Download as pdf or txt
Download as pdf or txt
You are on page 1of 10

Journal of Counseling Psychology Copyright 1992 by the American Psychological Association, Inc.

1992. Vol. 39. No. 1.71-80 0022-0167/92/tt.OO

A Typology of Career Decision Status:


Validity Extension of the
Vocational Decision Status Model
Connie R. Wanberg and Paul M. Muchinsky
Iowa State University

Three hundred ninety college students completed a battery of scales assessing 10 dimensions of
vocational indecision and 9 personality constructs. Cluster-analytic procedures were applied to a
subset of the scaled constructs, whereas other measures served as external descriptors. A 4-cluster
solution was obtained, with the clusters being differentiated on the basis of vocational decidedness
and personal concern over the stage of career decision making. The resulting cluster solution was
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

successfully replicated internally. Effect sizes in excess of .30 were obtained for 8 variables in the
This document is copyrighted by the American Psychological Association or one of its allied publishers.

cluster analyses and 2 external descriptors. The results are described in terms of their correspond-
ence to the vocational decision status model proposed by Jones and Chenery (1980).

The construct of vocational indecision purports to explain clear evidence that this is the best way to classify individuals
why some individuals are uncertain about their choice of a who have not specified a career choice.
college major or future career, whereas others are committed A second typology of career indecision, called the voca-
to their choices. Early research (e.g., Crites, 1969; Goodstein, tional decision status model, was proposed by Jones and
1965; Tyler, 1961) suggested that undecided individuals do Chenery (1980). They proposed that undecided individuals
not represent a homogeneous group, but rather multiple forms can be characterized according to three dimensions: (a) decid-
of indecision may exist. More recent research has been de- edness (the degree of an individual's indecision), (b) comfort
voted to defining and explicating various typologies of career level (how comfortable the individual is with his or her level
undecided individuals. of indecision), and (c) reasons (what reasons the individual
One typology differentiates undecidedness from indecisive- has for his or her indecision). They developed a scale—the
ness. Undecided individuals are hypothesized to be going Vocational Decision Scale (Jones & Chenery, 1980)—to mea-
through a normal and temporary stage of development. They sure these three dimensions. The Vocational Decision Scale
are unable to specify a career choice, but they do not feel was later revised and renamed the Career Decision Profile
pushed or stressed to make a decision. According to Salomone (CDP; Jones, 1989a). One way the CDP can be used is to
(1982), these individuals delay a career decision to gather classify individuals into four groups: (a) decided-comfortable,
more information about themselves, occupations, or the proc- (b) decided-uncomfortable, (c) undecided-comfortable, and
ess of decision making. Holland and Holland (1977) observed (d) undecided-uncomfortable. This classification recognizes
that one half of the undecided individuals in their study that some individuals may be decided in their career choice,
reported that "I don't have to make a decision right now." yet feel uncomfortable with their decision. The CDP can also
Conversely, individuals in the career indecisive group do not be used to differentiate individuals according to four reasons
appear to be going through a normal stage of development, they may have for being undecided: (a) lack of self-clarity
and they have not delayed their vocational choice to gather (indecision due to an individual's inability to understand his
more information. Instead, it is proposed that these individ- or her strengths, weaknesses, interests, and personality), (b)
uals possess attributes that will not allow them to make a lack of knowledge about occupations (indecision due to a lack
decision readily. Salomone (1982) noted that such individuals of information concerning occupations and educational pro-
were characterized by high levels of ambivalence, anxiety, grams), (c) indecisiveness (undecidedness due to a general
and frustration; an unclear sense of personal identity; low self- inability to make decisions), and (d) career choice importance
confidence and self-esteem; externalized locus of control; and (the extent to which respondents feel that choosing and work-
a tendency to blame others for their situation. Hartman, ing in an occupation is an important or unimportant part of
Fuqua, and Hartman (1983) labeled this condition as "chronic their life goals). Overall, the vocational decision status model
indecision." Although the differentiation between career un- seems promising as a means of specifying subtypes. Yet, there
decidedness and career indecision is recognized as potentially is little clear empirical research that has assessed the model
useful, Slaney (1988) commented that "very little progress has proposed by Jones and Chenery (1980).
been made thus far in demonstrating that the two constructs Four studies have used cluster-analytic procedures to clas-
are valid and discriminable" (pp. 44-45). Overall, there is no sify individuals according to their career decision status. Clus-
ter analysis is a statistical procedure that can be used to
identify homogeneous groups of individuals on the basis of
Correspondence concerning this article should be addressed to their scores on a set of related variables. Lucas and Epperson
Paul M. Muchinsky, Department of Psychology, Iowa State Univer- (1990) used measures of self-esteem, locus of control, career
sity, Ames, Iowa 50011. salience, state and trait anxiety, and role salience (relation-
71
72 CONNIE R. WANBERG AND PAUL M. MUCHINSKY

ships, work, and leisure) to investigate possible subtypes of sciousness measure was used as an external descriptor of the
undecided students with cluster analysis. Similarly, Larson, cluster solution, not as a variable within the cluster analysis.
Heppner, Ham, and Dugan (1988) classified undecided stu- Our study includes both career decided and career unde-
dents into clusters using measures of indecision (Career De- cided individuals. Including the scores of both decided and
cision Scale [CDS]; Osipow, Carney, Winer, Yanico, & Kos- undecided individuals allows us to compare ourfindingswith
chier, 1976), interests, problem-solving self-appraisal, and the Jones and Chenery (1980) vocational decision status
career planning difficulty. Both the Lucas and Epperson and model, which also includes both career decided and career
Larson et al. studies included only career undecided students undecided groups.
in their cluster analyses. Although this seems logical given the
goal of finding a typology of undecided students, excluding
decided students fails to recognize the possible existence of a Method
decided-uncomfortable group (as suggested by the vocational
decision status model) that may also need career counseling. Sample
The two other studies using cluster analysis to classify
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

individuals according to their career decision status focused The subjects were 390 undergraduate students enrolled in intro-
This document is copyrighted by the American Psychological Association or one of its allied publishers.

on both career decided and undecided students. Fuqua, Blum, ductory psychology courses at a large midwestern university (197
were women, 188 were men, and 5 did not indicate their gender).
and Hartman (1988) classified high school students into one
The median age was 18 years, with a range from 16 to 55 years. Three
of four types on the basis of their responses to the State-Trait hundred fifty-eight students had formally declared a major, whereas
Anxiety Inventory (STAI; Spielberger, Gorsuch, Lushene, 32 had not. Subjects received extra credit toward their course grade
Vagg, & Jacobs, 1983), the Identity Scale (Holland, Gottfred- for participating in the study.
son, & Nafziger, 1975), the Internal-External Locus of Con-
trol Scale (Rotter, 1966), and the CDS (Osipow, Carney,
Winer, Yanico, & Koschier, 1976). One cluster of students Indecision Measures
was characterized as being career decided and having a low
level of anxiety. The second cluster consisted of career unde- Because vocational indecision was the primary construct of interest
in this study, three instruments were devoted to its assessment.
cided individuals with moderate levels of anxiety and an
Career Decision Profile. The CDP (Jones, 1989b) was developed
internal locus of control. The third and fourth clusters exhib- as a multidimensional measure of career indecision. The measure has
ited high levels of indecision and an external locus of control. six subscales. The Decidedness scale (2 items) measures the extent to
These two clusters differed in their respective levels of anxiety, which a person has decided on an occupational choice. The Comfort
with the third cluster being moderate and the fourth cluster scale (2 items) measures the extent to which a person feels comfortable
being high. Fuqua et al. commented that a greater range of with his or her career decision status. The remaining four scales are
variables in the cluster analysis would have permitted a more concerned with reasons for indecision. These scales include Self-
detailed understanding of how the groups differed. Savickas Clarity (3 items), Knowledge About Occupations (3 items). Decisive-
and Jarjoura (1991) used cluster analysis to identify underly- ness (3 items), and Career Importance (3 items). All items in this
ing subgroups of students on the basis of their responses to measure are answered on an 8-point Likert scale ranging from strongly
disagree (1) to strongly agree (8). According to Jones (1989a), the
CDS items. Again, although these results were supportive of
alpha coefficients for these scales range from .68 to .85. Evidence
multiple types of undecided students, it is suggested that using supporting the convergent and discriminant validity of the CDP scales
a greater range of variables would further explicate the bound- was also shown by Jones (1989a).
aries of this construct. Career Decision Scale. The CDS (Osipow, Carney, Winer, Yan-
The purpose of this study was to further investigate a ico, & Koschier, 1976) consists of 19 items. Items 1 and 2 comprise
possible typology of career decision status by using cluster- the Certainty scale, which provides a measure of the respondent's
analytic procedures. Our research expands the range of vari- certainty surrounding career choice and choice of an academic major.
ables used in past studies to allow additional description and Items 3 through 18 constitute the Indecision scale and are designed
to assess aspects of career indecision. The 19th item permits individ-
delineation of how career subtypes might differ. Careful con- uals to elaborate on their career aspirations and was not used in this
sideration was put into choosing a constellation of variables study. Responses were made on a 4-point Likert scale ranging from
that would be useful in characterizing individuals within a not at all like me (1) to exactly like me (4). Osipow, Carney, and
typology of career decision status. Three instruments were Barak (1976) analyzed the reliability of the CDS with two samples of
devoted to the assessment of relevant aspects of career inde- college students over a 14-day period. They found test-retest corre-
cision, and four instruments were chosen to measure relevant lations for the sum of Items 3 through 18 of .90 and .82. Evidence
aspects of personality. The personality constructs we chose to for the validity of the CDS has been supported by studies of the
use included anxiety, locus of control, and self-esteem; our relationship of CDS scores with age, career counseling interventions,
choice was based on the constructs' empirical relation to the and career choice change (e.g., Osipow, Carney, & Barak, 1976).
indecision construct (cf. Fuqua et al., 1988; Fuqua, Seaworth, My Vocational Situation (MVS). The MVS (Holland, Daiger, &
& Newman, 1987; Maier & Herman, 1974). A measure of Power, 1980) consists of 20 items designed to measure vocational
self-consciousness (assessing the tendency of an individual to indecision. The measure includes three subscales. The first 18 items
comprise the Identity scale and are answered as true or false. The
direct self-attention inward versus outward) was also included score on the Identity scale is the total number of false responses, with
as an exploratory variable in this study, to investigate the higher scores indicating a clearer picture of one's goals, interests,
possible differences among subtypes in their tendency to self- personality, and talents. Validity studies have shown that scores on
reflect. However, as an exploratory variable, the self-con- this scale increase with age, training, and degree of specialization
CAREER DECISION STATUS 73

(Holland etal., 1980). The 19th item is the Occupational Information each item on a scale of extremely uncharacteristic (1) to extremely
scale; it assesses whether or not individuals need any of four kinds of characteristic (5). Test-retest reliability was computed by Fenigstein
career information (e.g., how to find a job). The 20th item is the et al., with a 2-week period between administrations. The reliability
Barriers scale, which assesses whether or not any of four kinds of coefficients were .79 for Private Self-Consciousness, .84 for Public
factors (e.g., money) have acted as a barrier to career choice. Accord- Self-Consciousness, and .73 for Social Anxiety. Private self-conscious-
ing to Holland et al. (1980), the Occupational Information and ness correlates with thoughtfulness (.48); individuals high in public
Barriers scales resemble checklists more than scales. Holland et al. self-consciousness have been shown to demonstrate high self-blame
(1980) reported internal consistency reliabilities for the MVS of .87 and conformity; and social anxiety correlates negatively with socia-
(Identity), .79 (Occupational Information), and .45 (Barriers). Be- bility. This and other validity information for the SCS is summarized
cause of its low reported reliability, the Barriers scale was not used in in Buss (1980).
this study.
Procedure
Personality Measures Subjects completed the seven questionnaires in groups of approx-
imately 50. After all subjects had arrived in the testing room, modified
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Four instruments were chosen to measure aspects of personality. consent forms were distributed. The experimenter read the consent
This document is copyrighted by the American Psychological Association or one of its allied publishers.

Measures of anxiety, locus of control, and self-esteem were chosen form aloud and then distributed the seven questionnaires. Subjects
on the basis of their empirical relation to career indecision. A measure were asked to provide the following demographic information on
of self-consciousness was included to explore the relationship between their answer sheets before completing the questionnaires: (a) educa-
career indecision and the tendency to self-reflect. tional major if one had been declared, (b) gender, (c) age, and (d)
State-Trait Anxiety Inventory. The STAI (Spielberger et al., year in college. Most subjects completed the questionnaires in ap-
1983) consists of both a State Anxiety scale and a Trait Anxiety scale. proximately 50 min.
The State Anxiety scale asks individuals to indicate how they feel
The seven questionnaires were assembled into four different com-
"right now . . . at this moment," whereas the Trait Anxiety scale asks
binations to balance possible order effects. The seven questionnaires
individuals to indicate how they "generally feel." Each scale consists
consist of three measures of vocational indecision and four measures
of 20 items. Subjects indicated the extent of their agreement with
of personality. The four combinations served to stagger the indecision
each of the items using four response alternatives: almost never,
measures so that no two were ever presented in sequence. The four
sometimes, often, and almost always. Spielberger et al. reported
orders of presentation were as follows (vocational indecision measures
Kuder-Richardson (K-R 20) coefficients for the STAI in three sam-
are italicized): (a) SCS, CDS, JF, CDP, STAI, MVS, IPC; (b) JF,
ples. For the State Anxiety scale, K-R 20 coefficients ranged from .86
CDS, STAI, CDP, IPC, MVS, SCS; (c) STAI, CDS, IPC, CDP, SCS,
to .95. For the Trait Anxiety scale, coefficients ranged from .89 to
MVS, JF; and (d) IPC, CDS, SCS, CDP, JF, MVS, STAI.
.91. Evidence of the concurrent, divergent, and construct validity of
the STAI scales is summarized in the STAI manual (Spielberger et
al., 1983). Statistical Analyses
Internal, Powerful Others and Chance Scales (IPC). The IPC
(Levenson, 1974) contains three measures of locus of control: the Intercorrelations among the scale scores were computed, and a
extent to which an individual feels personal control over his or her one-way analysis of variance (ANOVA) was conducted on the mean
life (Internal scale), the extent to which an individual feels that scale scores to assess any order effects associated with the four
powerful others are in control (Powerful Others scale), and the extent presentations of questionnaires.
to which an individual feels that life events are due to luck or chance Cluster analysis was used as a means to classify and differentiate
(Chance scale). Each scale consists of eight items to be rated on a the subjects in this study. Although there are several different methods
Likert scale ranging from strongly disagree (1) to strongly agree (5). of cluster analysis, Borgen and Weiss (1971) and Borgen and Barnett
Levenson reported K-R 20 reliabilities of .64 (Internal scale), .77 (1987) recommended the Ward (1963) method because of the high
(Powerful Others scale), and .78 (Chance scale). The Internal scale interpretability of the results and its frequency of use in the psycho-
correlates negatively with Rotter's Internal-External Locus of Control logical literature. Participants in this study were classified into clusters
Scale (—.41), whereas the Powerful Others and Chance scales produce on the basis of their standardized scores on 12 of the 19 subscales.
values of .25 and .56, respectively. Further validity research has been Six of the 10 indecision subscales were included in the cluster analysis:
conducted with the IPC scales to assess relationships with achieve- Decidedness (CDP), Self-Clarity (CDP), Knowledge About Occupa-
ment, interpersonal perception, and behavior (e.g., Levenson, 1981). tions (CDP), Decisiveness (CDP), Career Importance (CDP), and
Janis Field Feelings of Inadequacy Scale (JF). The revised ver- Identity (MVS). The remaining four indecision subscales (Comfort
sion of the JF (Janis & Field, 1959; revised by Eagly, 1967) was used [CDP], Certainty [CDS], Indecision [CDS], and Occupational Infor-
in this study to measure self-esteem. Subjects indicate their agreement mation [MVS]) were assigned as external descriptors of the cluster
with each of the 20 items on this scale usingfiveresponse alternatives: solution. The CDS subscales were chosen to be left out of the cluster
very often, fairly often, sometimes, once in a great while, and practi- analysis because it was believed to be useful to have a general measure
cally never. Responses to the items are summed to produce a single of undecidedness to use as an outside descriptor of the cluster analysis.
scale score. Eagly (1967) reported split-half reliabilities of .72 and .88. The Comfort scale (CDP) was chosen to be left out of the cluster
O'Brien (1985) found a correlation of .82 between the Eagly revision solution to allow an outside assessment of the degree of occupational
of the JF and the Rosenberg Self-Esteem Scale (Rosenberg, 1965). comfort associated with each of the identified clusters. The Occupa-
Self-Consciousness Scale (SCS). The SCS (Fenigstein, Scheier, & tional Information scale (MVS) was left out of the cluster analysis
Buss, 1975) assesses an individual's tendency to direct self-attention because of its status as a borderline scale or checklist (Holland et al.,
inward or outward. There are three subscales: Private Self-Conscious- 1980).
ness (10 items; e.g., "I'm always trying to figure myself out"), Public Six of the nine personality subscales were used in the cluster
Self-Consciousness (7 items; e.g., "I'm concerned about what other analysis, including the State Anxiety and Trait Anxiety scales (STAI),
people think of me"), and Social Anxiety (6 items; e.g., "I have the IPC scales, and the Self-Esteem scale (JF). The constructs mea-
trouble working when someone is watching me"). Respondents rate sured by these scales (anxiety, locus of control, and self-esteem) have
74 CONNIE R. WANBERG AND PAUL M. MUCHINSKY

been previously associated with career indecision, as discussed earlier. occurred, because it suggests that researchers cannot auto-
The three SCS subscales (Private Self-Consciousness, Public Self- matically assume that the State Anxiety scale, presented at
Consciousness, and Social Anxiety) were included as exploratory various points in a packet of questionnaires, will produce
variables and thus were only used as external descriptors of the cluster
equivalent results. This is not surprising given the construct
solution.
the scale measures.
The Ward method of cluster analysis forms clusters so that they
have minimum within-group variation and maximum between-
groups variation. Each of the 390 subjects is initially regarded as a Cluster Analysis
cluster, and the number of clusters is systematically reduced in an
iterative and hierarchical manner. Semipartial R2 values were inter- The 12 subscale scores marked for use in the cluster analysis
preted as error terms in making a decision about the number of were standardized, and Ward's method was used to perform
resultant clusters (Statistical Analysis System, 1985). In a plot of the
the clustering process. To determine the optimal number of
error terms by each potential cluster solution, the first point at which
the error term climbs sharply upward can be defined as the optimal clusters, changes in the R2 error term were examined. A
number of clusters. marked increase in the error term was first evident at the
merge from four clusters to three clusters, meaning a great
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

A one-way ANOVA was performed for each of the 12 variables


deal of information had been lost by collapsing four clusters
This document is copyrighted by the American Psychological Association or one of its allied publishers.

used within the cluster analysis across the identified clusters. This
procedure was used to help describe the cluster solution. Likewise, into three (see Figure 1). Because there were no other abrupt
this procedure was performed to test differences between the identified increases in the error term, a four-cluster solution was inter-
clusters on the subscales not included in the cluster analysis. The preted.
effect sizes (as measured by a>2) of variables used within the cluster Table 2 reports the raw subscale means and standard devia-
analysis and those not included were also calculated (Kirk, 1982). tions for each of the four clusters. Effect sizes (or), also shown
The stability of the resultant cluster solution was tested by ran- in Table 2, reveal that the variables of Identity, Trait Anxiety,
domly dividing the total sample into two subsamples (n — 204 and n
Self-Esteem, and Knowledge About Occupations were most
= 186). The cluster-analytic procedure was repeated for each
subgroup. The clusters resulting from these analyses were compared influential in the clustering process. To aid interpretation,
with each other and with the cluster solution based on the entire Figure 2 displays pictorially the standardized subscale means
sample. Finally, a chi-square analysis was conducted on the group for the four clusters, and Table 3 provides a comparison
membership of those students who had declared an academic major among the clusters with the standardized means being classi-
(n = 358) versus those who had not (n = 32). This analysis served to fied into high (.3 or greater), moderate (-.3 to .3), and low
help establish the validity of the resulting cluster solution. (—.3 or lower) categories.
Subjects in Cluster 1 (n = 77; 36 men and 40 women;
19.7% of the total sample) scored high, relative to members
Results of the other clusters, on Decidedness, Self-Clarity, Knowledge
About Occupations, Decisiveness, Identity, Internal Control,
Table 1 shows the intercorrelations among the 19 variables and Self-Esteem. The high scores on the Decidedness, Self-
in this study. Variables 1 through 10 represent the indecision Clarity, Knowledge, and Identity scales indicate that these
scales. Correlations among these scales ranged from . 16 (De- individuals have clear pictures of their interests and talents,
cisiveness and Importance) to —.78 (Indecision and Identity). they seem to be decided about what career they will follow,
The correlations provide evidence that the indecision scales and they believe they are well informed about the occupations
converge in their assessments, although some scales were more and educational programs. The high scores on the Decisive-
centrally related to indecision than others. Correlations be- ness, Internal, and Self-Esteem scales indicate that these in-
tween the indecision scales and the personality scales (Vari- dividuals believe they are able to make decisions without
ables 11 through 19) reveal many statistically significant re- difficulty, they feel they have personal control over their lives,
lationships. However, two of the self-consciousness scales and they generally feel good about themselves.
(Private Self-Consciousness and Public Self-Consciousness) The members of Cluster 1 had low scores on State Anxiety
had the lowest relationships (typically nonsignificant) with and Trait Anxiety, showing that they did not feel tension or
the indecision variables. apprehension at the time of the study and that they do not
The four different orders of questionnaire presentation were tend to be nervous and worrisome in other situations. This
analyzed by performing a one-way ANOVA on the 19 mean group also scored low on Powerful Others and Chance control,
scale scores. One variable (State Anxiety) showed a statistically meaning that there is not a feeling among these individuals
significant (p < .01) order effect. Subjects who completed the that their lives are controlled by people in powerful positions
STAI fifth in the sequence of questionnaires (Order 1, M - or by chance factors. Finally, the individuals in this cluster
34) reported less state anxiety than subjects who completed had intermediate scores on the Career Importance scale,
the STAI first (Order 2, M = 38) or third (Order 3, M = 38) indicating that they feel their future careers and work are at
in the sequence. These results suggest that some individuals least somewhat important to them at this time. This group
were anxious about the study at the start but that, as they was named "confident decided individuals."
completed the questionnaires, their anxiety abated. We felt Members of Cluster 2 (« = 154; 80 men and 72 women;
that this one significant order effect did not confound the 39.5% of the total sample) also scored high on Decidedness
interpretation of the other scales, because our counterbal- and Career Importance and relatively high on Identity. Yet,
anced design served to control for the error contributed by this cluster shows higher levels of state anxiety and trait
order. However, it is of interest that this one order effect anxiety and lower levels of self-clarity, decisiveness, and self-
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

S Si
CAREER DECISION STATUS

|S35S-|Sj5?e
75
76 CONNIE R. WANBERG AND PAUL M. MUCHINSKY

0.25 that choosing an occupation is not as important to this group


as it may be for the others. Cluster 3 members had moderate
scores on personality variables State Anxiety and Trait Anxi-
ety; Internal, Powerful Others, and Chance control; and Self-
Esteem. This group was named "indifferent undecided indi-
viduals."
Like Cluster 3, members of Cluster 4 (n = 60; 25 men and
35 women; 15.4% of the total sample) were career undecided,
as witnessed by their low mean scores on the Identity and
Self-Clarity subscales and their moderate score on the Decid-
edness scale. This group reported a low level of knowledge
about occupational and educational programs (Knowledge
About Occupations scale), and a moderate feeling that their
0.00 career choice is an important issue at this time (Career Im-
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

3 4 5 6 7 8 9 10 11 12 portance scale). In comparison with the other groups, these


This document is copyrighted by the American Psychological Association or one of its allied publishers.

Number of Clusters individuals are characterized by very high scores on State


Anxiety and Trait Anxiety and very low scores on Self-
Figure 1. Semipartial R2 error values by number of clusters. Esteem. This group scored high on the Powerful Others and
Chance scales and low on the Internal scale, meaning that
there is a feeling among these individuals that their lives are
esteem, than does Cluster 1. This group had intermediate controlled by people in powerful positions or by chance
scores on the Knowledge About Occupations, Internal, Pow- factors. This group was named "anxious undecided individ-
erful Others, and Chance scales, showing that members of uals."
this group feel that they have at least a moderate amount of
career information and a moderate belief in being controlled
by internal, powerful others, and chance factors. This group Differentiation of Clusters: External Variables
was named "concerned decided individuals."
Members of Cluster 3 (n = 99; 47 men and 50 women; Table 4 shows the raw mean scores of the seven variables
25.4% of the total sample) had low scores on the Decidedness, (Comfort, Certainty, Indecision, Occupational Information,
Self-Clarity, Decisiveness, and Identity scales. The low scores Public Self-Consciousness, Private Self-Consciousness, and
on these scales indicate that individuals in this cluster are Social Anxiety) that served as the external descriptors of the
unclear about their vocational interests and goals and that four clusters. The external descriptor scales Certainty, Inde-
they have difficulty making decisions. This group also scored cision, and Occupational Information can be used to reiterate
low on the Knowledge About Occupations subscale, indicat- the cluster analysis results. For example, high Certainty and
ing that these students feel they lack information about oc- low Indecision scores support the portrayal of Clusters 1 and
cupational and educational programs. Cluster 3 scored lowest 2 as career decided; low Certainty and high Indecision scores
of the four groups on the Career Importance scale, suggesting support the portrayal of Clusters 3 and 4 as career undecided.

Table 2
Means, Standard Deviations, and Analysis of Variance Results for the Variables Used Within the Cluster Analysis
Cluster 1 Cluster 2 Cluster 3 Cluster 4
(n = 77) (« = 154) (« = 99) (n = 60)
Subscale Range M SD M SD M SD M SD F" or
Decidedness (CDP) 2-16 14.19, 2.91 14.45, 1.79 10.04b 3.48 12.50c 3.20 57.97 .30
Self-Clarity (CDP) 3-24 18.26a 4.92 13.99b 5.14 8.90c 4.47 7.45C 4.10 83.17 .39
Knowledge (CDP) 3-24 19.70a 3.90 15.85b 4.34 10.07, 3.46 11.17, 4.45 100.14 .43
Decisiveness (CDP) 3-24 20.57 a 4.34 17.96b 4.60 14.17C 5.18 11.87d 5.37 48.94 .27
Importance (CDP) 3-24 21.58, 3.50 21.90 a 2.39 18.17b 4.23 19.93C 3.26 29.03 .18
Identity (MVS) 0-18 15.41, 2.79 12.75b 3.19 7.01 c 3.52 6.80c 3.47 145.61 .53
State Anxiety 20-80 28.34a 6.76 35.90, 10.77 37.54 b 10.48 48.42C 10.50 46.04 .26
Trait Anxiety 20-80 30.94a 5.82 38.83 b 8.12 40.20 b 7.65 54.43C 6.99 115.08 .47
Internal 8-40 33.12a 4.51 31.56 b 3.64 3O.13c 3.96 28.90 c 3.80 15.72 .10
Powerful Others 8-40 15.28a 3.72 20.32 b 4.29 20.39 b 3.84 25.87,. 4.25 76.36 .37
Chance 8-40 14.96a 3.79 19.23b 4.17 2O.3Ob 4.09 25.33, 4.13 74.47 .36
Self-Esteem 20-100 81.79, 7.15 71.89 b 8.45 69.00 b 8.48 56.03, 9.36 108.97 .45
Now. Across individual rows, means with different subscripts differ significantly with the Scheffe test (p < .05). Scales were scored such that
higher values indicate more of the construct measured. Indecision subscales are followed by scale labels for easier identification (CDP = Career
Decision Profile; MVS = My Vocational Situation). State Anxiety and Trait Anxiety subscales are from the State-Trait Anxiety Inventory;
Internal, Powerful Others, and Chance subscales are from the Internal, Powerful Others and Chance Scales; and the Self-Esteem scale is from
the Janis Field Feelings of Inadequacy Scale, revised version.
" All F ratios are significant at p < .001.
CAREER DECISION STATUS 77

• Cluster 1 •• — • Cluster 3 repeating the cluster analysis on each subsample separately.


) Cluster 2 o - - -o Cluster 4 Subsample 1 contained 204 individuals, whereas Subsample
2 contained 186 individuals. Although the results of these
1.50
cluster analyses are not presented in tabular form, to conserve
space, the results are summarized as follows.
A four-cluster solution clearly emerged in Subsample 2,
whereas both three-cluster and four-cluster solutions were
interpretable in Subsample 1. Because a four-cluster solution
c
CO in Subsample 1 was defensible, and to be consistent with the
other cluster solution, the results for both subsamples were
CO interpreted in a four-cluster solution. Cluster 1 (confident
c decided individuals), Cluster 3 (indifferent undecided individ-
CO
CO uals), and Cluster 4 (anxious undecided individuals) from the
total sample all clearly replicated in both subsamples. How-
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

ever, Cluster 2 (concerned decided individuals) from the total


This document is copyrighted by the American Psychological Association or one of its allied publishers.

sample did not replicate as clearly in the subsamples. One


[1
variable, Decisiveness, served to cloud the results. The indi-
a.
n a n a> —
viduals in Subsample 2 had a moderate level of decisiveness,
o o o _ O
To
whereas the individuals in Subsample 1 exhibited a lower
E O UJ
nee
ess

<D

level of decisiveness. Overall, we conclude that the cluster


"8
01 (75
-o
results from the total sample did replicate well in the subsam-
o ples, with the exception of Cluster 2 in relation to the variable
Decisiveness.
Scales Included in the Cluster Analysis A chi-square analysis was constructed on the group mem-
bership of those students who had declared an academic major
Figure 2. Standardized means of Cluster 1, Cluster 2, Cluster 3, and (n = 358) versus those who had not (n = 32). This analysis
Cluster 4. (CDP = Career Decision Profile; MVS = My Vocational served to help establish the validity of the total sample cluster
Situation. State Anxiety and Trait Anxiety subscales are from the solution. Of the 32 individuals who had not declared an
State-Trait Anxiety Inventory; Internal, Powerful Others, and Chance academic major, 26 (81%) were members of one of the two
subscales are from the Internal, Powerful Others and Chance Scales;
and the Self-Esteem scale is from the Janis Field Feelings of Inade-
undecided groups. Of the 358 individuals who had declared
quacy Scale, revised version.) an academic major, 224 (63%) were members of one of the
two decided groups, x2(3, N = 390) = 23.07, p < .001.

Also, according to the Occupational Information scale, mem-


bers of Cluster 1 feel they have a high amount of occupational Discussion
information; Clusters 2 and 3, moderate levels; Cluster 4, low
levels. The cluster variables showed the same pattern of Jones and Chenery (1980) developed the vocational deci-
results. sion status model to explain the concept of career indecision.
The external variables Comfort, Public Self-Consciousness, Jones (1989a) revised the Vocational Decision Scale, on which
Private Self-Consciousness, and Social Anxiety were used to the model was based, to form the CDP. The model suggests
provide additional information about the clusters. High Com- that it may be useful to classify individuals into four groups
fort scale scores for Cluster 1 members indicate that these of vocational decision status: (a) decided-comfortable, (b)
individuals are comfortable with their place in the vocational decided-uncomfortable, (c) undecided-comfortable, and (d)
choice process, but Cluster 2 members showed only moderate undecided-uncomfortable. It is interesting that both our study
levels of comfort. Members of Clusters 3 and 4 seemed and Jones's study reported finding four types of individuals,
uncomfortable with their vocational decision status; their two that seem to be career decided and two that seem to be
mean comfort scores were significantly lower than those of career undecided. Although our cluster-analytic study was not
Clusters 1 and 2. Private Self-Consciousness scores show that intended to be an explicit test of the Jones and Chenery
there are no significant differences among the four groups in model, it seems their model provides a basis for juxtaposing
their tendency to self-reflect. Low Public Self-Consciousness our findings. In comparing our results with their proposed
and Social Anxiety scores for Cluster 1 show that these typology, we found similarities. Yet, some major empirical
individuals tend not to be shy or anxious in social situations. and conceptual distinctions also emerged.
Clusters 2 and 3 scored in the moderate range on these two Both studies identified the existence of a group of individ-
variables, whereas Cluster 4 showed higher levels of Public uals who have clear vocational interests and who appear
Self-Consciousness and Social Anxiety. content with their decision. Jones and Chenery (1980) labeled
this group decided-comfortable individuals, whereas we la-
Stability and Validity of Cluster Solution beled this group confident decided individuals (Cluster 1).
These individuals are characterized by a strong sense of career
The stability of the four-cluster solution was addressed by decidedness, self-clarity, vocational identity, and self-esteem.
randomly dividing the total sample into two subsamples and Both studies also identified the existence of a group of indi-
78 CONNIE R. WANBERG AND PAUL M. MUCHINSKY

Table 3
Description of Clusters: High, Medium, and Low Mean Standard Scale Scores
High Intermediate Low
Cluster standard scores standard scores standard scores
(.3 and higher) (.3 to -.3) (-.3 and lower)
Cluster 1 Decidedness (CDP) Importance (CDP) State Anxiety
Self-Clarity (CDP) Trait Anxiety
Knowledge (CDP) Powerful Others
Decisiveness (CDP) Chance
Identity (MVS)
Internal
Self-Esteem

Cluster 2 Decidedness (CDP) Self-Clarity (CDP)


Importance (CDP) Knowledge (CDP)
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Identity (CDP) Decisiveness (CDP)


This document is copyrighted by the American Psychological Association or one of its allied publishers.

State Anxiety
Trait Anxiety
Internal
Powerful Others
Chance
Self-Esteem

Cluster 3 State Anxiety Decidedness (CDP)


Trait Anxiety Self-Clarity (CDP)
Internal Knowledge (CDP)
Powerful Others Decisiveness (CDP)
Chance Importance (CDP)
Self-Esteem Identity (CDP)

Cluster 4 State Anxiety Decidedness (CDP) Self-Clarity (CDP)


Trait Anxiety Importance (CDP) Knowledge (CDP)
Powerful Others Decisiveness (CDP)
Chance Identity (MVS)
Internal
Self-Esteem
Note. Indecision subscales are followed by scale labels (CDP = Career Decision Profile; MVS = My
Vocational Situation). State Anxiety and Trait Anxiety subscales are from the State-Trait Anxiety
Inventory; Internal, Powerful Others, and Chance subscales are from the Internal, Powerful Others and
Chance Scales; and the Self-Esteem scale is from the Janis Field Feelings of Inadequacy, revised version.

viduals who have unclear vocational interests and are not of vocational identity, self-clarity, and self-esteem and high
content with their state of indecision. Jones and Chenery levels of state and trait anxiety.
labeled the group undecided-uncomfortable individuals, However, it is with regard to the final two groups that the
whereas we labeled this group anxious undecided individuals two studies produced the greatest conceptual differences.
(Cluster 4). These individuals are characterized by low levels Jones (1989a) described a group he labeled as decided-un-

Table 4
Summary of Significant Differences Between Clusters on External Descriptor Scales
Cluster 1 Cluster 2 Cluster 3 Cluster 4
(« = 77) (n= 154) (n = 99) (« = 60)
Subscale Range M M M M F w2
Comfort (CDP) 2-16 13.79a 11.92b 7.90c 9.31c 64.03* .33
Certainty (CDS) 2-8 6.95a 6.60. 4.8 l b 5.62c 44.27* .25
Indecision (CDS) 16-64 23.69a 27.18 b 38.34c 35.90c 85.67* .39
Occupational Information (MVS) 0-4 2.78a 1.84b 0.91c 0.72c 44.35* .25
Public Self-Consciousness 7-35 24.65 a 26.38 ab 26.45 ab 27.83b 4.20* .02
Private Self-Consciousness 10-50 33.04a 33.51 a 33.38a 35.20a 2.15 .01
Social Anxiety 6-30 14.74a 17.45b 18.45b 21.42t. 25.69* .16
Note. Across individual rows, means with different subscripts differ significantly with the Scheffe test
(p < .05). Scales were scored such that higher values indicate more of the construct measured. Indecision
subscales are followed by scale labels for easier identification (CDP = Career Decision Profile; CDS =
Career Decision Scale; MVS = My Vocational Situation). Public Self-Consciousness, Private Self-
Consciousness, and Social Anxiety subscales are from the Self-Consciousness Scale. For all F statistics,
df= 3, 386.
CAREER DECISION STATUS 79

comfortable. We described a group we labeled concerned tween our two groups of undecided individuals (indifferent
decided individuals (Cluster 2). Both studies found that these undecided and anxious undecided) on the Self-Clarity and
individuals had decided on a career, but we disagree as to the Knowledge About Occupations factors. The two reason fac-
appropriate descriptor for the affective dimension. In our tors that did differentiate our two undecided groups were
study, these individuals indicate self-doubt about the correct- Career Importance and Decisiveness (Table 2). Overall, we
ness of their vocational choice. Yet, we did not identify these found the reason scales useful. Many career indecision scales
individuals as uncomfortable with their vocational decision have not provided subscales to indicate possible reasons for
status, but rather as concerned. Although they show higher indecision. Although more research on these reason scales
levels of anxiety and lower levels of self-esteem than the would be beneficial, future work in this area should also focus
individuals in Cluster 1 (the confident decided individuals), on how the career counselor might best use this information.
their levels of anxiety and self-esteem are not as pronounced
as those of the members of Cluster 4 (the anxious undecided
group). These concerned decided individuals also scored sig-
nificantly higher (more comfortable) than the two undecided Conclusion and Comments
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

groups on the CDP Comfort scale (see Table 4).


This document is copyrighted by the American Psychological Association or one of its allied publishers.

The group with the fewest members identified by Jones This study included both career decided and career unde-
(1989a) was labeled undecided-comfortable. In our study, we cided subjects in a cluster-analytic investigation of career
labeled members of Cluster 3 indifferent undecided. Both decision status subtypes. Our study showed what may be
studies found that these individuals had yet to settle on a clear considered as clear support for Jones and Chenery's (1980)
vocational path, but again we differ on the descriptor for the decided-comfortable and undecided-uncomfortable groups.
affective dimension. We did not find these individuals to be Yet, we did not find a group clearly identified as decided-
comfortable in their indecision, but rather they seemed not uncomfortable or undecided-comfortable. Because little em-
to care or mind that they had yet to settle on a career choice. pirical research has assessed the typology proposed by Jones
Although they had low scores on Identity, Decidedness, and and Chenery, perhaps further studies should attempt to un-
Decisiveness, they exhibited moderate scores on Trait Anxi- derstand whether their model is a meaningful way to type
ety, State Anxiety, and Self-Esteem. They are less anxious individuals according to career decision status.
than the members of Cluster 4 (anxious undecided) and more Our study suggests that it is useful to include both unde-
anxious than the members of Cluster 1 (confident decided). cided and decided individuals in a typology. For example,
What differentiates the members of Cluster 2 (concerned most individuals in our study were members of a group that
decided) and Cluster 3 (indifferent undecided) is not only is rarely discussed in the vocational and counseling literature,
their scores on the vocational decision scales but also the that is, concerned decided individuals. Although we disagree
scores on Career Importance and Knowledge About Occu- that these individuals are uncomfortable with their career
pations (Tables 2 and 4). Indifferent undecided individuals decision status (e.g., Jones & Chenery), we do agree that they
lack information about occupational and educational pro- may need career counseling. These individuals have settled
grams, yet they feel their career choice is less of an important on a vocational path, yet exhibit self-doubt over their decision.
issue to them. The emergence of this group may be attributable to social and
Although it may be tempting to attribute the differences cohort effects placed on freshmen to select a major and be
between the Jones and Chenery (1980) model and the present committed to a vocational goal early in their college career.
study to semantics, we think that this is not the case. We It is our experience that many freshmen select majors partly
agree that confident and comfortable are equivalent in concep- on the basis of their assessment of anticipated favorable job
tual meaning, as are anxious and uncomfortable. Thus, Clus- market conditions on graduation. Rather than selecting vo-
ters 1 and 4 in our study seem to show strong support for cational paths on the basis of their own interests and abilities,
Jones and Chenery's decided-comfortable and undecided- these students' choices are more heavily influenced by per-
uncomfortable groups. However, we do not believe that in- ceived opportunities for employment. This sentiment is in
dividuals who are concerned are necessarily uncomfortable, contrast to student cohorts of the past who seemed somewhat
nor are indifferent individuals necessarily comfortable. We oblivious to the economic realities of obtaining employment
believe these latter terms are not synonymous but reflect in certain majors. Much of the popularity of business as a
differences in the substantive significance of the meaning of major today is attributed to the perception of readily accessi-
our identified clusters. Thus, we do not believe our cluster ble employment on graduation. We believe that a significant
results directly support Jones and Chenery's decided-uncom- proportion of individuals in this concerned decided group
fortable and undecided-comfortable groups. represent those who are questioning the soundness of their
Jones (1989a) also suggested that it may be useful to differ- decision to pursue a vocation for the types of reasons described
entiate individuals according to four reasons they may have herein. We also believe that this group is, in part, responsible
for being undecided. His CDP measure, used in this study, for the high percentage of college students who switch aca-
includes four scales concerned with reasons for indecision: demic majors. In short, we hypothesize that those who switch
Self-Clarity, Knowledge About Occupations, Decisiveness, majors are composed not only of undecided students but also
and Career Importance. Jones reported some overlap among of students who become disenchanted with the major to which
the four reason scales in his study. He showed, for example, they once were committed. We believe it could be particularly
that some individuals score low on more than one reason advantageous to study this group of concerned decided indi-
scale. In our study, we found nonsignificant differences be- viduals in a longitudinal design.
80 CONNIE R. WANBERG AND PAUL M. MUCHINSKY

References Kirk, R. E. (1982). Experimental design. Monterey, CA: Brooks/


Cole.
Borgen, F. H., & Barnett, D. C. (1987). Applying cluster analysis in Larson, L. M., Heppner, P. P., Ham, T., & Dugan, K. (1988).
counseling psychology research. Journal of Counseling Psvchologv, Investigating multiple subtypes of career indecision through cluster
34, 456-468. analysis. Journal of Counseling Psychology, 35, 439-446.
Borgen, F. H., & Weiss. D. J. (1971). Research methodology: Cluster Levenson, H. (1974). Activism and powerful others: Distinctions
analysis and counseling research. Journal ofCounseling Psvchologv, within the concept of internal-external control. Journal of Person-
75,583-591. ality Assessment, 38, 377-383.
Buss, A. H. (1980). Self consciousness and social anxiety. San Fran- Levenson, H. (1981). Differentiating among internality, powerful
cisco: Freeman. others, and chance. In H. M. Lefcourt (Ed.), Research with the
Crites, J. O. (1969). Vocalional psychology: The study of vocational locus of control construct (Vol. 1, pp. 15-63). San Diego, CA:
behavior and development. New York: McGraw-Hill. Academic Press.
Eagly, A. H. (1967). Involvement as a determinant of response to Lucas, M. S., & Epperson, D. L. (1990). Types of vocational unde-
favorable and unfavorable information. Journal of Personality and cidedness: A replication and refinement. Journal of Counseling
Social Psychology Monographs, 7(3, Whole No. 643). Psychology, 37, 382-388.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Fenigstein. A., Scheier, M. F., & Buss, A. H. (1975). Public and Maier, D., & Herman, A. (1974). The relationship of vocational
This document is copyrighted by the American Psychological Association or one of its allied publishers.

private self-consciousness: Assessment and theory. Journal of Con- decidedness and satisfaction with dogmatism and self-esteem. Jour-
sulting and Clinical Psychology. 43, 522-527. nal of Vocational Behavior, 5, 95-102.
Fuqua, D. R., Blum, C. R., & Hartman. B. W. (1988). Empirical O'Brien, E. J. (1985). Global self-esteem scales: Unidimensional or
support for the differential diagnosis of career indecision. Career multi-dimensional? Psychological Reports, 57, 383-389.
Development Quarterly, 36, 365-373. Osipow, S. H., Carney, C. G., & Barak, A. (1976). A scale of
Fuqua, D. R., Seaworth, T. B., & Newman, J. L. (1987). The educational-vocational undecidedness: A typological approach.
relationship of career indecision and anxiety: A multivariate ex- Journal of Vocational Behavior, 9, 233-243.
amination. Journal of Vocational Behavior, 30, 175-186. Osipow, S. H., Carney, C. G., Winer, J. L., Yanico, B., & Koschier,
Goodstein, L. D. (1965). Behavior theoretical views of counseling. In M. (1976). The Career Decision Scale (Rev. ed.). Odessa, FL:
B. Stefflre (Ed.), Theories of counseling (pp. 140-192). New York: Psychological Assessment Resources, Inc.
McGraw-Hill. Rosenberg, M. (1965). Society and the adolescent self-image. Prince-
Hartman, B. W., Fuqua, D. R.. & Hartman, P. T. (1983). The ton, NJ: Princeton University Press.
predictive potential of the career decision scale in identifying Rotter, J. B. (1966). Generalized expectancies for internal versus
chronic career indecision. Vocational Guidance Quarterly, 32, 103- external control of reinforcement. Psychological Monographs,
108. 80(\, Whole No. 609).
Holland, J. L., Daiger, D. C, & Power, P. G. (1980). My Vocational Salomone, P. R. (1982). Difficult cases in career counseling: II—The
Situation. Palo Alto, CA: Consulting Psychologists Press. indecisive client. Personnel and Guidance Journal, 60, 496-500.
Holland, J. L., Gottfredson, D. C, & Nafziger, D. H. (1975). Testing Savickas, M. L., & Jarjoura, D. (1991). The Career Decision Scale as
the validity of some theoretical signs of vocational decision-making a type indicator. Journal of Counseling Psychology, 38, 85-90.
ability. Journal of Counseling Psychology, 22, 411-422. Slaney, R. B. (1988). The assessment of career decision making. In
Holland, J. L., & Holland, J. E. (1977). Vocational indecision: More W. B. Walsh & S. H. Osipow (Eds.) Career decision making (pp.
evidence and speculation. Journal of Counseling Psychology, 24, 32-72). Hillsdale, NJ: Erlbaum.
404-414. Spielberger, C. D., Gorsuch, R. L., Lushene, R. E., Vagg, P. R., &
Janis, I. L., & Field, P. B. (1959). Sex differences and personality Jacobs, G. A. (1983). State-Trait Anxiety Inventory manual. Palo
factors related to persuasibility. In C. I. Hovland & I. L. Janis Alto, CA: Consulting Psychologists Press.
(Eds.), Personality and persuasibility (pp. 55-68). New Haven, CT: Statistical Analysis System. (1985). SAS user's guide: Statistics. Cary.
Yale University Press. NC: SAS Institute, Inc.
Jones, L. K. (1989a). Measuring a three-dimensional construct of Tyler, L. E. (1961). The work of the counselor (2nd ed.). New York:
career indecision among college students: A revision of the Voca- Appleton-Century-Crofts.
tional Decision Scale—The Career Decision Profile. Journal of Ward, J. H. (1963). Hierarchical grouping to optimize an objective
Counseling Psychology, 36, 477-486. function. Journal of the American Statistical Association, 58, 236-
Jones, L. K. (1989b). The Career Decision Profile. Unpublished scale. 244.
(Available from Lawrence K. Jones, North Carolina State Univer-
sity, Box 7801, Raleigh, North Carolina 27695)
Jones, L. K., & Chenery, M. F. (1980). Multiple subtypes among Received December 17, 1990
vocationally undecided college students: A model and assessment Revision received June 10, 1991
instrument. Journal of Counseling Psychology, 27, 469-477. Accepted June 17, 1991

You might also like