Career planning validity of self-estimates and test estimates of work-relevant abilities.

This study sought to determine whether self-estimates of work-relevant abilities can improve upon the validity of test estimates (scores). Validity for career planning applications was the focus. The study (final sample of 1,620 college students) used 9 self-estimates and test estimates to predict certainty-screened occupational choices grouped by J. L. Holland (1997) types. The addition of ability self-estimates to test estimates substantially improved upon the validity of test estimates alone. Study results (together with previous research) suggest that ability self-estimates can provide efficient additions or alternatives to test estimates during career planning, while substantially broadening the range of abilities under consideration.

**********

As Donald Super noted over 40 years ago, "In choosing an occupation one is, in effect, choosing a means of implementing a self-concept" (1957, p. 196). What might be called "Super's Dictum" has implications for both career counseling and career development, more generally considered. Regarding the former, self-estimates provide a way to make career-relevant self-concepts evident to the counselor and counselee, and thus, open for discussion, clarification, and application. Regarding the latter, self-concepts are relevant to understanding career choice.

The general purpose of this study was to determine whether self-estimates of work-relevant abilities can improve upon the validity of test estimates (i.e., test scores) when both are used to facilitate career exploration and planning, hereafter called "career planning." The study's validity analyses addressed the use of self-estimates of work-relevant abilities in a comprehensive, work-world search for occupations with counselee-compatible work tasks, that is, in career planning "for the many." The use of self-estimates in personnel selection, college admissions, and so forth was not investigated, nor is it proposed.

In what follows, the term work-relevant abilities includes noncognitive abilities in addition to cognitive abilities, and it subsumes "basic and cross-functional skills" (Peterson, Mumford, Borman, Jeanneret, & Fleishman, 1999, p. 51). Many of these abilities are relevant to career planning, but relatively few are routinely assessed by tests. For a number of years, self-estimates have been used to address this problem. However, computer-based literature searches using PsycINFO (http://www.csa.com) and the Social Science Citation Index (http://www.webofscience.com) located only five studies (summarized below) that determined whether self-estimates can improve upon the career planning validity of test estimates. Because both self-estimates and test estimates of abilities are readily available to career counselors, research on this topic would appear to be important to the practice of career counseling.

Estimates of skill self-confidence and self-efficacy were not addressed in this study because no one appears to view them as ability estimates. For example, Betz, Borgen, and Harmon (1996) noted that the Skills Confidence Inventory "is not a measure of actual abilities" (p. 21). Regarding the Campbell Interest and Skill Survey, see Campbell, Hyne, and Nilsen (1992, p. 41). In a study of the relationship between self-efficacy beliefs and self-rated abilities, Brown, Lent, and Gore (2000) concluded that the constructs "are empirically distinguishable.... [They] may, thus, serve complementary, rather than competing, roles relative to ... career choice making" (p. 233). The computer-based literature searches revealed no studies comparing the career planning validity of skill self-confidence and self-efficacy estimates with that of test estimates of work-relevant abilities.

Rationale for the Use of Ability Self-Estimates

Prediger (1999b) discussed general considerations regarding the assessment of abilities for the purpose of facilitating career planning. Much of this section draws on that discussion.

As evidenced by common sense (e.g., Jones, 1996), theory (e.g., Holland, 1997), and research (e.g., Peterson et al., 1999; Prediger, 1989), occupations differ on a wide range of work-relevant abilities. Consider, for example, the different types of abilities needed by sales workers, machinists, commercial artists, actuaries, and career counselors. Table 1 lists the 15 work-relevant abilities used in the study reported here. Their bases in theory and research (beginning with Abe & Holland, 1965) were reported by ACT, Inc. (ACT, 2001). A primary consideration was comprehensive coverage of career clusters paralleling Holland's (1997) typology. Certainly, it would be possible to add abilities to the list. But their intended use with counselees suggested otherwise.

Unfortunately, test estimates for many work-relevant abilities are seldom available (e.g., see Abilities 8-15 in Table 1). Hence, self-estimates are needed if those abilities are to be considered in career planning. Also, the test estimates that are available may have limitations, for example, narrow operational definitions for some abilities and the time and expense involved. In contrast, ability self-estimates can systematically address a wide range of broadly defined work-relevant abilities in a relatively brief amount of time.

The various ways in which ability self-estimates are obtained are illustrated by the Self-Directed Search (Holland, 1997), the Ability Explorer (Harrington & Harrington, 1996), the O*NET Web site (http://online.onetcenter.org) Skills Search, and the Inventory of Work-Relevant Abilities (IWRA), which is used in the Career Planning Survey (ACT, 2001) and DISCOVER (ACT, 1999). Prediger (1999b) emphasized the importance of obtaining self-estimates that are informed by experience (direct or vicarious) and by feedback over the years from parents, peers, teachers, and employers. Test scores, if available, can also be considered. Prediger (1999b) noted the need to use normed scores to report self-estimates in order to address general differences in self-estimate "optimism" across abilities and to facilitate ability level comparisons across abilities.

Evidence regarding self-estimate validity is sometimes sought by correlating self-estimates with test estimates of ability. However, it is important to keep in mind that, even when the ability's name is the same for both types of estimates, they may operationally define different aspects of the ability. In addition, neither self-estimates nor test estimates are perfectly reliable, and they are subject to similar types of distortion (e.g., due to limited experience and response style). Hence, if correlations between the two types of estimates are low, both may be the problem.

Regarding self-estimate and test-estimate validity, studies of validity for career planning applications appear to be much more relevant than correlations between the two types of ability estimates. The results of five such studies (four published) were summarized by ACT (2001). These were the only such studies found in the literature searches. The studies' five samples consisted of 1st-and 2nd-year college students (one sample) and students in Grades 11, 11-12, and 12 (two samples). Sample sizes ranged from 356 to 1,620. In each of the studies, the self-estimates were obtained via an early edition of IWRA.

As documented by Prediger (1998), career planning validity is commonly determined by finding the percentage of occupational group members whose membership is predicted by an assessment's scores, that is, by finding the "hit rate." In effect, this validation model asks whether persons in a given occupational group would have been referred to that group by their scores, which is consistent with the goal of identifying personally relevant occupational options. When there are many occupations, they are often grouped by Holland's (1997) six types. Prediger (1998) also noted that occupational choice "has long been used and defended" (p. 207) as the basis for determining occupational group membership and that sometimes the choices are first screened for certainty.

In each of the five studies, ability estimates were obtained for the 15 abilities listed in Table 1, and occupational group membership was based on certainty-screened occupational choices assigned to Holland's (1997) types (see Table 1). Test estimates were available for Abilities 1-6 in Table 1. All 15 abilities listed in Table 1 were used to obtain composite ability scores for Holland's types. Each composite was based on the four abilities flagged by Holland type in Table 1. In four studies, the six ability composites were based on test estimates for Abilities 1-6 plus self-estimates for Abilities 7-15. In the fifth study, the composites were based on self-estimates for all 15 abilities.

For each study, the average occupational group hit rate for the ability composites was greater than that for the test estimates alone (medians of 42% and 34%, respectively). Thus, the composites had greater validity for career planning applications. The hit rates for the composites were similar to those typically reported in interest inventory validation studies using Holland-type occupational groups (e.g., see Swaney, 1995, for a summary).