Instruments to measure opportunities to satisfy needs, and degree of satisfaction of needs, in the workplace.
This requirement has been met in vocational psychology (e.g. Dawis & Lofquist, 1984; Holland, 1985) and research on Holland's model can be used to exemplify how it has been operationalized. In that research, correspondence of dimensions has been achieved by measuring environments indirectly through people (Prediger & Vansickle, 1992). People in various occupations complete questionnaires which yield their profiles in terms of Holland's hexagon. The Holland profile of people in each occupation is then adopted as the profile for that occupation. Thus, the occupation (environment) is assigned a profile with dimensions which correspond to those used to measure people. One advantage of this approach is that it ensures a close conceptual correspondence between people measures and environment measures. In addition, it provides face-valid information for guiding people to occupations. People are advised to enter occupations populated by people who are like themselves.
Although this approach has its merits, one of its weaknesses is that it assumes the great majority of people are in occupations for which they are well suited. Prediger & Vansickle (1992) capture this point nicely with their suggestion that the Holland profiles of satisfied and dissatisfied people in a given occupation should be compared. They suggest that the profiles should be different. Even if most people do find and remain in suitable occupations, there will always be some error introduced by the minority who are not well placed.
One way to improve research on Holland's (1985) model and other theories of person-environment fit is to directly measure the environments themselves and not the people in them (Bretz & Judge, 1994; Caldwell & O'Reilly, 1990; Davis-Blake & Pfeffer, 1989). Caldwell & O'Reilly (1990) have taken this approach using a Q-sort technique to measure people and environments separately, but on corresponding dimensions, as have Bretz & Judge (1994) using Likert scales.
A second weakness of the indirect approach is that it has not facilitated a coordinated link between occupational guidance and job selection. Thompson, Avery & Carlson (1968) define a job as 'a localized version of an occupation which fixes the practice of the occupation in time and space'. Vocational psychology has developed methods for guiding people into occupations but does not offer much help in guiding people towards specific jobs within those occupations. Given that organizational and other environmental factors can cause considerable variations in how an occupation is practised in different locales, the jobs available within an occupation can vary considerably (Davis-Blake & Pfeffer, 1989). Guiding people to occupations on the basis of occupational profiles does improve their probability of finding suitable jobs over the probability of them finding suitable jobs without such guidance. However, even more congruent placements could be made if the idiosyncrasies of specific jobs were matched with the idiosyncrasies of particular individuals, which would require the direct measurement of those jobs.
The direct measurement of jobs has been given considerable attention by those interested in selection so it is useful to consider the relationship between vocational psychology and selection here. Selection, like vocational psychology, is concerned with person-environment fit. In vocational psychology the environments under consideration are occupations. In selection the environments are jobs. The traditional approach to selection, as described in standard texts such as that of Schneider & Schmitt (1986), begins with a job analysis that leads to a description of the characteristics needed by people in order to do the job effectively. Thus, the characteristics of people are organized along dimensions set by the description of the job. People are then selected for the job on the basis of those characteristics. The usual approach of vocational psychology is to begin with the characteristics of people (e.g. their needs and interests) and then derive measures of environments based upon them. This process is evident in the case of Holland's (1985) model described above. Therefore, the conceptual transition from occupational counselling to job selection is not smooth. The transition must bridge two different conceptual systems which are built around two different descriptor systems, one focused upon jobs, the other upon people.
Conceptual and practical activities involving the transition from vocational guidance to job selection would be greatly facilitated if there were a single conceptual system to encompass both. Such a system would need to include direct and separate measures of people and environments. One way to build such a system would be to extend the conceptual systems from vocational psychology to the description of jobs, including instruments which measure jobs directly using dimensions derived from the description of people. Another approach would be to extend the conceptual systems from selection back through to the description of people, incorporating direct measures of people. Ultimately, some combination of these will probably lead to a conceptual system which amalgamates concepts from both approaches.
In this paper the first of the two approaches described above will be taken. A conceptual system built around human needs for achievement, power and affiliation will be extended to the description and measurement of jobs. In addition, this paper will present instruments to measure worker satisfaction of each of those three needs on the job. These instruments will allow an even more precise evaluation of person-job fit.
The needs for achievement, power and affiliation
The needs for achievement, power and affiliation have been given considerable attention as possible determinants of person-occupation fit. Despite some methodological concerns, it has been shown empirically that levels of these needs predict job satisfaction and competence in a number of occupations, particularly management (Chusmir, 1985; House, 1988; McClelland, 1985; McClelland & Boyatzis, 1982; McClelland & Burnham, 1976; Medcof, 1985, 1990; Medcof & Wall, 1990; Medcof & Wegener, 1992; Stahl, 1986; Steers & Braunstein, 1976). Jackson (1989) describes a person with a high need for achievement as follows: 'Aspires to accomplish difficult tasks; maintains high standards and is willing to work toward distant goals; responds positively to competition; willing to put forth effort to attain excellence'. Jackson (1989) describes a person with a high need for power (dominance) as follows: 'Attempts to control environment, and to influence or direct other people; expresses opinions forcefully; enjoys the role of leader and may assume it spontaneously'. Jackson (1989) describes a person with a high need for affiliation as follows: 'Enjoys being with friends and people in general; accepts people readily; makes efforts to win friendships and maintain associations with people'.
Research on person-occupation fit and the needs for achievement, power and affiliation has generally taken the methodological approach found in the occupational guidance literature. For example, the occupation of manager is usually defined operationally on the basis of the job titles of the persons being studied with no attempt to directly measure the management jobs themselves in a way which makes specific reference to the needs for achievement, power and affiliation. For instance, it is generally proposed that management jobs require people to exercise power more than non-management jobs do, and that managers as a group have a higher need for power than most other workers. This proposition is supported by studies showing that people occupying positions titled 'manager', or one of its variants, have a higher average need for power than people in positions with other titles (e.g. House, 1988; Medcof, 1990).
Although research on these needs has taken the methodological approach of vocational psychology, researchers have been more concerned with selection than with vocational guidance (McClelland, 1985; Medcof & Wegener, 1992; Stahl, 1986; Steers & Braunstein, 1976). As a result, the system for selecting people for narrowly defined environments (jobs) has been developed based on a methodology designed to guide people to broadly defined environments (occupations) with the outcome that the lack of specific direct measures for describing jobs may have weakened the results.
Studies of the needs for achievement, power and affiliation have also lacked measures of the degree to which workers experience satisfaction of the three needs. It is usually assumed that people do jobs well if effective performance of those jobs leads to satisfaction of their needs and that people perform poorly if job performance does not lead to need satisfaction. Although measures of general job satisfaction are sometimes used, the precision of theoretical predictions and empirical verifications could be greatly improved if the satisfaction of specific needs could also be measured.
This paper presents and evaluates instruments that measure opportunities to satisfy needs on the job and the level of satisfaction of those needs. Instruments of high quality (with respect to reliability and validity) should enable more effective research on the role of the needs for achievement, power and affiliation in job selection and vocational guidance. In addition, the instruments will provide an example of the successful extension of a conceptual system based upon the measurement of human needs to the measurement and conceptualization of jobs. Such a demonstration should be of general interest to those concerned with person-environment fit.
The sample consisted of 1155 people working in a variety of jobs. These work groups were made available to the senior author through personal contacts and the contacts of associates. Table 1 lists the various work groups in the sample and the number of respondents in each. The groups were chosen to represent a wide variety of occupations and Table 1 shows that that goal was accomplished, although no procedures were used to ensure that the groups were statistically representative of the working population. Table 2 shows the occupational and demographic characteristics of the sample.
The questionnaire used in this study consisted of several parts. Section One, immediately following the cover page, requested occupational and demographic information. Section Two, on pages 2-5, asked respondents to describe their jobs and included the questions about the degree to which the job offered opportunities to satisfy the needs for achievement, power and affiliation. Section Three, pages 6-10, consisted of questions from the Personality Research Form (PRF), Form E, developed by Jackson (1989), which measures the needs for achievement, power (dominance) and affiliation. Section Four, pages 11-14, asked about satisfaction with a number of aspects of the job including satisfaction of the needs for achievement, power and affiliation.
Opportunities to satisfy needs on the job were measured on five-point scales with anchors ranging from 'mostly inaccurate' to 'mostly accurate'. Respondents indicated how well each of several statements described their own jobs. For example, one statement in the achievement opportunities scale read, 'On this job I can work towards clear, challenging goals'. One statement for power opportunities read, 'On this job I spend a great deal of time getting other people to do things'. One statement for affiliation opportunities read, 'My job is a "people" job'.
Levels of satisfaction of needs were rated on five-point scales with anchors extending from 'dissatisfied' to 'satisfied'. Respondents indicated their satisfaction with various listed aspects of their jobs. For example, one statement for satisfaction of the need for achievement read, 'The number of opportunities I have to take on difficulties at work'. One statement for power satisfaction read, 'The degree to which I am able to organize and direct the activities of others'. One statement for affiliation satisfaction read, 'The degree to which I can talk to others about non-work matters'.
The respondent groups were obtained by initially contacting members of the work groups or their managers. After permission for a group's participation in the study was obtained through appropriate channels, researchers visited the work site to distribute and collect the questionnaires. Since the groups varied considerably (e.g. in work schedules, working conditions, and amount of time allotted for their participation), the procedures for the distribution, completion and return of the questionnaires varied across groups. In some cases there was direct contact between the researcher and respondents and in other cases there was no direct contact. Some respondents filled out the questionnaires during work hours whereas others completed them on their own time.
Table 2. Occupational and demographic characteristics of the sample used in this study
N 1155 Blue collar (%) 6 Managers (%) 26 Professionals (%) 29 Secretarial or clerical (%) 8 Technical (%) 11 Other occupations (%) 20 Male (%) 50 Average number of years in workforce 14.4 Average number of years on present job 6.0
Although it was realized that these variations would probably raise the error variance in the data, this was the only practical way to collect data from such a variety of different groups. The variety in work groups is important for demonstrating the generalizability of the scales. Furthermore, the variety in administration methods reflects expected differences in organizational practices that may impact on scale reliability and validity and thus provides an accurate evaluation based on the intended use of the scales.
Preliminary development of the scales
Questionnaire items were developed using Jackson's (1989) definitions of the needs for achievement, power (dominance) and affiliation as a conceptual guide. Jackson's definitions were adopted because his instrument (the PRF) and the definitions associated with it have been well validated empirically and are widely used (Jackson, 1989). Further, as described above the PRF was used in this study to measure needs, so the use of Jackson's definitions to guide the development of the new instruments provides theoretical consistency among the instruments. Eight questions were developed for each of the six scales and these questions were administered to a sample of about 200 respondents in a variety of jobs. Exploratory factor analyses and reliabilities provided data to eliminate some of these questions. Some new questions were generated, again using Jackson's definitions for conceptual guidance. The final six scales, of eight to 10 questions each, which had some statistical and face validity, were used in this study.
Respondents without missing data were randomly split into two groups of 509 respondents each. The first group was used for the primary analyses and the second as a 'hold back' sample for cross-validation. Exploratory factor analyses and Cronbach's alphas were used on the first sample to select questions for each scale which yielded good scale properties. This included four questions for power opportunities, three for affiliation opportunities, five for achievement opportunities, four for power satisfaction, five for affiliation satisfaction, and five for achievement satisfaction. These questions were used in the confirmatory factor analyses described below.
Confirmatory factor analysis
Confirmatory factor analysis attempts to describe a set of observed variables using a smaller number of factors (Manly, 1986). Conceptually, it is an extension of the general linear model in which an observed score reflects the contribution of a true value and error. Any test score consists of two elements: a contribution from the latent construct tapped by the item and error.
James, Muliak & Brett (1982) recommend that studies assessing the factor structure of items test a number of models and not just that favoured by the researcher. Accordingly, the confirmatory factor analyses done on these data using EQS (Bentler, 1989) evaluated four competing models to determine which fit the data best. Each model was developed to represent possible relationships between the items and their corresponding scales. The first model tested here, and that proposed by the authors to be the most appropriate one, was a six-oblique-factor model which states that the items load on achievement opportunity, power opportunity, affiliation opportunity, achievement satisfaction, power satisfaction and affiliation satisfaction. Support for this model would indicate that the distinction between opportunities and satisfactions for each of the needs (achievement, power and affiliation) are meaningful. Since it is possible that all the items may appear to be related to each other due to common method variance, the second model was based on the possibility that all items loaded on one factor. Support for this model would suggest that none of the hypothesized distinctions was relevant. The third model proposed two oblique factors which reflected the possibility that the scales measured general opportunities and satisfactions rather than three unique types of each. Support for this model would indicate that the distinction between achievement, power and affiliation was not meaningful. The fourth model proposed three oblique factors representing achievement, power and affiliation. Support for this model would indicate that the distinction between satisfactions and opportunities was not meaningful. In summary, the four models represent different possible relationships among the items with increasing differentiation as one moves from the one-factor model to the six-factor model. By comparing the fit indices across these models the best one can be identified.
The covariance matrix from each sample was analysed using the maximum likelihood criterion. Based on the recommendation by Bollen (1989) and Breckler (1990), that multiple indices should be used to evaluate models, four indices were chosen to determine the best fit. These were: chi-square, normed fit index (NFI), non-normed fit index (NNFI), and comparative fit index (CFI).
Bentler (1989) discussed the four indices and their advantages and disadvantages. Although the chi-square test is often affected by large sample sizes, Bentler (1989) suggests that it can provide a good baseline index against which other indices can be evaluated. NFI (Bentler & Bonett, 1980) evaluates each model with higher values indicating a better fit of model and the sample covariance matrix. NNFI is used in the analysis because sample size influences NFI (Bearden, Sharma & Teel, 1982). CFI was computed because Bentler (1990) demonstrated that CFI avoids underestimation in small samples and has less sampling variability than NNFI.
Confirmatory factor analyses
All four fit indices for both samples (primary and cross-validation) are shown in Table 3 and it is seen that the fit indices improve significantly as one moves from the null model to the six correlated factors model for both samples (except between the two-factor and three-factor models for Sample 1), indicating that the six-factor model is the best one to describe the data. The NFI, NNFI and CFI are all above .9 for Sample 1, indicating a good fit (Bentler & Bonett, 1980). Sample 2 shows the same pattern with the only meaningful difference between the two samples being that the NFI fell below .9 for the validation sample, although it was still very close (.856). In short, Table 3 provides support for the existence of six unique yet correlated constructs.
Table 4 shows the factor loadings and error components for the six correlated factors model. Item loadings were moderate to large (ranging from .549 to .849) across the six scales. The proportion of item variance accounted for by each factor loading ranged from 30 to 72 per cent (the square of the factor loading is the proportion of variance accounted for by the loading). There was considerable error variance with the item error components representing 28 to 70 per cent of the variance across the sample. This moderate to high [TABULAR DATA FOR TABLE 3 OMITTED] error variance may be due to the heterogeneous sample and questionnaire administration methods.
The factor intercorrelations above the diagonal in Table 5 reveal considerable correlations between the three opportunity measures (correlations of .58, .53 and .70) and between the three satisfaction scales (. 66, .78 and. 55). These suggest that jobs which provide opportunities to satisfy one need tend to satisfy other needs.
Cronbach's alphas were calculated for the six scales to provide another measure of their psychometric quality. As Table 5 shows, all but two had alphas above .80 and the lowest alpha was .72, indicating that the scales were reliable.
In summary, these data attest to the respectable psychometric properties of these six newly developed scales. Table 3 provides strong evidence for the six correlated factors model. Table 4 shows moderate to high factor loadings of scale items on their respective factors. Table 5 shows that there are intercorrelations among the various scales but that these are not inordinately high. Table 5 also reveals that the scales have respectable reliabilities.
Two general approaches were taken to evaluate scale validities. In the first, the extent to which the correlations between variables are consistent with theoretical assumptions about the relationships between variables was examined. In the second approach, known group validities were evaluated.
Validities and intercorrelations
One fundamental assumption made by researchers in this area is that people gravitate to, and tend to remain in, jobs which fulfil their needs (Chusmir, 1985; House, 1988; McClelland, 1985; Medcof, 1985, 1990). This assumption has provided the basis for the many studies that assess the needs of job incumbents to determine if needs are prerequisites to effective job performance. Although researchers acknowledge that not all people are able to move to, and remain in, jobs which satisfy their needs, it is assumed that as a general rule a significant level of matching of people and jobs does occur.
Table 4. Factor loadings and error components for the six correlated factors model (Sample 1)
N ACHOP POWOP AFFOP SATACH SATPOW SATAFF Error
1 .606 .795 2 .692 .722 3 .740 .673 4 .643 .766 5 .617 .787 6 .668 .744 7 .797 .604 8 .789 .614 9 .683 .730 10 .549 .836 11 .756 .655 12 .766 .643 13 .723 .691 14 .751 .660 15 .765 .644 16 .819 .573 17 .826 .564 18 .849 .528 19 .833 .553 20 .706 .708 21 .711 .703 22 .799 .602 23 .629 .778 24 .800 .599 25 .630 .777 26 .658 .753
Key. ACHOP = opportunity to satisfy the need for achievement; POWOP = opportunity to satisfy the need for power; AFFOP = opportunity to satisfy the need for affiliation; SATACH = satisfaction of the need for achievement; SATPOW = satisfaction of the need for power; SATAFF = satisfaction of the need for affiliation.
This assumption, that people gravitate to jobs which provide opportunities to fulfil their needs, suggests that there should be correlations between needs and opportunities. For example, people with a high need for achievement should be found in jobs with high opportunities to satisfy that need and people with a low need for achievement should be found in jobs with low opportunities to satisfy that need. The correlation matrix in Table 5 shows correlations between needs and opportunities. The correlations of the needs for achievement, power and affiliation with their corresponding job opportunities are .27, .35 and .21, respectively. This is consistent with the assumptions and supports the validity of the scales. These correlations are not very high, which suggests that the person-job matching process is not highly effective.
[TABULAR DATA FOR TABLE 5 OMITTED]
A second basic assumption is that the greater the opportunity to satisfy a particular need on a job, the higher will be the reported level of satisfaction of that need by the incumbents of that job. It follows that there will be positive correlations between corresponding opportunities and satisfactions. The correlations in Table 5 between opportunities to satisfy needs (for achievement, power and affiliation) and their corresponding satisfactions are, respectively, .56, .29 and .26, which provide support for the validity of the scales.
Known group validities
The known group validities were tested by determining how well the opportunity scales discriminated between groups of workers holding jobs which were believed to be, on the basis of earlier research, different in the opportunities they provided to satisfy needs. Tests were done to compare managers, nurses, teachers and professionals.
In the first test of known group validity, opportunities to satisfy needs on the job were assessed to identify if they distinguished between managers and the rest of the sample. Theorists (House, 1988; McClelland, 1985; Medcof, 1990; Stahl, 1986) maintain that management positions provide greater opportunities to satisfy power needs than most other jobs. They also maintain that management positions provide greater opportunities to satisfy the need for achievement, although this difference is not as great as that for power opportunities. A discriminant analysis was done with two groups, one consisting of respondents in this study who identified themselves on the questionnaire as being managers (N = 165), and the second consisting of all other respondents in the sample. The discriminant variables were opportunities to satisfy the needs for achievement, power and affiliation. Table 6 shows that the function discriminated between the two groups ([[Chi].sup.2] = 133.8, p [less than] .001). In the 'Variable statistics' part of the table it is seen that all three types of opportunity discriminated between managers and the other work groups. As predicted, power opportunity was by far the best discriminator (F = 133.0, p [less than] .001), followed by achievement opportunity (F = 15.3, p [less than] .001). The significant effect of affiliation opportunity (F = 5.8, p [less than] .05) was unexpected and will be discussed below. A classification analysis revealed that the discriminant function correctly classified 68 per cent of the sample (Table 7). These data are consistent with the assumption that managerial jobs provide more opportunities to satisfy the needs for power and achievement than jobs in general do, thus supporting the validity of the scales.
Table 6. Discriminant analyses of occupational groups using needs and opportunities as discriminators
Discriminant function statistics
Occupational Eigen- Wilk's Chi- group value lambda square d.f. Sig.
Opportunities Managers .12 .89 133.8 3 [less than].001 Nurses .07 .93 83.0 3 [less than].001 Teachers .12 .89 136.2 3 [less than].001 Managers and .22 .82 98.3 3 [less than].001 professionals
Needs Managers .02 .98 25.7 3 [less than].001 Nurses .04 .96 50.9 3 [less than].001 Teachers .01 .99 6.5 3 n.s. Managers and .06 .94 28.8 3 [less than].001 professionals
Occupational Wilk's group Variable lambda F Sig.
Opportunities Managers ACHOP .99 15.3 [less than].001 POWOP .90 133.0 [less than].001 AFFOP .99 5.8 [less than].05
Nurses ACHOP .99 5.2 [less than].05 POWOP .99 5.4 [less than].05 AFFOP .96 46.5 [less than].001
Teachers ACHOP .96 46.5 [less than].001 POWOP .90 130.8 [less than].001 AFFOP .95 57.7 [less than].001
Managers and ACHOP .99 0.0 n.s. professionals POWOP .86 82.8 [less than].001 AFFOP .99 0.1 n.s.
Needs Managers NACH .99 9.7 [less than].01 NPOW .98 22.0 [less than].001 NAFF .99 0.3 n.s.
Nurses NACH .99 3.9 [less than].05 NPOW .99 17.7 [less than].001 NAFF .98 23.7 [less than].001
Teachers NACH .99 4.0 [less than].05 NPOW 1.00 0.0 n.s. NAFF .99 2.1 n.s. Managers and NACH .99 6.0 [less than].05 professionals NPOW .96 22.9 [less than].001 NAFF .99 2.0 n.s.
In the second test of known group validity nurses were compared to the rest of the sample. Theorists (Chusmir, 1985; Medcof & Wegener, 1992; Stahl, 1986) maintain that nursing jobs provide greater opportunities to satisfy affiliation needs than most other jobs. Therefore, research shows that nurses as a group have a higher than average need for affiliation. A discriminant analysis was done with two groups, one consisting of all of the nurses in the sample, identified by their work groups shown in Table 1 (N = 147), the second consisting of all other respondents in the sample. The discriminant variables were opportunities to satisfy the needs for achievement, power and affiliation. Table 6 shows that the function discriminated between the two groups ([[Chi].sup.2] = 83.0, p [less than] .001) as did all three types of opportunity ('Variable statistics' part of the table). Affiliation opportunity (F = 46.5, p [less than] .001) was the strongest discriminator, as predicted. Achievement opportunity (F = 5.2, p [less than] .05) and power opportunity (F = 5.4,p [less than] .05) were of about equal, relatively weak, strength as discriminators. Classification analysis showed that the discriminant function correctly classified 68 per cent of the sample (Table 7). These data show that nursing is distinguished from other occupations primarily by affiliation opportunities.
Table 7. Classification analyses of occupations using job opportunities and needs to classify (Percentages in parentheses)
Predicted group membership % No. of Correctly Groups cases 1 2 grouped
1. Managers 165 129 (78) 36 (22) 68 2. All others 1001 340 (34) 661 (66)
1. Nurses 172 121 (70) 51 (30) 68 2. All others 1030 338 (33) 692 (67)
1. Teachers 102 86 (84) 16 (16) 70 2. All others 1100 343 (31) 757 (69)
1. Managers 165 131 (79) 34 (21) 69 2. Professionals 339 122 (36) 217 (64)
1. Managers 165 99 (60) 66 (40) 56 2. All others 1001 447 (45) 554 (55)
1. Nurses 172 113 (66) 59 (34) 62 2. All others 1030 396 (38) 634 (62)
1. Teachers 102 57 (56) 45 (44) 58 2. All others 1100 465 (42) 635 (58)
1. Managers 165 103 (62) 62 (38) 60 2. Professionals 339 142 (42) 197 (58)
In the third test of known group validity, opportunities to fulfil needs on the job were used to discriminate between teachers and the rest of the sample. Medcof (1985) proposed that teaching jobs have greater opportunities to satisfy power needs than most other jobs because in the classroom teachers are generally in a controlling role. A discriminant analysis was done with two groups, one consisting of all of the teachers in the sample, identified by their work groups shown in Table 1 (N = 102), and the second consisting of all other respondents in the sample. The discriminant variables were opportunities to satisfy the needs for achievement, power and affiliation. Table 6 shows that the function discriminated between the groups ([[Chi].sup.2] = 136.2, p [less than] .001) as did all three types of opportunity. Power opportunity was by far the best discriminator (F = 130.8, p [less than] .001), as predicted. Affiliation opportunity (F = 57.7, p [less than] .001) was the second best discriminator and it was marginally stronger than achievement opportunity (F = 46.5, p [less than] .001). A classification analysis showed the discriminant function correctly classified 70 per cent of the sample (Table 7). These data show that the primary distinguisher of teaching from other work is power opportunity.
In the fourth test of known group validity, managers were compared to professionals. According to Stahl (1986), managers have more opportunities to satisfy power needs than professionals, but do not differ from them in achievement opportunities. This was tested with a discriminant analysis in which one group consisted of those who identified themselves as managers on the questionnaire (N = 165) and the other group consisted of those who identified themselves as professionals (N = 339). The results are shown in Table 6. The function discriminated between the groups and, as predicted, only power opportunity (F = 82.8, p [less than] .001) significantly distinguished the two groups. The function correctly classified 69 per cent of the sample (Table 7).
In summary, evaluations of the known group validities of the opportunities scales were supportive. The data are consistent with theoretical assumptions about the opportunities that most strongly distinguish the occupations of management, teaching and nursing, and distinguish between management and the professions. Some unexpected significant discriminators, albeit of much smaller magnitude than the predicted ones, will be discussed below.
The strengths of opportunities and needs as discriminators
An important assumption in many studies of the needs for achievement, power and affiliation has been that the measured needs of workers can be used as surrogates for direct measurements of opportunities to satisfy those needs on the job. For example, it has been assumed that since managers as a group have a higher than average need for power, management as a profession provides greater than average opportunities to satisfy the need for power.
The premise of this paper is that direct measures of opportunities are better discriminators of occupations than measures of incumbents' needs. This was checked by comparing the sensitivity of the two kinds of measures as discriminators of occupations by repeating the discriminant analyses described above but with the three needs as the discriminators rather than the opportunities. The results of the discriminant analyses using needs can be compared to the results obtained using opportunities (see Tables 6 and 7).
The results of these comparisons support the assumption that opportunities are better discriminators than needs. When distinguishing managers from the rest of the population, the chi-square for the opportunities function was 133.8, that for needs was 25.7. For nurses, the opportunities chi-square was 83.0, for needs 50.9. For teachers, the opportunities chi-square was 136.2, for needs 6.5. In the comparison between managers and professionals, the opportunities chi-square was 98.3, for needs 28.8. When the numbers of cases correctly classified are compared it was also found that opportunities fare better than needs. For managers, the opportunities discriminant function correctly classified 68 per cent of the cases while that with needs correctly classified 56 per cent. For nurses, opportunities correctly classified 68 per cent, needs 62 per cent. For teachers, opportunities correctly classified 70 per cent, needs 58 per cent. For managers and professionals, opportunities correctly classified 69 per cent, needs 60 per cent. These results support the assumption that needs measurement is an imprecise surrogate for direct measures of opportunities.
Overall, there was support for the six scales. Confirmatory factor analyses favoured the six-oblique-factors model. The statistical properties of those six factors were good despite the moderate to high amounts of unexplained variance. The unexplained variance may be due to sample and data collection variability. Cronbach's alphas of the six scales were adequate. Moderate correlations among opportunities scales and among satisfaction scales indicated that jobs high in opportunities to satisfy one need tended to be high for others. Correlations among the scales followed patterns consistent with theoretical assumptions, since, as predicted, there are positive correlations between needs and corresponding job opportunities and between job opportunities and the corresponding satisfactions. Tests of known group validities of the scales were supportive in that the strongest discriminators of occupational groups were consistent with predictions based on the literature.
Some of the unexpected findings, such as that power, achievement and affiliation opportunities all discriminated teaching from other occupations, when only power was expected to, may be explained by the increased sensitivity of the scales. In other words, opportunities may be more sensitive discriminators among occupations than needs. The predictions made in this paper about which opportunities should discriminate between occupations were based upon past studies. In those past studies, only the needs of job incumbents, not job opportunities, were measured. As a result of being more sensitive the instruments measuring job opportunities used in the present study may have picked up differences among occupations which the less sensitive measures of needs in past studies were unable to detect.
These more sensitive instruments open up wide avenues of research involving the needs for achievement, power and affiliation. Past studies which measured needs can be replicated using these new opportunity and satisfaction measures to verify past findings, and to search for new, more subtle, differences among occupations. The possibility of using the new methodology for selection should also be explored. By combining the use of need, opportunity and satisfaction measures, quite precise evaluations of particular job placements are now possible.
It is now also possible to evaluate vocational guidance and job selection activities which are based upon a single conceptual framework. Vocational guidance can now be based upon occupations characterized, not by the needs profiles of occupation incumbents, but by the opportunities profiles of samples of jobs within those occupations. Those opportunities profiles could be based upon direct measures of the jobs using the instruments presented in this paper. Once an occupation is selected through vocational guidance, job selection could take place by matching applicants' needs profiles to the opportunities profiles of the jobs being considered for placement. Subsequently, the satisfaction questionnaires could be used to evaluate the fit.
Despite these benefits, needs profiles and job opportunities alone are unlikely to be sufficient to provide effective occupational guidance and job placement. Bretz & Judge (1994) point out that person-job fit must consider not only needs but also knowledge, skills and abilities, value orientations and personality. Combining the tools based upon needs that are presented here with tools for these other characteristics is a research challenge which should give both theoretical and practical yields.
This study has provided promising support for the opportunities and satisfactions scales but further evaluation of them is needed. The sample used in this study was large and diverse but it was a sample of convenience. This study should be replicated using more systematically identified and randomized samples. This study also did not evaluate the stability of the measures over time. Future research should address this issue. Finally, with the preliminary support for these scales that this study provides, future research should focus on expanding the associated nomological network with respect to possible covariates of needs profiles, such as gender, ethnicity and age.
This study joins those of Caldwell & O'Reilly (1990) and Bretz & Judge (1994) as a demonstration of the successful independent measurement of jobs and people for the purpose of person-job fit. Such demonstrations suggest that the activities of vocational guidance and job selection can be brought closer together as partner activities in the improvement of person-environment fit. Further development of Bretz & Judge's (1994) work with the Theory of Work Adjustment (Dawis & Lofquist, 1984), and the application of this approach within the context of Holland's (1985) model, should provide a way to merge the best elements of selection and vocational guidance.
The authors thank Rick Hackett, Simon Taggar and anonymous reviewers for their helpful comments on earlier versions of this paper. This research was partially funded by the Arts Research Board of McMaster University. For copies of this paper and normative data for the scales please contact John W. Medcof.
Bearden, W. O., Sharma, S. & Teel, J. E. (1982). Sample size effects on chi-square and other statistics used in evaluating causal models. Journal of Marketing Research, 19, 425-430.
Bentler, P. M. (1989). EQS: Structural Equations Program Manual. Los Angeles, CA: BMDP Statistical Software.
Bentler, P. M. (1990). Comparative fit indices in structural models. Psychological Bulletin, 107, 238-246.
Bentler, P. M. & Bonett, D. G. (1980). Significance tests and goodness of fit in the analysis of covariance structures. Psychological Bulletin, 88, 588-606.
Bollen, K. A. (1989). Structural Equations with Latent Variables. New York: Wiley.
Breckler, S. J. (1990). Applications of covariance structure modelling in psychology: Cause for concern? Psychological Bulletin, 107, 260-273.
Bretz, R. D. & Judge, T. A. (1994). Person-organization fit and the Theory of Work Adjustment: Implications for satisfaction, tenure and career success. Journal of Vocational Behaviour, 44, 32-54.
Caldwell, D. C. & O'Reilly, C. A. (1990). Measuring person-job fit with a profile-comparison process. Journal of Applied Psychology, 75, 648-657.
Chusmir, L. H. (1985). Matching Individuals to Jobs. New York: Amacom Books.
Davis-Blake, A. & Pfeffer, J. (1989). Just a mirage: The search for dispositional effects in organizational research. Academy of Management Review, 14, 385-400.
Dawis, R. V. & Lofquist, L. H. (1984). A Psychology of Work Adjustment. Minneapolis, MN: University of Minnesota Press.
Holland, J. L. (1985). Making Vocational Choices, 2nd ed. Englewood Cliffs, NJ: Prentice-Hall.
House, R. J. (1988). Power and personality in complex organizations. In B. Staw & L. Cummings (Eds), Research in Organizational Behaviour, 10, 305-357.
Jackson, D. N. (1989). Personality Research Form Manual, 3rd ed. Port Huron, MI: Research Psychologists Press, Inc.
James, L. R., Muliak, S. A. & Brett, J. M. (1982). Causal Analysis: Assumptions, Models and Data. Beverly Hills, CA: Sage.
McClelland, D. C. (1985). Human Motivation. Glenview, IL: Scott Foresman.
McClelland, D. C. & Boyatzis, R. E. (1982). Leadership motive pattern and long term success in management. Journal of Applied Psychology, 67, 737-743.
McClelland, D. C. & Burnham, D. H. (1976). Power is the great motivator. Harvard Business Review. March-April, 100-110, 159-166.
Manly, B. F. J. (1986). Multivariate Statistical Methods: A Primer. New York: Chapman and Hall.
Medcof, J. W. (1985). The power motive and organizational structure: A micro-macro connection. Canadian Journal of Administrative Sciences, 2(1), 95-113.
Medcof, J. W. (1990). The need for dominance scale of the Manifest Needs Questionnaire: Its reliability and validity. Applied Psychology: An International Review, 39, 307-322.
Medcof, J. W. & Wall, R. W. (1990). Work technology and the motive profiles of nurses. Canadian Journal of Nursing Research, 22(3), 51-65.
Medcof, J. W. & Wegener, J. G. (1992). Work technology and the needs for achievement and nurturance among nurses. Journal of Organizational Behaviour, 13, 413-423.
Osipow, S. H. (1987). Applying person-environment fit theory to vocational behaviour. Journal of Vocational Behaviour, 31,333-336.
Prediger, D. L. & Vansickle, T. R. (1992). Locating occupations on Holland's hexagon: Beyond RIASEC. Journal of Vocational Behaviour, 40, 111-128.
Schneider, B. & Schmitt, N. (1986). Staffing Organizations, 2nd ed. Glenview IL: Scott Foresman.
Stahl, M. J. (1986). Managerial and Technical Motivation, New York: Praeger.
Steers, R. M. & Braunstein, D. N. (1976). A behaviourally based measure of manifest needs in work settings. Journal of Vocational Behaviour. 9, 251-266.
Thompson, J., Avery, R. & Carlson, R. (1968). Occupations, personnel, and careers. Education Administration Quarterly, 4, 6-31.
RELATED ARTICLE: Table 1. Work groups included in this study
1. Full-time workers in a university book store (25) 2. Part-time workers in a university book store (25) 3. Nurses in a cancer clinic (16) 4. Radiography technicians in a hospital (18) 5. Police constables with a city police force (16) 6. Police above the rank of constable with a city police force (18) 7. Professional engineers in an aerospace company (41) 8. Supervisors of technicians in an aerospace company (44) 9. Salaried, white-collar workers in a liquid air company (19) 10. Hourly, blue-collar workers in a liquid air company (19) 11. Student volunteers at a hospital (16) 12. Adult volunteers at a hospital (19) 13. Long-haul truck drivers (22) 14. Office support staff in a trucking company (22) 15. High school teachers (25) 16. Community college professors (25) 17. Public health nurses (20) 18. Staff nurses working in a hospital on various wards (25) 19. Staff nurses working in a hospital on various wards (26) 20. Head nurses working in a hospital (18) 21. Salespersons in a retail computer store (25) 22. Managers in a retail computer company (25) 23. Managers and professionals in a manufacturing company (20) 24. Non-managerial workers in a manufacturing company (20) 25. University student nurses with RNs and nursing work experience (22) 26. University student nurses without RNs or work experience (21) 27. Elementary school teachers (21) 28. Secondary school teachers (31) 29. Elementary and secondary school teachers (25) 30. Elementary and secondary school principals and vice-principals (25) 31. White-collar workers in a machinery parts manufacturer (26) 33. Managers in a restaurant chain (24) 34. Non-managerial, full-time workers in a restaurant chain (25) 35. Marketing managers in a steel company (22) 36. Production managers in a steel company (20) 37. Salaried life insurance agents (26) 38. Insurance brokers on commission (24) 39. Lifeguards for city swimming pools (29) 40. Supervisors of lifeguards for city swimming pools (25) 41. Hospital volunteers (21) 42. Administrative staff in a hospital (20) 43. Managers and professionals in a purchase refinery (31) 44. Technicians and clerks in a petroleum refinery (67) 45. Nurses on medical-surgical wards (20) 46. Radiotherapy technologist (20) 47. White-collar workers in a hydro plant (14) 48. Blue-collar workers in a hydro plant (16)
Note: Numbers of parentheses are the number of people in each group.
|Printer friendly Cite/link Email Feedback|
|Author:||Medcof, John W.; Hausdorf, Peter A.|
|Publication:||Journal of Occupational and Organizational Psychology|
|Date:||Sep 1, 1995|
|Previous Article:||Effect of unemployment on school leavers' self-esteem.|
|Next Article:||Achievement motivation and female entrepreneurs.|