Abstract
Since its introduction, Schein’s Career Orientations Inventory (COI) has been widely used to assess and advise individuals who are making both initial and specialty career choices. Employed by both practitioners and researchers, the COI was designed to identify the person’s “career anchors,” also known as their career identities. Individuals are expected to have the highest job satisfaction and performance when their job specialization matches their career anchors. However, little published research has evaluated the psychometric properties of the COI. To address this gap, we conducted an exploratory factor analysis, calculated the subscale reliabilities, and correlated the career anchors with Holland’s career interests to examine its convergent and discriminant validities. In this article, we report on the psychometric properties of the COI based on our sample of 165 midcareer college graduates and make several suggestions on the use and interpretation of the COI.
Vocational research helps to enhance individuals’ understanding of their educational and career options and may lead to better career planning and eventually a more personally meaningful life (Flum & Blustein, 2000). A number of theories elucidating the psychological processes underlying individuals’ career decision making have been developed, and these theories have contributed immensely to the field of vocational psychology and career counseling (Kidd, 2006). Traditionally, research in this area has focused heavily on studying the initial career choices of high school and college students, and there are many career scales developed to assess young adults’ readiness to make decisions (Betz & Borgen, 2000), or to classify them into various career interests (Betz, 1991; Holland, 1973). However, as Leong and Geisler-Brenstein (1991) noted, for several occupational fields such as business, medicine, or psychology, there is a need for individuals to choose an area of specialty after entering their profession. For instance, a business student could major in finance, marketing, human resource management, and so on, while a doctor could choose to be a neurosurgeon, gynecologist, and so on. In such cases, individuals face career decisions beyond their initial career choices, and there has been little evidence indicating that such specialty decisions are easier than initial career choices. Consequently, several researchers have developed career measures to assess how individuals make specialty choices throughout the span of their careers.
The Careers Orientations Inventory (COI) was developed by Schein (1985) and has been widely used in vocational research to study how individuals make career specialty choices (Beck & La Lopa, 2001; Brunch, 1998; DeLong, 1982). The COI identifies an individual’s career anchor, and are made up of one’s career related values, abilities, interests, motives, and needs (Schein, 1985). Like personality, career anchors are stable, and serve as a guide for individuals when they make specialty career choices (thus termed as anchor). The underlying hypothesis is that individuals would be most satisfied and productive when there is a fit between their career anchors and their jobs (Schein, 1985). The COI has been used frequently in vocational research in the past 30 years (Beck & La Lopa, 2001; Brunch, 1998; DeLong, 1982; Igbaria, Kassicieh, & Silver, 1999), and applied in numerous cultural settings such as United States (Beck & La Lopa, 2001; Igbaria et al., 1999), Israel (Danziger & Valency, 2006), and China (Wong, 2007). The COI has been used to assess specialty choices of individuals after they have entered a field, such as how some individuals choose to be take on managerial positions while some prefer a more technical role in the same company or industry (Erdogmus, 2004; Igbaria & McCloskey, 1996; Petroni, 2000).
The COI has thus been widely used, and has been reviewed in the mental measurements yearbook, which evaluated it to be an excellent diagnostic tool to help individuals with midcareer specialization decisions (Brunch, 1998; Robertson, 1998). However, the reviewers also noted that more research should be conducted to assess the validity and reliability of the COI because it was initially developed through qualitative interviewing, and the sample used for that was a restricted sample made up of well-educated men (Brunch, 1998; Robertson, 1998). In response to this, our article seeks to evaluate the psychometric properties of the COI. In particular, we conducted exploratory factor analysis to examine if items are loading on their respective subscales, tabulated the internal reliabilities, and assessed the convergent and discriminant validities. The validation of the COI is important to help determine the level of confidence individuals and the field, including practitioners and researchers, can have in its utilization.
Background of COI
According to Schein (2006), one’s career anchor refers to one’s career self-concept made up of one’s perceived career-related talents and abilities, basic values, and motives and needs. The formation of these anchors is dynamic, and anchors are molded by one’s life and career experiences. However, once formed, they are usually stable and could guide career decisions of individuals (thus termed as anchor). Schein (1978) also noted that helping people identify their career anchors would help them clarify their preferences and assist them to make career choices more wisely. Therefore, to enable the use of career anchors in vocational research and counseling, Schein (1985) developed the COI.
Over a 10-year period beginning from 1961, Schein administered periodical questionnaires to management graduates from the Sloan School of Management of Massachusetts Institute of Technology. In these questionnaires, he requested participants to identify milestones in their careers, to discuss reasons for those events, and to report their feelings about them. From their responses, Schein (1985) observed that individuals showed consistent patterns in their career choices and their feelings toward the choices they made. These patterns were shaped by the educational, career, and life experiences of individuals, and guided individuals during their career decision making. In other words, they served as individuals’ career anchors (Schein, 1985). From the responses, Schein identified nine career anchors—technical competence, general managerial competence, autonomy, job tenure security, geographical security, entrepreneurship/creativity, service to a cause, pure challenge, and lifestyle integration. He also observed that individuals were happiest and most productive when they worked in an area consistent with their career anchors, for instance when someone with a general managerial competence was most satisfied and productive when he was in a managerial position. Anchors need not be mutually exclusive within individuals; for example, it is possible for an individual to value both technical competence and general managerial competence in his or her work. However, an individual’s career anchor refers to the single dimension which he or she perceives as the most important of the nine, and on which an individual bases his or her career decisions primarily on (Schein, 1985). The sections below describe the nine anchors and characteristics associated with each of them.
Technical Competence
Individuals whose career anchor is technical competence are primarily concerned with exercising, developing, and displaying their technical skills on their jobs, such as engineering or accounting proficiency (Schein, 1985). Doing so enables them to derive satisfaction and to reinforce their sense of identity. These individuals are happiest and perform best on jobs that challenge them to hone their expertise and would lose interest if their jobs do not pose any potential for them to develop professionally.
General Managerial Competence
Individuals with the general managerial competence career anchor have the desire to take on leadership roles in their organizations (Schein, 1985). These individuals are interested in utilizing interpersonal skills to coordinate and integrate the efforts of others toward a larger goal (Schein, 1985).
Autonomy
The autonomy anchor is characteristic of individuals who emphasize the importance of maintaining independence and asserting control on their jobs (Schein, 1985). They are most satisfied when they have the freedom to decide when to work, what to work on, and how hard to work. While much existing research has demonstrated that all individuals need some degree of autonomy to perform well and feel satisfied (Judge, Thoresen, Bono, & Patton, 2001; Weaver, 1977), individuals with autonomy anchor perceive autonomy as a central factor influencing their career decisions, and might turn down advancement opportunities to retain their right and say on their jobs (Schein, 1985).
Job Tenure Security
The main concern of individuals with anchor in job tenure security is to derive stability from their jobs (Schein, 1985). Their key objectives include having employment security, consistent financial rewards, and good pension and retirement plans. These individuals prefer jobs that could guarantee them stable long-term employment.
Geographical Stability
Individuals with geographical stability anchor deem it essential to stay settled at a specific geographical location (Schein, 1985). These individuals would forgo advancement or career opportunities which require them to relocate.
Entrepreneurship/Creativity
Under this entrepreneurship/creativity anchor, individuals are concerned with creating new ideas and products on their jobs (Schein, 1985). These individuals have a strong desire to extend their identities through the ideas and products they have generated or invented.
Service to a Cause
Individuals with an anchor in service to a cause are committed to making the world a better place through their jobs (Schein, 1985). They would pursue career options that resonate with their personal values on how to serve the world.
Pure Challenge
The pure challenge anchor is typical in individuals who enjoy tackling challenging problems on their jobs (Schein, 1985). They are focused on overcoming problems and/or winning opponents. They perceive jobs that do not contain challenges or competition as boring.
Lifestyle Integration
Individuals with a lifestyle integration anchor want to maintain a balance between various domains of their lives, such as family, leisure, and work (Schein, 1985). They are inclined to take on jobs that allow them to achieve this equilibrium, where each domain does not interfere or disrupt the other domains.
Use of COI in Existing Research and Practice
Career anchors provide scholars and organizations with insights into how to motivate employees. For instance, an employee with the geographical security anchor would not be motivated by an opportunity that would require him or her to relocate to another place. In addition, according to Schein (1985), career anchors are relatively stable across the life span of individuals, and are thus reliable predictors of job satisfaction and performance of individuals.
Several existing studies have used the COI for the above purposes. DeLong (1987) administered the COI to individuals who entered the teaching profession and later made specialty choices within the field. He found that individuals who went into teaching reported a variety of career anchors and assigned different weights to these anchors. He then discussed how individuals’ career anchors could inform schools of how to design and assign specialty responsibilities which would help result in better performance and higher satisfaction in these individuals. For instance, teachers with a technical competence anchor would derive greater satisfaction from exhibiting their expertise in a subject, and could thus be assigned a heavier teaching load, while those with the anchor general managerial competence could be designated to be mentors to junior teachers or students.
In another study, Igbaria, Kassicicieh, and Silver (1999) applied the COI to study career success among research, development, and engineering (RD&E) professionals. RD&E professionals reported diverse career anchors, and correlations between their career anchors and other variables showed that different career anchors were associated with unique antecedents (such as job involvement and job discretion) and outcomes (such as organizational commitment, intention to stay, and career satisfaction). For example, scores on geographical security anchor predicted intention to stay on the job (r = .28, p < .05), and scores on pure challenge anchor was positively associated with job involvement (r = .34, p < .01; Igbaria et al., 1999).
Beck and La Lopa (2001) used the COI to examine the careers of U.S. Hotel Executive Operating Committee (EOC) members. They found that EOC members in different disciplines (such as marketing, food and beverages [F&B], and rooms and operations) had different career anchors. Those who scored highest on the technical competence anchor were most likely to be in marketing than rooms/operations and F&B (F = 20.73, p < .001), while those with general managerial competence were more likely to be in F&B and rooms/operations (F = 11.84, p < .001). They explained that this was because marketing positions required and enabled employees to exercise specialized technical skills such as designing and publicizing, while rooms/operations or F&B positions allowed employees to exercise coordination and integration skills, and hone their proficiency to become general managers. These findings illustrate how career anchors are associated with specialty choices within a single industry.
More importantly, these studies were conducted at different time points across three decades since the introduction of COI, and across varying professions and cultures, thus providing some evidence that the COI has general utility as an instrument in vocational research. However, it should be reiterated that the COI was initially developed by Schein (1985) based on face validity from individuals’ interview responses, and its psychometric properties have not been fully assessed. As a measure that is frequently used in research and practice, as has been noted by other researchers (Beck & La Lopa, 2001; Feldman & Bolino, 1996), it is important to assess the validity and reliability of the COI in greater depth. To date, only limited validation studies of the COI have been published, and these studies do not provide a comprehensive and adequate psychometric evaluation of the COI. For example, Beck and La Lopa (2001) found the internal reliabilities of the anchors to be relatively low (in the .30 s). They also examined the intercorrelations between the anchors and found a positive correlation between the autonomy and entrepreneurship/creativity anchors (r = .27, p < .01), and negative correlations between entrepreneurship/creativity and job tenure security anchors (r = −.18, p < .01), and between technical competence and general managerial competence anchors (r = −.20, p < .01). The studies that came closest to a comprehensive psychometric examination of the COI were the studies by Nordvik (1996), as has been noted by other researchers (Beck & La Lopa, 2001; Feldman & Bolino, 1996), and by Danziger, Rachman-Moore, and Valency (2008). To examine the convergent and discriminant validities associated with COI, Nordvik (1996) tested the relationship between career anchors, Myers-Brigg’s personality types, and Holland’s (1973) typology of career interests. Nordvik (1996) developed an ipsative response scale for the COI where respondents had to choose only one of the four anchor items presented simultaneously to them in seven forced-choice questions instead of rating each item on a Likert-type scale. Respondents were then scored on the anchors based on the frequency they selected the anchor item. From his regression results, Nordvik found that there were no relations between the career anchors and the Myer-Briggs personality types.
More recently, Danziger et al. (2008) attempted a psychometric evaluation of the COI using a 6-point Likert-type response scale. Their confirmatory factor analysis showed adequate support for a nine-factor model. However, their supported nine-factor model was not the same as the original one proposed by Schein (1985). In Danziger et al.’s (2008) model, job tenure security and geographical security anchors were combined, but the entrepreneurship/creativity anchor split into separate entrepreneurship and creativity anchors. Doing so provided a better fit than the original structure (Danziger, Rachman-Moore, & Valency, 2008). They also assessed interitem correlations between items, and verified that items correlated highly with other items in the same anchor, and poorly with items from other anchors. Such results provided support for the subscales’ internal reliabilities.
Shortcomings of Existing Psychometric Studies and Contributions of Current Study
Both Nordvik’s (1996) and Danziger et al.’s (2008) studies have contributed to our understanding of the COI, but there are limitations to their studies which call for a more comprehensive study to adequately evaluate the psychometric properties of the COI. First, most research with COI used a Likert-type response scale but Nordvik (1996) used an ipsative scale. Utilizing different response scales can impact factor analytic results (Tinsley & Tinsley, 1987), hence Nordvik’s study might not be the most applicable to verify the COI’s validity and reliability. Also, the use of an ipsative scale might distort correlations of the COI with other variables since rating on each item was dependent on other items presented (Tinsley & Tinsley, 1987). Such nonindependence between items could influence psychometric findings associated with the scale (Tinsley & Tinsley, 1987). Therefore, in order to make psychometric findings generalizable to existing studies that used the COI with a Likert-type response scale, we would be using a Likert-type scale for the COI during data collection in our study.
Also, Danziger et al. (2008) reported interitem correlations as an examination of convergent and divergent validities of COI. They concluded that high interitem correlations between items of the same anchor and low interitems correlations between items of different anchors illustrated convergent and discriminant validities. However, this would be more of a test of internal reliabilities of the COI anchors than an examination of convergent and discriminant validities of the COI. To assess the convergent and discriminant validities of the COI, we would examine how the COI anchors relate to another established career scale: Holland’s (1973) typology of six career interests, realistic, investigative, artistic, social, enterprising, and conventional. According to Holland, these career interests are expressions of individuals’ personalities, and individuals would choose occupations that fall into the same category as their personality types. Although the career interests are primarily used to predict individuals’ initial career choices and not specialty choices, they are associated to the career anchors because both the anchors and interests are stable constructs stemming from individuals’ career personalities and identities (Holland, 1973); hence, we contend that the career interests could provide us with insights into the convergent and discriminant validities of COI.
In addition, we collected data from middle-aged men who have had a substantial amount of work experience. They represent an ideal sample for examining the validity of career anchors because most, if not all, of them would have had the experiences of having to make specialty choices at some points during their careers. Consolidating these points, our study would provide a more comprehensive and accurate psychometric evaluation of the COI.
Method
Respondents
One hundred and sixty-five men from Dartmouth College’s graduating class of 1962 responded to our survey during a college reunion workshop on midlife career issues. Their ages ranged from 45 to 58, with a mean of 46.2 (SD = 1.1). All of them were males, 3.0% were single, 68.7% were married, 9.8% were divorced, and 17.7% were remarried. The number of jobs held by the men ranged from 0 to 20, with a mean of 4.7 (SD = 3.7). The number of career changes ranged from 0 to 10, with a mean of 1.0 (SD = .8). Their average tenure at their current jobs was 8.6 years (SD = 6.5). They worked in a variety of occupational settings, with 47.2% in small organizations, 28.2% in medium organizations, and 24.5% in large organizations.
Procedure
Respondents received a questionnaire by mail as part of preworkshop assessment. The questionnaire asked for demographic information such as age and marital status, and number of and tenure on past and present jobs. The questionnaire also contained career anchors, career interests, and job satisfaction. After completing the questionnaire, respondents returned it by mail. At the workshop, respondents received their individual scores on the various measures. Three guest speakers with expertise in vocational research and counseling gave presentations to assist respondents to interpret and understand their scores, which were used as a basis for discussion of midlife career issues during the workshop.
Measures
Career Anchors
Respondents’ career anchors were assessed by the 41-item COI (Schein, 1985). Respondents rated their perceived importance on the first 21 items on a 10-point scale (0 = of no importance; 10 = centrally important). A sample item included, “Being able to use my skills and talents in the service of an important cause.” For Items 22–41, respondents rated how each statement was true of themselves on a 10-point scale as well (0 = not at all true; 10 = completely true). A sample item was, “I will accept a management position only if it is in my area of expertise.” Of the nine anchors, seven anchors (technical competence, general managerial competence, autonomy, entrepreneurship, service, pure challenge, and lifestyle integration) contained 5 items each, and two anchors (job tenure security and geographic security) contained 3 items each. Items from all anchors were randomly ordered, as per Schein’s (1985) format. Reliabilities of the anchors would be discussed in detail under the Results section.
Career Interests
We assessed respondents’ career interests using Form B of Holland’s (1973) Vocational Preference Inventory (VPI). The six career interest dimensions are realistic, investigative, artistic, social, enterprising, and conventional. The VPI consists of 42 items that are occupational titles. Respondents responded whether they liked, disliked, or were undecided about each occupation. The occupations listed are associated with one of the six interest dimensions. Respondents’ scores on each interest dimension increased when they indicated liking for each occupational title associated with that interest dimension. Respondents’ scores on the career interest dimensions were used to test the hypotheses on convergent and discriminant validities of the COI. For our sample, Cronbach’s αs for the interest dimensions were adequate, at .78 for realistic, .85 for investigative, .85 for artistic, .82 for social, .76 for enterprising, and .68 for conventional.
Results
To assess the psychometrics properties of COI, we conducted an exploratory factor analysis to examine the factor structure of the scale. An exploratory factor analysis was conducted instead of a confirmatory factor analysis because there was insufficient prior psychometric validation of the COI as discussed earlier and hence a lack of empirical support for Holland’s theorized nine-factor structure. Therefore, it was necessary to empirically explore instead of confirming the COI’s factor structure in this article. We also tabulated Cronbach’s reliabilities to evaluate the internal reliability for each anchor, and used Pearson’s correlations between the COI and Holland’s (1973) VPI, and between COI and the job satisfaction measures to test the hypothesized convergent and discriminant validities of the anchors.
Exploratory Factor Analysis
The 41-item COI yielded .76 on Kaiser–Meyer–Olkin measure of sampling adequacy (MSA). According to Kaiser (1970), an MSA above .60 suggests adequate common variance among the items and justifies conducting a factor analysis. Using principal axis factoring extraction, we found 11 factors with Eigenvalues of above 1.00. We also conducted parallel analysis to help determine how many factors to extract. Results revealed eight factors with raw data Eigenvalues higher than 95th percentile raw data permutation Eigenvalues. Since this was an exploratory factor analysis, we decided to adopt a liberal approach and extracted 11 factors. These 11 factors accounted for 72.49% of the variance in the items. Table 1 documents the factor loadings of items sorted by their sizes of loading after Varimax with Kaiser rotation. Following the advice of Kim and Mueller (1978), we chose orthogonal (varimax) rotation because the anchors were created to be unrelated to one another. As revealed by our results and shown in the table, Items 3, 8, 11, 15, 19, 27, 31, and 35 loaded onto Factor 1 with loadings ranging from .35 to .73. Most of these items, except Item 8, 15, and 31, were hypothesized to measure individuals’ preference for autonomy on their jobs, and thus Factor 1 was determined to be the autonomy anchor. Items 5, 13, 21, 29, 37 loaded on Factor 2 with loadings from .48 to .84. These item loadings were expected and made up the service to a cause anchor. Items 2, 10, 18, 26, 34 hypothesized to measure general managerial competence all loaded onto Factor 3, with loadings ranging from .60 to .82. Next, Items 20, 28, and 41, loaded commonly on Factor 4 from .84 to .92, and thus made up the geographical security anchor. Factor 5 contains 4 items (Items 16, 24, 32, and 40) with loadings .52 to .89, and was determined to be the entrepreneurship/creativity anchor. However, Item 8 (“To be able to create or build something that is entirely my own product or idea is”), an item supposedly under entrepreneurship/creativity anchor, did not load highly on this factor. Instead, it loaded .35 on Factor 1 (autonomy anchor). Items 1, 9, 17, 25, and 33 loaded on Factor 6 with loadings from .55 to .70. Thus, Factor 1 was the technical competence anchor. Factor 7 contained all 3 items (Items 4, 12, and 36) under job tenure security anchor, with loadings from .72 to .79. Items 14 and 30 loaded onto Factor 8 with loadings of .75 and .76, and Items 6, 22, 38 loaded onto Factor 9 with loadings from .52 to .70. These 5 items were supposed to be items measuring the pure challenge anchor, so both Factors 8 and 9 make up the pure challenge anchor. Factor 10 had two lifestyle integration items (Items 7 and 23) loading on it with loadings of .58 and .74. However, Items 15 (“Developing a career that permits me to continue to pursue my own lifestyle is”) and 31 (“A career is worthwhile only if it enables me to lead my life in my own way”), which were also supposed to be lifestyle integration items, only loaded .03 and .25 respectively on this factor. Both of these items, instead, loaded on Factor 1 (autonomy anchor) with loadings of .49 and .70, respectively. Finally, Item 39 (“Choosing and maintaining a certain lifestyle is more important than career success”) of lifestyle integration anchor loaded on its own on Factor 11.
Factor Loadings of Items on 11-Factor Structure.
Note. Principal axis factoring with varimax rotation. Factors had Eigenvalues of above 1.00. Items sorted by size of loadings. Boldface numbers indicate highest factor loading of each item. +Item did not load on expected factor.
Correlations Between COI Career Anchors and Holland’s (1973) Typology of Career Interests.
Note. COI = Career Orientations Inventory.
Alphabetical letters represent typology of career interests where R = Realistic; I = Investigative; A = Artistic; S = Social; E = Enterprising; C = Conventional.
*p < .05. **p < .01.
Since a number of items showed cross-loading on more than one factors, we examined the factor structure of items of the original nine factors by conducting a lower order factor analysis within each anchor. Notably, items under each anchor revealed a single-factor structure except those under the anchors pure challenge and lifestyle integration. Items in the pure challenge anchor yielded a two-factor structure. Two items (14 and 30) loaded on one factor, and 3 items (6, 22, and 38) on another. Items 14 and 30 assess how much individuals value winning (e.g., “Competing and winning out over others [is important to me],” while Items 6, 22, and 38 focused more on overcoming challenges and problems at work (e.g., “Working on problems that are almost insoluble [is important to me].”The pure challenge anchor appears to include both the intention to overcome challenges and the motivation to win. Similarly, the lifestyle integration anchor yielded a two-factor structure. Items 7 and 23 loaded on one factor and Items 15, 31, and 39 loaded on another. Items 7 and 23 reflect a preference to balance career and family (e.g., “I have always tried to give equal weight to my family and to my career,” while Items 15, 31, and 39 emphasize the importance of maintaining a personal lifestyle (e.g., “A career is worthwhile only if it enables me to lead my life in my own way”).
Reliability Assessment
For reliability assessment, we tabulated the subscale Cronbach’s α scores following the original nine-factor structure as proposed by Schein (1985). This nine-factor structure is consistent with how this scale has been used in research and practice. Respectively, the technical competence, general managerial competence, autonomy, job tenure security, geographical security, entrepreneurship/creativity, service to a cause, pure challenge, and lifestyle integration subscales yielded Cronbach’s αs of .80, .87, .81, .84, .91, .85, .87, .72, and .63. Cronbach’s α for the entire 41-item scale was .76. Other than the lifestyle integration subscale, the reliabilities coefficients of all subscales fell in the acceptable range of .70s to .90s as advised by Nunnally (1978). The relatively lower reliability of the lifestyle integration subscale was not entirely surprising considering how several items that were expected to load on this anchor loaded on other factors instead, as found in our prior exploratory factor analysis. However, removing any item from the lifestyle integration subscale would further lower the internal reliability of the subscale to the .50 range.
Validity Assessment
To assess the convergent and discriminant validities of the career anchors, we ran correlations between the career anchors and Holland’s (1973) career interests subscales (realistic, investigative, artistic, social, enterprising, conventional). Table 3 presents the correlations between the variables. As shown, the technical competence anchor of COI was positive related to the investigative subscale, and negatively related with the conventional subscale. General managerial competence was positively related to Holland’s enterprising subscale. Next, autonomy anchor was negatively related to both social and enterprising subscales of Holland’s typology. Entrepreneurship/Creativity anchor had a positive correlation with investigative subscale. The service to a cause anchor was positively linked to realistic, investigative, and social subscales. The job tenure security, geographical security, pure challenge, and lifestyle integration scales showed no significant relations with Holland’s career interest subscales. Several of these significant correlations were expected, such as those between technical competence and investigative and general managerial competency and enterprising. These relationships provided support for COI’s convergent validities and are further discussed in the next section.
Significant Correlations Between COI Anchors and Holland’s Career Interests.
Note. COI = Career Orientations Inventory.
Alphabetical letters represent typology of career interests where R = Realistic; I = Investigative; A = Artistic; S = Social; E = Enterprising; C = Conventional.
Results indicated are significant at p < .05.
Discussion
In this article, we examined the psychometric properties of Schein’s (1985) COI by conducting exploratory factor analysis, reliability assessment, and tests of convergent and discriminant validities. Here, we discuss the results and evaluate the psychometric strengths and weaknesses of the COI.
Strengths
Our factor analysis has provided overall support for the factor structure of the COI. Items that were expected to load on six of the anchors (service to a cause [Factor 2], general management competence [Factor 3], geographical stability [Factor 4], entrepreneurship/creativity [Factor 5], technical competence [Factor 6], and job tenure security [Factor 7]) all loaded on their respective anchors. Nordvik (1996) reported a four-factor structure for the COI using an ipsative response scale, which raised the question of how valid the nine-factor structure proposed by Schein (1985) was. Our study has demonstrated that using Likert-type scale versus ipsative scale yielded factor analysis results that are more consistent with Schein’s original conceptualization. Since most existing studies that have used the COI used a Likert-type response scale, we contend that results from our factor analysis are appropriate for evaluating the psychometric properties of COI. In other words, besides endorsing the factor structure of COI as proposed by Schein (1985) as opposed to Nordvik’s (1991) four-factor structure, our results support the validity of existing studies that have used the Likert-type scale for responding to COI.
Next, our reliability assessment indicated that most of the anchors except lifestyle integration anchor have good internal reliability. Our reliability results also showed that removal of any item from each anchor would reduce the internal reliability of the respective anchor. This means that the current COI, despite having 41 items, is a parsimonious measure that should not be further shortened.
Finally, our validity assessment has demonstrated that COI does possess some degree of convergent and discriminant validities with existing career measures, namely the Holland’s (1973) typology of career interests. For instance, technical competence anchor was positively related to the investigative subscale of Holland’s (1973) framework. This was expected because investigative individuals are often trying to develop and acquire professional knowledge on their jobs (Holland, 1996), which is also descriptive of individuals with a technical competence anchor. Also, our results revealed a significant positive association between the general managerial competence anchor and Holland’s (1973) enterprising subscale. Enterprising individuals express interest in persuading or directing others, and often define their success by their power status in a social setting (Holland, 1996), traits that are characteristic of individuals with a general managerial competence anchor who enjoy managing and coordinating the efforts of others. On the other hand, the lifestyle integration anchor was not found to be related to any of Holland’s (1973) subscales. Lifestyle integration focuses on the need to maintain balance across multiple domains of one’s life, such as family, leisure, and work. This is not a dimension that Holland’s (1973) typology assesses; hence, it was not surprising that lifestyle integration anchor did not show any significant relationships with any of Holland’s (1973) subscales. Taken together, these findings demonstrated some degree of convergent and discriminant validities of the COI.
Weaknesses
Despite the above strengths, our results have also uncovered some weaknesses associated with the COI. First, in our factor analysis, the 5 items designed to measure the anchor of pure challenge loaded on two factors, Factors 8 and 9. This suggests that the pure challenge anchor could possibly be further categorized into two categories. Also, items that were supposed to represent the lifestyle integration anchor loaded on Factor 1 (which largely contained items from autonomy anchor), Factor 10, and Factor 11. These cross-loadings across three factors suggest that items under lifestyle integration do not unequivocally measure a single construct. In fact, several of the lifestyle integration items cross-loaded onto the autonomy anchor. Autonomy may reflect the preference for having freedom at work to establish one’s identity, while lifestyle integration emphasizes the freedom to balance work and personal demands. These two anchors appear to be somewhat overlapping. Such findings imply that the lifestyle integration anchor is not entirely stable, and this is further supported by the lifestyle integration anchor having the lowest Cronbach’s reliability of .63 among all the anchors. This is a weakness of the COI, and calls for caution to be exercised by individuals, career counselors, or managers when making recommendations to individuals who appear to have a lifestyle integration anchor, because scoring high on these items may not necessarily mean individuals view lifestyle integration as central in making career specialty decisions. Without further study, it is difficult to characterize what essential career anchor these items assess.
Also, while our correlation results supported that the COI does have some degree of convergent and discriminant validity, not all findings were clear in this regard. For example, the technical competence anchor was negatively correlated with Holland’s (1973) conventional subscale. This contrasts Nordvik’s (1991) finding of a positive correlation between the technical competence anchor and the conventional subscale. Also, we found that job tenure security and geographical security anchors were not related to any of Holland’s subscales, while Nordvik found both anchors to be positively related to the conventional subscale. These discrepant results could either be due to limited validity of the COI, or because we used a Likert-type scale while Nordvik used an ipsative scale for responding to the COI items. In addition, these discriminant findings could be due to sample variance. Our sample was made up of middle-aged American males, while Nordvik’s sample consisted of Norwegian males and females. Nordvik’s sample was also more than 10 years younger than our sample on average. These demographic differences between samples could result in different responses to the COI items. In any case, such contrasting findings suggest the need to reevaluate and examine what exactly these career anchors mean conceptually, how they should actually relate to existing career measures or job types, and also whether the anchors can be generalized cross culturally. Such understanding would help researchers and counselors better use the career anchors to predict individuals’ career interests and make recommendation for their career specialty options. Further studies are required to answer these questions.
Limitations and Future Directions
We also acknowledge the following limitations in our study and make some suggestions to address them. First, our study did not examine the test–retest reliability of the scale. Test–retest reliability helps to establish the temporal reliability of scales to ensure that findings associated with these measures could apply across time (Guttman, 1945). Unfortunately, due to constraints posed by our data collection method, we did not manage to get follow-up data from respondents. Fortunately, Nordvik (1991) found that the test–retest reliability of the COI anchors using an ipsative scale had a mean 2-month test–retest reliability of .82, providing evidence that the COI is a temporally stable measure. Nonetheless, it would be ideal for future research to examine the test–retest reliability of the COI using the Likert-type scale that is used in our study.
In addition, our sample was made up of the 1962 graduating class of Dartmouth College and consisted only of males. This could raise the question whether the results are applicable across genders. In existing studies with COI using both male and female participants, no gender differences have been reported. Therefore, we have reasons to believe that having an all-male sample does not affect the utility of our results. However, future psychometric evaluation studies on COI might still want to expand the sample to include both genders.
Future psychometric studies on COI could also conduct confirmatory factor analysis to examine the fit of the nine-factor structure. If future studies also find an 11-factor structure, it would support the exploratory factor analysis findings of the current study and imply that both the pure challenge anchor used might fully be divided into two anchors (overcoming challenges and winning), and the lifestyle integration anchor divided into personal lifestyle and work–family balance anchors.
In addition, future research might want to examine the incremental validity of the COI beyond existing career measures in predicting more distal outcomes such as job performance and job satisfaction. This could possibly be done by longitudinal studies individuals working in careers matching or not matching, and assessing job performance and job satisfaction over time.
Finally, our study did no evaluate the clinical utility of the COI, that would include factors such as scale administration, training requirements for its use, and the meaning and relevance of the information obtained from the scale (Smart, 2006). For a scale to possess value, it should ideally possess both psychometric soundness and clinical utility. It is not within the scope of the current article to examine the clinical utility of the COI, but its wide application in existing research and practice, suggests its clinical utility. Future research should examine how the COI is being used in applied settings, and systematically evaluate its clinical utility.
In conclusion, our study reveals both psychometric strengths and weaknesses of the COI. Specially, the COI is an internally consistent instrument, with a factor structure rather similar to the one originally proposed by Schein (1985) almost 30 years ago. However, our study also leaves questions about the validity and interpretation of the scale, and the limits of its utility in guiding individuals’ specialty choices within a field. Results of our study suggest the need for individuals, career counselors, and organization managers to be prudent when interpreting individuals’ results on the COI, and not choose, advise, or assign specialty paths based solely on COI results. Ultimately, career decisions, including specialty choices, are dynamic processes (Feldman & Bolino, 1996). The COI is a career measurement tool that aims to ease the difficult process of career specialty choice by making individuals more aware of their preferences. However, it is necessary to also recognize how various personality and situational factors not measured by the COI can interact to influence such specialty choices. A comprehensive evaluation of all these factors is necessary for good career advice and decisions.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
