|
4.6.1 Differences Between Boys and Girls Gender differences in interpersonal needs have been reported in studies. Men were found to have difficulty in intimate relationships more than women (McGill, 1985; Rubin, 1979, 1985). A recent study reports that females value closeness (Openness) in relationships more than males do and males value autonomy (Control) more than females do (Bakken & Romig, 1992). The present study shows significant differences in four of the FIRO variables, namely, WEI, PRO, WRC and WRO. Table 4.6.3 indicates that, although there is no difference between the boys and girls in their current (expressed) inclusion behaviour, the boys want to initiate inclusion more than girls (p<.05); while the two groups are equally low on expressed openness with others, girls seem to receive as well as want to receive more openness in their relationships than boys (both p<.001); although the present level of expressed control (one form of autonomy) is low in both the groups and both want to have much higher levels of it than what they have at present, the girls want more autonomy than the boys want in terms of being freed from external control (p<.05). 4.6.2 Differences Among FIRO-Strength Groups of Students In order to examine whether students of varying FIRO strengths differed in their interpersonal behaviour patterns, we carried out an analysis of variance by that grouping variable. The results are presented in Table 4.6.4. Students with a low overall FIRO seem to be most concerned with their received openness & received control and least concerned with the dimension of Inclusion. The medium-FIRO group shows a dominant concern for all the three dimensions of Inclusion, Control and Openness, with special emphasis on a particular aspect of each of them: It is WRI in Inclusion, WEC in Control and WRO in Openness. The group is moderate on other aspects. The within-group variation in the high-FIRO group shows least scores on received control and tops on WRO and WRI, with high scores on other variables. It is also noted from the table that ten of the F-ratios are significant, meaning that the three FIRO-strength groups truly differ in ten of the twelve interpersonal variables. The two FIRO variables they do not show a difference on are PRC and WRC; that is, irrespective of the intensity of their overall interpersonal orientations, all the students uniformly experience being controlled by others and all of them would equally like to be rid of external control. 4.6.3 Differences Between Regional Sub-groups of Students Unlike the managers (see 4.5.5 above), students do not manifest significant regional differences in any of their interpersonal needs (Table 4.6.5). We also saw that the academic backgrounds of managers and students (Tables 4.5.2 and 4.6.2, respectively) did not show any differences in their interpersonal needs. Given these observations, one wonders if the differences in the needs of managers and students could be accounted for by the nature of work performed by either group. The inter-departmental differences observed in the study (presented and discussed in an earlier paragraph) might be a pointer to such a possibility. 4.6.4 Differences Between Poor and Better Students Top graders and bottom graders among the students did not show a significant difference in any of their interpersonal needs (see Table 4.6.6). It may be recalled that effective managers were higher on expressed control and received openness than ineffective managers (Table 4.5.6). Again, one wonders if academic success and managerial success are different in their nature. The author's own observation, over the years, of the on-the-job performance of academically high-scoring management students leads him to conjecture a low correlation between the two. The results indicate that, while student success is independent of their interpersonal behaviour, managerial success seems to be significantly dependent on the manager's interpersonal behaviour -- particularly on the expressed control and received openness dimensions. 4.7.0 Summary of Analyses and Test of Hypotheses Our analysis of data began with a simple computation of frequencies of managers and students, falling in the ten different score positions (from 0 to 9) on each of the twelve FIRO scales. That showed us the various proportions of respondents occupying various positions on the respective scales. From a summary of these frequency distributions, it was found that a great majority of managers were at the higher end of the following scales: WEI (80%), WEC (74%), WRI (70%), WRO (64%) and PEI (60%); and at the lower end of WRC (71%) and PEO (63%). The comparable distributions of students were: WEI (77%), WEC (72%), WRI (73%), WRO (74%), and PRO (63%) at the higher end of the scales; and WRC (85%), PEC (64%) and PEO (62%) at the lower end. Vast majorities of managers as well as students want to interact and associate with people very much, both at their own initiative and at that of others. A great majority of both the groups, however, maintain their personal contacts at a superficial level and seem to be wary of being personally close and open in their relationships. About half the number of managers and about two-thirds of the students feel powerless at present, about three-forth of both the groups desire to take charge, initiate and exercise power in their interpersonal relations. After the descriptives, we took a look at the mean scores of all the variables as another simple way to make sense out of the mass of data and compare groups on that basis. While, as a measure of central tendency, the means could be used for simple comparison of several groups, they would be inadequate for the purpose of testing any hypothesis, because the variances in the groups were ignored in their means. We, therefore, proceeded to apply univariate statistical tools, which would consider the means and the variances in the analysis. The t- and F-ratios for the various groups helped us test the hypotheses we had set. Hypothesis H1 stated that managers and management students did not differ in any of their interpersonal needs. The t-ratios in Table 4.4 (page 69) urged rejection of the hypothesis with regard to five of the interpersonal need variables: PEC, WRC and WRO at p<.001; PRI and PRO at p<.01. Managers and management students were thus found as having differences in their interpersonal needs. Hypothesis H2 stated that managers of different age groups did not differ in any of their interpersonal needs. This hypothesis is also rejected on the basis of the F-ratios discussed in Section 4.5.1 above. Different age groups of managers differed in their interpersonal needs, with regard to PEI at p<.001; PEC and PRI at p<.01; and WEI at p<.05. Hypothesis H4 is rejected, for the F-ratios discussed in Section 4.5.3 above showed significant differences among departments in the Inclusion, Control and Openness dimensions of their interpersonal behaviour. Specifically, the differences were in PEI at p<.01; in PEC, PEO, WEC and PRI at p<.05. On the basis of the t-ratios, discussed in Section 4.5.5, managers of North and South regions were found to differ in PEO and PRC at p<.01; in PEI, PRI and WRC at p<.05, thus rejecting H5. The hypothesis about successful and not-so-successful managers (H6) was rejected, for the two groups differed significantly in four interpersonal aspects: PEC at p<.001 and WEC at p<.01; PRO and WRO at p<.05 (for discussion, see Section 4.5.6 ). By rejecting H1, managers and management students are said to be different in their interpersonal needs. While hypotheses H2 to H6 related to differences within managers, H7 to H11 were about differences within students. Only one of the latter five hypotheses was rejected, the one relating to gender (H9). That is to say, the interpersonal needs of the students are the same, irrespective of the latter's age, academic discipline, region of origin and academic performance; but the male and female students differ in WEI and WRC at p<.05; in PRO and WRO at p<.001. The fact that hypotheses H3 and H8 could not be rejected may indicate that pursuit of, and training in, a particular academic discipline does not alter the interpersonal need patterns of people. The popular belief that students of Arts are better at human relations than, for example, engineers may, therefore, be questioned. In order to examine possible associations among the FIRO variables themselves, a bivariate statistical tool was used in the form of Product Moment Correlations. The coefficients of correlation showed us that the interpersonal variables were indeed associated with one another. A few differences were observed in the variables that were associated in the managers, compared to those that were so in the students (see discussion in Section 4.4). Openness and Control, for instance, were found to be correlated in the managers, while they showed no such relationship in the students. 4.8.0 Discriminant Analysis Although the preceding analyses sufficed to test all the hypotheses that were pursued in the present study, it may be noted that all the techniques used for the purpose considered only one variable at a time (except for Pearson's Correlation, which treated any two of the variables simultaneously). Groups that differ (or do not differ) on a particular variable, when the latter is considered in isolation, may behave differently on the variable when the same variable collaborates with other variables. Univariate analyses cannot capture this "conjoint" phenomenon, whereas multivariate analyses do. In order to examine at one shot, therefore, as to which of the twelve FIRO-need variables, when taken together, differentiate the various groups, the data were subjected to a multiple discriminant analysis, which would extract the best combination of the predictor variables to maximise differences among groups. A step-wise (Wilks) discriminant analysis would select variables on the basis of their discriminating power, construct a vector and then, on the basis of the vector, classify subjects into exclusive groups. This technique can, thus, help us identify the one combination of the FIRO variables which can classify, for example, existing managers into the manager group and existing students into the student group. The degree of correct classification or the "hit ratio" would tell us of the goodness or the suitability of the discriminant function. If, by the criterion of correct classification, the function turns out satisfactory, then the variables included in that function would be the discriminating variables. The absolute magnitude of the discriminant function coefficients can also (tentatively) indicate the relative importance of the variables. The summary results of these analyses for select groups are presented and discussed in the following sub-sections. 4.8.1 Managers and Management Students The results of the discriminant analysis for the managers and students are summarised in Table 4.8.1. Besides upholding the rejection of hypothesis H1, these results give us the additional information on the particular combination of FIRO needs that discriminate the managers and the management students. A combination of eight variables discriminated the two groups. The variables were: PEI, PEC, PRI, PRO, WEI, WRI, WRC and WRO. However, four of these variables (PEC, WRO, PRI and WEI) seem to have contributed more to the discriminant function than the rest of them, as indicated by the standardised function coefficients (Table 4.8.1). The managers exert Control, receive Inclusion and want to socialise more than the students seem to do in those three interpersonal need areas, but their desire to receive openness from others is lower than found in the students. Table 4.8.1 : Summary of Discriminant FIRO for Managers & Students
Mgr. Mean Stu. Mean Function Wilks' Signi.
............... n = 253 n = 322 Coeff.
PEI 5.01 4.73 -.25382 .87323 .0000
PEC 4.34 3.40 .51181 .91999 .0000
PRI 4.38 3.71 .48544 .88918 .0000
PRO 4.64 5.19 -.27227 .88485 .0000
WEI 6.12 5.93 .47288 .88268 .0000
WRI 5.76 6.12 -.37570 .87619 .0000
WRC 3.23 2.32 .38671 .89613 .0000
WRO 5.24 6.25 -.50128 .96081 .0000
------------------------------------------------------------------------
Group Centroids: Managers = .42910; Students = -.33715
------------------------------------------------------------------------
Classification Results
Actual Group No. of Cases Predicted Predicted
...........................................Membership Membership
Managers Students
Managers 253 140 113
55.3% 44.7%
Students 322 77 245
23.9% 76.1%
-------------------------------------------------------------------------------------------------------------
4.8.2 Successful and Not-so-successful Managers In addition to providing further supportive evidence to the rejection of hypothesis H6, Table 4.8.2 tells us that six of the twelve FIRO variables (PEI, PEC, WEC, WEO, PRC and WRO) discriminate successful managers from not-so-successful ones. The accuracy of the discriminant function used for Table 4.8.2 is very high: its hit ratio for the successful managers is 91 per cent and for the not-so-successful ones, 84 per cent. While the earlier univariate t-ratios of Table 4.5.6 showed PRO to be discriminating and PEI, WEO and PRC to be non-discriminating, when tested separately, the multiple discriminant analysis has dropped the former and included the latter. The most discriminating variable between the good and the poor managers is PEC, followed by WEO and WRO (see coefficients in Table 4.8.2). The interpersonal behaviour of successful managers, thus, appears to be typically oriented toward making an impact on others; these managers seem willing to be open with others and have a great desire to receive openness from others. 4.8.3 North and South Managers While our earlier univariate analysis showed that managers of North and South Indian origin differed on five of the variables (see Table 4.5.5), the results of multivariate analysis (Table 4.8.3) shows PRC, PEO and PRI to be the three typical variables that discriminate the two groups of managers. Hypothesis H5 remains rejected. Table 4.8.2: Summary of Discriminant FIRO for Poor & Good Managers
Mean Mean Function Wilks' Signi.
Poor Good Coeff. Lamda
managers managers
n = 31 n = 34
PEI 5.10 5.76 -.32338 .48484 .0000
PEC 3.55 7.03 1.20090 .59829 .0000
WEC 5.26 7.18 -.27839 .47516 .0000
WEO 4.61 4.21 -.46722 .52067 .0000
PRC 5.35 4.41 -.34754 .54982 .0000
WRO 4.68 6.06 .46021 .49802 .0000
-----------------------------------------------------------------------
Group Centroids: Poor Managers = -1.08359; Good Managers = .98798
-----------------------------------------------------------------------
Classification Results
Actual Group No. of Predicted Predicted
Cases Membership Membership
Poor Managers Good Managers
Poor Managers 31 26 5
83.9% 16.1%
Good Managers 34 3 31
8.8% 91.2%
----------------------------------------------------------------------
Percentage of cases correctly classified: 87.69%
______________________________________________________________________
Table 4.8.3 : Summary of Discriminant FIRO for North & South Managers
Mean Mean Function Wilks' Signi.
North mgrs. South mgrs. Coeff. Lamda
n = 137 n = 81
PEO 3.76 4.49 .56612 .96685 .0070
PRI 4.25 5.02 .37793 .92840 .0010
PRC 4.50 5.38 .66236 .93732 .0012
-----------------------------------------------------------------------
Group Centroids: North Managers = -.21255; South Managers = .35950
-----------------------------------------------------------------------
Classification Results
Actual Group No. of Predicted Group Predicted Group
Cases Membership Membership
......................................North Managers South Managers
North Managers 137 122 15
89.1% 10.9%
South Managers 81 59 22
72.8% 27.2%
------------------------------------------------------------------
Percentage of cases correctly classified: 66.06%
__________________________________________________________________
4.8.4 Managers of Different Departments Table 4.8.4 shows that, of the six variables which discriminate the various departmental managers, PEI, PRO and PEO seem to have contributed significantly to discriminate the production managers from the other three departmental managers. They are comparatively low on socialising and are less open than the others. The P&I managers, compared to the marketing and quality-control managers (Function 2), seem to have a characteristically high desire to influence (WEC) people; they also seem to influence people at present (PEC) more; they are (comparatively) also more willing to be influenced by others (WRC). The marketing managers and the quality-control managers have been differentiated on PRO, PEC and PEO (Function 3), but this finding is to be taken with some caution: the classification results for these two groups (marketing and quality control) have turned out to be very poor and, as may also be noted, their group centroids (on which the classification was based) are rather too close for the two groups to be really different. The first two functions cumulatively explained about 96 per cent of the variance, whereas function 3 did less than 4 per cent. Table 4.8.4: Summary of Discriminant FIRO for Departments.........Mean Mean Mean Mean Function Function Function Wilks' Signi. .........Score Score Score Score Coeff. Coeff. Coeff. Lamda .........Prodn. P & I Mktg. Q.C. Fn. 1 Fn. 2 Fn. 3 .........n = 70 n = 56 n = 17 n = 15 PEI 4.30 5.61 5.06 4.73 .58 -.10 .03 .91230 .0027 PEC 3.60 4.93 4.65 4.33 .24 -.75 -.69 .81566 .0019 PEO 3.54 4.48 4.29 3.60 .41 .17 -.48 .78032 .0042 WEC 6.03 6.79 6.33 4.71 .17 1.08 -.10 .86567 .0011 PRO 4.67 4.70 4.76 4.13 -.44 .21 .82 .79839 .0031 WRC 3.11 4.07 2.94 3.27 .27 .49 .26 .84033 .0016 -------------------------------------------------------------------------------------- Group Centroids: Prdn.= -.38304; P&I= .49480; Mktg= -.01444; QC= -.04336 --------------------------------------------------------------------------------------
Canonical Discriminant Functions
Fcn Eigen- % of Cumu Canonical After Wilks' Chisquar DF Signi.
value Variance % Fcn Lambda e
Corr
0 .7803 37.705 18 .0042
1* .1559 59.15 59.15 .3673 1 .9020 15.680 10 .1092
2* .0978 37.11 96.25 .2985 2 .9902 1.494 4 .8276
3* .0099 3.75 100.00 .0989
Classification Results Actual No. of Predicted Predicted Predicted Predicted ....Group Cases Membership Membership Membership Membership ............................Prdn. P & I Mktg. Q.C. Prdn. mgrs.... 70 51 (72.9%) 17 (24.3%) 2 (2.9%) 0 (0.0%) P & I mgrs........ 56 21 (37.5%) 33 (58.9%) 2 (3.6%) 0 (0.0%) Mktg. mgrs ....... 17 10 (58.8%) 6 (35.3%) 1 (5.9%) 0 (0.0%) Q.C. mgrs ........ 15 11 (73.3%) 4 (26.7%) 0 (0.0%) 0 (0.0%) ----------------------------------------------------------------------------- Percentage of cases correctly classified: 53.80% _____________________________________________________________________________Continue ... |