O0.0 OH O0.0 O woleIU OO.N O OO.H m OO.N OH O0.0 w OH.O OH ow.m m HOIHIU OO.N O O0.0 O OO.N w OHIOIO HH.O O O0.0 OH OO.H O OO.H O OO.N OH OO.H O UOSIOIU O0.0 O O0.0 O OH.H O OO.N O O0.0 OH O0.0 O HOIOIO mammz z mammz z mammz 2 names 2 memo: z memo: z HOmquHcooO HHH z--.wH mHnOH 61 five such analyses; 2 Aff at two-levels as the independent variable, and the other two motives at two- and three- levels. Thus the same interaction and main effect can be significant in a variety of analyses. This is the case in this analysis. The 32 significant results in reality represent 16 different effects. Six of these 16 effects represent interactions between two motives alone and ac- count for 11 significant results. These two-motive in— teractions show essentially the same relationship as found by the correlations between the motives presented in Table 6: high Q Pow is associated with high a Inf and low g_Aff and vice versa, low E Pow is associated with low 3_Inf and high §_Aff. These relationships are straightforward and will not be discussed here. The other ten separate effects are more interesting and pre- sent a problem in interpretation. If a similar main ef— fect is significant in two out of five possible analyses, and a different main effect significant in four out of five possible analyses, can the same degree of confidence be placed in both results? To resolve this difficulty, the following Accept- ance Rule was used: In order to be reported, over one- half of the possible analyses relating to a main effect or interaction must be significant at the 10% level. If exactly one-half of the possible analyses are significant 62 at the 10% level, at least one result must have a prob- ability level of 5% or less. Utilizing this rule, five effects are eliminated for not satisfying the requirements; five effects remain as significant and will be reported. See Table 15 for a summary of the ten effects and analyses related to each effect. Results Considered Significant by the Acceptance Rule Analyses of Variance In all the analyses related to Result No. 1, ex- cept the two-level E Inf, the Civic group has higher a Pow than the Party group. The Party group has significantly higher 2 Aff scores than the Civic group in half of the analyses re- lated to Result No. 2. Means are given in Table 15. Result No. 3, an interaction between E_Inf, Party vs. Civic and E_Aff is significant in two-thirds of the analyses and is shown in Table 16. Inspection of cell means suggests that a person with low E_Inf and high E_Aff will be affiliated with a political party. N Ach and Active-Inactive interact with g Pow in all analyses related to Result No. 4. Visual inspection of cell means (Table 17) indicates that two combinations of E_Pow and §_Ach are associated with activity (a) high 63 Table 15.--Summary of ten significant effects and the analy- ses relating to them, and probability levels of effects if appropriate. Effects Reported in the Text of this Paper 1. Party vs. Civic, £_Pow main effect (A) g Inf as n Ach " in f n n n u A Ch I: u u n Af f u u n n Three level Three level Two level n Two level Two Level the independent variable E E. 2. Party vs. Civic, §_Aff main effect (A) Three Three Three level level the independent variable I! ll level _ Two level Two level Two level In f II " II II ACh I! H II I! lbflflb 3. Party vs. Civic, a Inf, g Aff interaction (AC) Two level n Aff as Three leveI n Inf " Two level E_Inf " " " " the independent variable II II 4. Active vs. Inactive, n Pow, n Ach interaction (BC) ‘ " Three level a Ach as Two level n Ach " Three leveI n Pow " " " " Two level Q_Pow " " " " the independent variable 5. Party vs. Civic, Active vs. Inactive, §_Pow and E_Aff interaction (ABC) Two level E_Aff as the independent variable Three level E_Pow " " " u Two level a Pow " " n n '0'!) "O'U 'U'U'U'U < .10 < .05 < .10 < .10 < .10 < .10 < .05 .10 .10 .10 .05 AAAA 64 Table 15.--(Continued) A Effects Not Rsported in the Text of this Paper 1. Active vs. Inactive, §_Ach main effect (B) Three level a Pow as the independent variable Three level n Inf " " " " Two level E_Pow " " " " Two level n Inf " " " " Two level :n: Aff " " " " Active vs. Inactive, g Aff main effect (B) Three level n Pow as the independent variable Three level 5 Inf " " " " Three level E Ach " " " " TWO level 2 Pow II II II II Two level n Inf " " " " Two level E Ach " " " " Active vs. Inactive, g Inf, 2 Aff interaction (BC) Two level g_Aff as the independent variable Three level a Inf " " " u Two level £_Inf " " " n Party vs. Civic, Active vs. Inactive, 3_Pow and g_Ach interaction (ABC) Three level g_Ach as the independent variable Two level n Ach " " " " Three leveI n.Pow‘ " " " " Two level §_Pow " " " " Party vs. Civic, Active vs. Inactive, g Inf and E_Aff interaction (ABC) Two level E_Aff as the independent variable Three level a Inf " " " u Two level E Inf " " n n 65 Table 16.--Cell means of an interaction between Party- Civic, g Inf and n Aff when either motive is the dependent.varIable. Party: Civic N Inf as the Dependent Variable N Aff High 14.8 18.2 Low 18.0 17.6 N Aff as the Dependent Variable N Inf High 2.4 2.0 Low 4.5 2.4 66 Table l7.--Ce11 means of an interaction between Active- Inactive, g_Pow, and §_Ach when either motive is the dependent variable. Active InaCtive"* N Pow as the Dependent Variable N Ach High 6.6 5.8 Low 5.3 6.5 N Ach High 7.9 5.7 Med 4.9 6.2 Low 5.6 7.0 N Ach as the Dependent Variable N Pow High 10.0 6.0 Low 7.6 8.4 N Pow High 10.3 6.6 Med 9.3 7.1 Low 6.6 10.0 67 a Pow and high E_Ach and (b) low E_Pow and low §_Ach. The other combinations, high g_Pow-1ow §_Ach and low a Pow— high E_Ach are more often related to inactivity. These interactions suggest that g Pow and §_Ach are positively correlated in the active group and negatively correlated in the inactive group. This is the case when rank order correlations are computed: the Active correlation is +.37 (n=55) the Inactive correlation is -.16 (n=39). Result No. 5, a triple interaction between Party vs. Civic, Active vs. Inactive, §_Pow and ngff is signifi- cant in two-thirds of the analyses. Inspection of cell means shown in Table 18 suggest that the most unusual group regarding these two motives is the Party Inactives. The difference in §_Aff scores across the g Pow levels is largest for the Party Inactives. The Civic Actives' group scores also change across g_Pow levels, but not as-much as the Party Inactives' scores. The Party Actives and Civic Inactives remain fairly consistent across all levels of g Pow and also are similar to each other when the means in Table 18 are considered. The uniqueness of the Party Inactives is also suggested by the large negative rank order correlation between §_Aff and E_Pow, -.66. The cor- relation between £_Aff and g_Pow for the other groups are: Party Actives, -.30; Civic Actives, -.23: and Civic In- actives, -.06. 68 Table 18.--Cell means of the interaction between Active- Inactive, Party-Civic, E_Aff and a Pow when E_Aff is the dependent variable. High N Pow Low N Pow Party 'Civic Party Civic Two-Level n Pow Active 2.7 0.7 3.0 2.7 Inactive 1.6 3.1 5.0 3.2 High NPow Medium Pow Party 7 Civic Party Civic. Three-Level n Pow 9 Active 3.0 0.6 3.2 1.7 Inactive 0.8 3.4 4.0 2.7 Three-Level n Pow Active 3.3 2.5 Inactive 5.8 2.8 69 Non—Parametric Results By chance 5 results can be expected to be signifi- cant at the 10% level. As reported below, 7 results are significant, a number not much greater than that amount expected by chance. This many indicate that the follow- ing results are random and should be regarded only as suggestive. When compared with the Ward group, the Electives have higher a Pow (p = .06, two-tailed) and higher a Inf (p = .07, two-tailed). Those Party Leaders and Party Sub- leaders who self—initiated their entry into politics have higher g_Pow (p = .06, two-tailed), higher a Ach (p = .02, two-tailed) and lower §_Aff (p = .02, two-tailed) than those Party Leaders and Party Subleaders who were recruited by the Party. When the party offices are categorized by the power potential of the office (described in Browning, 1964) those men who held offices with high power potential have higher a Pow (p = .03, two-tailed) than those men holding offices with low power potential. Those Party Leaders and Party Subleaders whose father was not active in politics had higher 2 Pow scores than those with politically active fathers (p = .08, two-tailed). 70 Cell means of the significant results are presented in Table 19. The means and probability levels of all the subgroup comparisons with each other are given in Table 20. Possible Interpretations (1) One possible interpretation of Results No. 1, 2, and 3 given in Table 15, is in terms of the expectation of each office and the rewards contingent on the behavior of each man. The Party politician must be elected by and re- main on good terms with his constituents to further his career. This involves campaigning, meeting the public, in general being a man of the peOple. Obvious attempts at influencing or pressuring the public are frowned upon by the public and may have consequences at the next election. A wise politician would not be likely to publically show or give any indication of g Pow or a Inf strivings, but would emphasize the E_Aff strivings in all his relationships. A businessman involved in urban renewal is not elected and is required and encouraged by his position to accentuate his power concerns in dealing with other busi- nessmen and the "consumer." A businessman is rewarded for showing 2_Pow concerns but would infrequently be rewarded by his superiors for showing great concern with E_Aff strivings in business deals. Here is a difference in the expected roles of a politician and businessman which could 71 Table l9.--Cell means of seven significant results computed by non-parametric tests. Elective Subleaders vs. Ward Subleaders N Pow N Inf Elective 6.9 18.4 Ward 3.7 14.8 Self-Initiated vs. Recruited Party Members N Pow N Ach N Aff Self-Initiated 6.8 9.8 1.3 Recruited 4.7 5.9 4.4 Power Potential of the Office, High vs. Low N Pow Low 4.5 High 7.4 Father Active vs. Inactive in Party Politics N Pow Father Active 4.6 Father Inactive 6.0 72 Table 20.--Ce11 means and probability levels of the compari— sons of subgroups among themselves, with the four motives as dependent variables. Matched Comparisons including mean of inactive group, E of comparison, T-value, and two-tailed p-value of the Wilcoxon Matched Signed Ranks Test Inactives N Mean N T p Ward 7Ac-4I g Pow 4 3.58 8 2.0 .27 g Inf 4 13.48 7 0.0 .11 g_Ach 4 9.50 7 1.0 .29 g_Aff 4 6.00 8 4.0 .72 Elective llAc-lOI r_1_ Pow 10 6.33 20 22.0' .58 §_Inf 10 17.61 20 20.0 .44 g_Ach 10 6.20 20 11.5 .10 g_Aff 10 2.40 18 13.5 .52 Appointive 5Ac-4I g Pow 4 6.50 8 4.0 .72 g_Inf 4 18.00 8 3.0 .46 g_Ach 4 6.75 8 5.0 1.00 E_Aff 4 2.75 8 3.0 .46 Table 20.--(Continued) 73 Subgroup Comparisons by Mann Whitney T-tests including mean of groups, U-value, and two-tailed probability levels Groups N Mean Comparisons “U p 11222: Ward 7 3.73 Elective-Ward 17.5 $232 Elective 11 6.89 Elective-Appointive 23.5 .65 Appointive 5 6.26 ward-Appointive 9.0 .17 N Inf Ward 7 14.81 Elective-Ward 18.5 $223; Elective 11 18.39 Elective-Appointive 16.5 .21 Appointive 5 16.48 Ward—Appointive 13.0 .53 N Ach Ward 7 5.86 Elective-Ward 21.0 .110 Elective 11 8.27 Elective-Appointive 25.5 .82 .Appointive 5 8.40 Ward-Appointive 15.0 .68 N Aff Ward 7 4.00 Elective-ward 31.5 .55 Elective 11 3.55 Elective-Appointive 26.5 .94 Appointive 5 3.80 Ward-Appointive 13.3 .56 74 be the basis for the three results. The Party politician is rewarded for §_Aff strivings and would be expected to continue them; a businessman active in civic affairs would be rewarded for a Pow concerns, which he would be expected to continue. These would be the motivations which each would be willing to demonstrate in contacts with the public and with the experimenter. These are then reflected in the TAT stories and scores. One point to note concerning Results No. l and 2: These same main effects were not significant in the single motive analyses. In these designs, the addition of a second motive as an independent variable accounts for some variance previously labelled error variance. With some variance eliminated, effects previously hidden are significant. (2) The E Pow-E_Ach interaction (Result No. 4) is interesting in that activity appears to be the outcome of two different motive arrangements, either high E_Ach and high Q Pow or low 2 Ach and low g_Pow. This would imply that two very different persons would be active in party or civic affairs. A possible link between g_Ach- 2_Pow interaction and those persons who are recruited or self-initiate their entrance into politics is suggested by the Mann-Whitney U Tests in the exploratory analyses. It was tentatively found that those persons who self- initiate their entrance into politics have higher a Pow 75 and E Ach and lower g_Aff than those recruited. This would correspond to the g Pow—§_Ach interaction: the self-initiators may be those with high Q Pow and §_Ach, the recruits may be those with low g_Pow and E_Ach. (3) Result No. 5 is difficult to interpret. It is obvious that the inactive group matched to the Party Sub— leaders contributes a great deal to the negative correla- tion between E Pow and E_Aff found in the entire sample. The significance of this result is unclear and no explana- tion can be offered. (4) Since the results gathered from the non-parametric tests must be regarded as random, they will not be con- sidered further. APPENDIX C Cell means, F-values and probability levels for the analyses of variance computed with uncorrected motive scores and story length. 76 O0.00w O0.0 O0.0 O0.0H H0.0 OO HOHoe O0.00w O0.0 O0.0 O0.0H O0.0 Om HOOHOHOS Onsnm>HuomcH OO.HHO H0.0 O0.0 O0.0H O0.0 OO HOOHOHOz Ousuw>HuoH OO.OOO OO.O OH.O OO.OH OO.O Ow HmcHOumz OHs-oH>Ho O0.0ew O0.0 H0.0 O0.0H OH.O ee HOOHOHOS Ouzumuumm O0.000 O0.0 O0.0 O0.0H OH.O OH cums m>HHomcHn0H>Ho O0.000 OH.O O0.0 H0.0H O0.0 OO cams m>HHomuoH>Ho HH.OOO H0.0 O0.0 OO.eH H0.0 OH cums m>HHoOOHnmunmm O0.00e H0.0 O0.0 O0.0H O0.0 OO cams m>HH64uOHuOm OHOOOH.OnoHO HHH 2 One 2 HcH z 30m z 2 meat: HHmo Om HO Hmuoa wm HOHHH OO.v OOO.H OO.v HOw.H HH0.0 OO0.0 OO0.0 H OH OO0.0 eO0.0 ONO» HO0.0 OOO.H OO0.0 H m>HHomcHnm>HHoHnm OO.v OHO.O OOO.O OO.v. OOO.O OH.v. OOe.O OO.v OHO.e H oH>Housuummne OmcHnsoo mumOmmHnsm UGO muovmmq suHs mmthmsd a msHm>|m Am msHO>Im a msHmblm m msHsrlm m msHm>|m O HO mousom :HOcmH Ououm HHO 2 rod 2. h, HsH z 30m z .sm>Hm mum Hmv mHm>mH OHHHHQOQOHQ can mmsHO>Im .HHOV Eocemum Ho moonmmc .msmofi HHwU mm QHOGmH muoum use mm>HuoE Omuomunooss Odom on» nuH3 mosmHHm> Ho mmthmseln.HN anOB .mmHQmHum> uswwsommw 77 O0.000 O0.0 O0.0 O0.0H H0.0 OO Hmuoa OO.OOO OO.O O0.0 O0.0H O0.0 Om HOOHOHmz-m>HHomaH O0.0HO O0.0 OH.O O0.0H O0.0 we HmcHOHmzuo>HHoH OO.OOO O0.0 OO.O O0.0H OO.O Ow HOOHOHmzuoH>Ho OO.OOO OO.O OO.O OO.OH OH.O Oe HOOHOquuspumm O0.000 O0.0 O0.0 O0.0H OH.O OH cums m>HHomcHuoH>Ho O0.000 O0.0 O0.0 O0.0H O0.0 OO cmms m>HHo¢noH>Ho HH.OOO H0.0 O0.0 OO.eH H0.0 OH cums m>HHomcHuOHHOm O0.000 O0.0 O0.0 OH.OH O0.0 OO cum: mpHHoHuOHHOH OHOcmH Ououm HHH z OoH z HcH z 30m z z mammz HHmo OO HO Hmuos mm. HOHHW OO0.0 HO0.0 OO0.0 OO0.0 OOH.O H mm OOw.O OO0.0 OOO.H on.H :_ OO0.0 H m>HHomcHum>HHoHum OO.v OOH.O OO0.0 OO.v OHO.e OH.v..wO0.0 OO.v OOO.H H oH>HouOHHOmuO mesouw Odom "OHso mumOOmHhsO nqu mmmOHssm m msHm>Im m msHm>Im m osHm>Im m 05HO>Im m msHO>|m HO monsom gumnma Ououm HHH z :oH z HOH z 30m z HOmscHHaooOuu.HO mHnms MICHIGAN STATE UNIVERSITY LIBRARIES IIII IUJIIIILIII'I 3 1293 24 III 3