Validating and Using the Career Beliefs Inventory

J o h n L. Holland

Johns Hopkins University

J o s e p h A. J o h n s t o n N. Francis A s a m a S u z a n n e M. Polys

University of Missouri-Columbia

For many years career counselors have observed that a person's beliefs about potential vocational alternatives can either foster or interfere with his or her decision-making, but until the development of the Career Beliefs Inventory (Krumboltz, 1991) there has been no formal tool for assessing career beliefs. The Career Beliefs Inventory (CBI) grew out of Krumboltz's (1983) thinking about destructive career beliefs. This inventory characterizes a client's beliefs so that the client and counselor have a convenient summary of the beliefs that may foster or hinder a person's vocational decision making. This report has two goals, (a) to examine how the CBI scales are correlated with the NEO Personality Inventory (NEO, Costa & McCrae, 1985), The Self Directed Search (SDS, Holland, 1985), The Vocational Identity scale (Holland, Daiger & Power, 1980), and the Preconscious Activity scale (PA, Holland & Baird, 1968). We expected that the scale correlates of the CBI would generally support the CBI scale labels or interpretations. We also wanted (b) to learn how clients react to interpretations of their results in informal career counseling.

Correspondence concerning this article should be addressed to John L. Holland, 111 St. Albans Way, Baltimore, Maryland 21212.
Journal of Career Development, Vol. 19(4), Summer 1993 1993 Human Sciences Press, Inc.
233

234

Journal of Career Development

Method

Subjects
Adults were obtained from two career planning programs in Kansas and Missouri (37 men and 88 women). The two career planning samples were combined to obtain more stable statistics. Inspection of the means, standard deviations, and correlations for demographic and test variables were similar but not identical. The largest differences were by gender rather than by other variables; so all subsequent analyses were performed for the total sample and by gender. All participants were adults who presented themselves for one of these programs. For the total sample (N= 125) the average age was 40.1 with a standard deviation of 9.6. The average educational level was about two years of college and ranged from the ninth grade to a graduate degree; 58% were employed, 18% were unemployed, 15% were homemakers, and 8% were students.

Procedures & Measures

All adults were administered the following inventories and scales: The Career Beliefs Inventory (CBI, Krumboltz, 1991) consists of 25 brief scales ranging from 2 to 8 items per scale. Their retest reliability correlations range from .35 to .73 with a median correlation of .58 for a sample of 51 high school students over a 1-month interval. See the CBI Manual (Krumboltz, 1991) for more retest information about college samples. The CBI scales were created rationally and by a series of factor analyses using large samples of high school and college students as well as adults. The CBI also includes an Administrative Index of 25 matched pairs of items to detect inconsistent responses. A short form of the Dogmatism scale (Rokeach, 1960). It is a 20item form (Troldahl & Powell, 1965) which has been shown to correlate .94 with the 40-item form. The split-half reliability of the long form is about .84 (Troldahl & Powell, 1965). The NEO Personality Inventory (Costa & McCrae, 1985), assesses five domains of personality--Neuroticism, Extraversion, Openness, Agreeableness, and Conscientiousness. In other samples, the domain scales (N,E,O,A,C) have shown a high degree of internal consistency (.91 to .93) and retest reliability (.86 to .91) and have an impressive history of research based on adult samples, correlations with other inventories, sentence completion devices, and convergent correlations

J o h n L. Holland, et al.

235

between self-reports and spousal ratings as well as peer ratings. The alpha for the five 12-item scales has ranged from .74 to .89. The Self-Directed Search (Holland, 1985a) has six scales of 38 items each whose internal consistency (corrected split-halt) ranges from .83 to .95. The SDS manual (Holland, 1985b), the Manual Supplement (Holland, 1987), and a recent bibliography (Holland & Gottfredson, 1990) provide extensive evidence for its reliability, validity, and theory. The Vocational Identity Scale (VI, Holland Daiger and Power, 1980) is composed of 18 true-false items. A recent review of 37 investigations (Holland, 1991) indicates that this scale has substantial construct validity, satisfactory internal consistency (about .86), and retest reliability for a two-week interval (.63 to .93) for samples of high school and college students as well as adults. The Preconscious Activity scale (PA, Holland & Baird, 1968), an originality measure, was stimulated by Kubie's theory (Kubie, 1958) of the creative process. The first form of the scale was positively correlated with artistic interests (VPI), self-ratings of originality, Barron's Complexity-Simplicity scale (1953), and faculty ratings of the originality and independence of undergraduates. The internal consistency (KR 21) was .70 for 544 men and 275 women (Nichols & Holland, 1963). For the present study, the 38-item PA scale was reduced to 26 items (half true and half false) to obtain a shorter and more reliable scale. An earlier item analysis (Nichols & Holland, 1963) was reviewed and the weakest items were eliminated. The internal consistency (KR 21) of the 26-item PA scale was .75 for the total sample (N = 125). With the Missouri sample (n = 86), the CBI, NEO, Dogmatism, Preconscious Activity, and the SDS were administered by mail as a continuing voluntary assessment of program participants. In the case of the SDS, about 50% of the participants had taken the SDS 2 to 12 months earlier and were not retested. These discrepancies in testing time imply that the correlations between the SDS and other variables are conservative values. The Kansas sample (n = 39) took all instruments at the same time when they first presented themselves for career counseling.
Results

Correlational Analysis
The means and standard deviations for the scale variables (CBI, SDS, NEO, PA, VI, and Dogmatism) by gender are shown in Table 1.

236

Journal of Career Development

Table 1 Means and Standard Deviations of Scale Variables for M e n a n d W o m e n

Men (n = 35---37) Scale M SD Women (n = 8 6 - 8 8 ) M SD t

CBI Adm. Administrative Index SC1. Employment Status SC2. Career Plans SC3. Acceptance of Uncertainty SC4. Openness SC5. Achievement SC6. College Education SC7. Intrinsic Satisfaction SC8. Peer Equality SC9. Structured Work Environment SC10. Control S C l l . Responsibility SC12. Approval of Others SC13. Self-other Comparisons SC 14. Occupation/College Variation SC15. Career Path Flexibility SC16. Post-training Transition SC17. Job Experimentation SC18. Relocation SC19. Improving Self SC20. Persisting While Uncertain SC21. Taking Risks SC22. Learning Job Skills SC23. Negotiating/Searching SC24. Overcoming Obstacles SC25. Working Hard SDS Realistic Investigative

61.0 37.5 28.5 28.0 39.3 37.5 32.5 36.8 20.3 32.5 29.5 29.8 25.0 34.3 39.8 29.3 32.8 34.4 29.6 38.5 38.2 40.3 37.5 39.3 35.1 38.1 31.3 26.2

14.1 11.5 10.5 9.0 6.0 7.5 10.0 5.8 6.0 7.5 7.0 4.0 7.5 6.5 3.8 6.3 7.0 4.3 7.6 8.5 6.2 6.0 5.0 5.5 4.9 4.9 8.2 11.9

63.6 35.5 31.5 29.0 38.8 39.0 30.0 37.6 22.0 33.5 32.0 28.4 25.5 35.0 41.0 30.5 33.4 34.3 27.0 38.5 39.6 40.8 36.5 42.0 34.9 39.1 18.2 18.9

14.0 14.0 10.5 9.0 7.0 6.5 9.5 5.0 6.0 7.5 8.5 4.8 7.5 5.3 3.6 5.8 5.6 3.3 8.6 7.5 5.6 5.5 6.5 4.0 4.4 5.0 9.4 9.8

0.97 0.77 1.37 0.40 0.31 1.22 1.55 0.75 1.31 0.79 1.44 1.61 0.48 0.59 1.58 1.21 0.58 0.30 1.60 0.01 1.21 0.49 0.58 3.10"* 0.23 1.10 7.25** 3.46**

J o h n L. Holland, et al.

237

Table 1 (Continued)
Men (n = 35-37) Scale M SD Women (n = 86-88) M SD t

Artistic Social Enterprising Conventional NEO Neuroticism Extraversion Openness Agreeableness Conscientiousness Preconscious Activity MVS Vocational Identity Dogmatism
*p < .05. **p<.01.

18.2 26.5 27.1 21.8 21.3 28.2 29.2 31.1 32.5 12.2 8.9 - 5.5

10.8 10.3 9.6 10.6 7.7 6.2 6.5 4.4 6.6 4.9 4.4 12.4

20.0 30.4 24.8 27.5 22.2 28.6 28.9 33.9 34.6 11.9 8.2 -13.9

11.0 8.6 9.6 11.2

0.82 2.17" 1.18 2.56*

8.2 0.55 6.0 0.38 5.9 0.33 4.9 2.95** 5.5 1.81 4.8 0.35 4.2 0.81 15.6 2.98**

The SDS scales reveal some traditional gender differences especially for the Realistic and Investigative scales. In contrast, the NEO and CBI scales show only small or trivial differences. The PA scale is not related to gender, but the Dogmatism scale implies that women are less dogmatic than men. The main findings are presented in Table 2 and Table 3. In general, the correlates of the CBI support the meanings attributed to the scales. For instance, Scale 25, Working Hard (Hard work will bring success), is related to not being neurotic (-.51), being extroverted (.34), open (.18), agreeable (.40), enterprising (.19), having a clear sense of identity (.40), and not being dogmatic (-.40). Scale 10, Control (Career path is self-determined), is related to not being neurotic (-.35), extroverted (.20), agreeable (.28), conscientious (.21), having low Realistic interests ( - .26), and low dogmatism ( - .26). The following scales (Persisting While Uncertain, 20; Taking Risks, 21; Negotiating/Searching, 23; Overcoming Obstacles, 24; and Working Hard, 25) share similar correlational patterns with the NEO PI scales and the Identity scale. To summarize, these scales are posi-

II

I l l

II

Q~

II

II

~S

II

II

Z
I III III I

I l l l

I l l

I I I I

L~ J I I I I II

I l i l l

II

II ~ ~ = l ~ r . O

I I I I I I

, ~ ~'~

~.~
~'~,

M 1= ~,--4 b . e o

I .~.~< = 2~I~

I I I I

~ II ~

~~

~.~

o'~. ~ = ~v

I I I

It

II

III

II

II

II

I l l l l l l

II

II

II

I ~ 1

~4

I
c~

I I I I I

II

FII

II

I l l

II

II

I I I I I c~
c~

o.~, 0 ~

II

I I I I I I

o"~

.r~

~,..~

c/~ ~ ' - ) ; " ~

~ ~ ~

m II

--

0 '.-.' ~ 0 0 . ]

>~ I t l

~ ~

~,.~0

~ ~ ~.~.~
il ~,-,~.,.,~ o ~

r.~ ~

l>

. ~ - -. ',~ . . = c',,1 o ~ ~

I l l

'-"

0 0 0 0 0 0 0

~ II ~ . a . . , ~

,~,-.e

o"b
~

"

242

Journal of Career Development

tively correlated with Consistency, Extroversion, Agreeableness, Conscientiousness, and Identity, and they are negatively correlated with Neuroticism. These correlates imply that the CBI scales have impressive construct validity, but they also imply that scales 20, 21, 23, 24, and 25 are measuring the same variable or factor. In fact, the factor analysis in the CBI Manual (Krumboltz, 1991, page 23) reveals that these five scales have high loadings on the first factor labeled "Belief that work is v a l u a b l e . . . " Several other findings appear noteworthy. The Administrative Index (AI), which is a measure of the tendency to give consistent responses when similar ideas are worded differently, appears to be a good measure of general competency and positive outlook. It is significantly correlated with 14 of the 25 CBI scales and in directions that always imply desirable personal characteristics such as hardworking, open, belief in self-determination, and so on. The correlates of this index with the non-CBI scales are also impressive and positive. Consistent responses (AI) are - . 4 4 with Neuroticism, .20 with Extroversion, .33 with Agreeableness, .34 with Conscientiousness, .32 with Identity, and - . 2 7 with Dogmatism. In addition, gender is related only to Scale 23, Negotiating/Searching, indicating that women m a y be more inclined to negotiate work changes or seek a new job t h a n men (r = .27). Finally, the four correlations between age and the CBI scales appear difficult to interpret and are probably of little value. If readers wend their way through the remaining scales, they will find most correlates appear to be consistent with the CBI labels, but m a n y scales have only one or two sigr~ificant correlates. For instance, Enterprising interests are correlated with Achievement (r = .25), Self-Other Comparisons (r = -.32), Relocation (r = .39) and Overcoming Obstacles (r = .23). These correlations replicate the same correlations reported in the CBI Manual for a college sample. Similar patterns of construct validity can be observed for the Investigative, Social, and Conventional scales of the SDS. In short, CBI and SDS scales have some weak but plausible intercorrelations. Another example is Scale 9, Structured Work Environment (Prefer standard hours with supervision). This scale is correlated with Extraversion (.23), not Open ( - .30), Conscientiousness (.32), low Artistic interests ( - .28), Conventional interests (.21), and low Originality ( - .33).

Counseling Experience
The second phase of this study was to work with some small groups of adults and undergraduates who had not participated in the correla-

J o h n L. Holland, et al.

243

tional study. Our aim here was to explore how to use the CBI, to learn if individual scale interpretations had validity, and to secure an appreciation of the effects of the test-taking and interpretive discussions. This work then might lead to a more defensible or experimental evaluation. In the first informal evaluation two authors met with two groups of adults ( N = 10) who were considering career changes. Much to our surprise, these hour-long sessions went well. The participants were told that they had taken a new inventory and that we were unclear how to proceed and needed their help. All participants were asked to identify themselves and their current career situations. Then we proposed to ask about their responses to the CBI beginning with Scale 25. Each person had a CBI interpretative profile; so we went around a table asking for scores and suggesting interpretations. With each scale, discussions became more informal and open. When interpretations missed the mark, participants were asked to clarify the meaning of their scores. Before each session was over, we (the authors) believed that we had a useful understanding of each participant because they participated so actively and frankly, and because they talked at length about themselves and their particular career problem or potential transition. In the second informal evaluation, Polys (1991) administered and interpreted the CBI to a group of 12 adult volunteers (ages 22 to 42). She reported that the participants found the CBI helpful, that the inventory did identify facilitating and hindering beliefs, and that the CBI promoted other career activity such as talking with friends, spouses, or family about career plans. Small group interpretations went well and participants readily entered discussions of their beliefs. Generally, items were viewed as relevant and non-threatening.

Discussion Although this preliminary study of the CBI is based on small adult samples, its correlates with some well-established inventories and scales imply that most CBI scales and its Administrative Index have at least moderate construct validity. In short, scale correlates appear consistent with CBI scale labels or interpretations. At the same time the CBI scales appear to be measuring variance not currently assessed by the other instruments. Parenthetically, Scale 25, Working Hard, appears to be an excellent measure of Self

244

Journal of Career Development

Efficacy. The similar correlational patterns found for scales 20, 21, 23, 24, and 25 imply a need to reconsider combining or reorganizing these scales. At the same time, the current scales can be justified because of their specificity and their usefulness in stimulating client explorations. Along with its stated purpose as a diagnostic inventory, the CBI should be studied for its therapeutic value. The authors anticipated some difficulty in interpreting scale meanings and found instead that the CBI fostered client talk and group interaction. Experimental studies of the CBI as a career intervention appear desirable now.

References
Barron, F. (1953). Complexity-simplicity as a personality dimension. Journal of Abnormal and Social Psychology, 48, 163-172. Costa, Jr., P.T., & McCrae, R.R. (1985). The NEO Personality Inventory. Odessa, FL: Psychological Assessment Resources. Costa, Jr., P.T., & McCrae, R.R. (1989). Manual Supplement for the NEO P1/FF1. Odessa, FL: Psychological Assessment Resources. Holland, J.L., Daiger, D.C., & Power, P.G. Some diagnostic scales for research in decision-making and personality: Identity, information and barriers. Journal of Personality and Social Psychology, 1980, 39, 1191-1200. Holland, J.L. (1991). The Vocational Identity Scale: A Diagnostic and Treatment Tool. San Francisco, CA: Paper presented at APA convention. Holland, J.L. (1985b). Professional manual for the Self-Directed Search. Odessa, FL: Psychological Assessment Resources. Holland, J.L. (1985c). Manual for the Vocational Preference Inventory. Odessa, FL: Psychological Assessment Resources. Holland, J.L. (1987). The 1987 Manual Supplement for the Self-Directed Search. Odessa, FL: Psychological Assessment Resources. Holland, J.L. & Baird, L.L. (1968). The Preconscious Activity scale: The development and validation of an originality measure. Journal of Creative Behavior, 2, 217-225. Krumboltz, J.D. (1991). Manual for the Career Beliefs Inventory. Palo Alto, CA: Consulting Psychologists Press. Polys, S.M. (1991). Career Beliefs Inventory: Its impact on users and suggestions for further development. Unpublished manuscript. University of Missouri--Columbia. Rokeach, M. (1960). The Open and Closed Mind. New York: Basic Books. Troldahl, V.C., & Powell, F.A. (1965). A short-form Dogmatism Scale for Use in Field Studies. Social Forces, 44, 211-214.