Facilitating Career
Decision-Making
Itamar Gati
The Hebrew University of Jerusalem
In this presentation, I will
Discuss the decision-theory viewpoint
 Present the PIC 3-stage cdm model
 Introduce the CDDQ
 Describe the CDSQ – cdm style
 Demonstrate MBCD - Making Better Career Decisions
 Review research and demonstrate applications
 Highlight the unique features of our approach

2
Unique features of career decisions

Quantity of Information:
Often large N of alternatives and factors, within-occupation
variance  information is practically unlimited

Quality of Information:
soft, subjective, fuzzy, inaccurate, biased

Uncertainty about:
the individual’s future preferences, future career options,
unpredictable changes and opportunities, probability of
implementing choice

Non-cognitive Factors:
emotional and personality-related factors, necessity for
compromise, actual or perceived social barriers and biases
3
From decision theory to
career counseling practice

Many factors contribute to the complexity and
difficulties involved in career decision-making
The basic claim:

Career counseling may be viewed as
decision counseling, which aims at facilitating
the clients' decision-making process, and
promotes better career decisions
4
If so evident, why was decisiontheory not adopted until recently?
Because
 Normative decision theory (how individuals should
make decisions) is –





too rational
too arbitrary
too quantitative
exceeds human’s information-processing capability
Descriptive decision theory (how individuals
actually make decisions) is not helpful either –
it mainly documents human weakness


heuristics, biases, and fallacies
limited information-processing capabilities
5
The Proposed Approach
–

By adopting decision theory and adapting it to the
unique features of career decisions, theoretical
knowledge can be translated into practical
interventions to facilitate individuals’ career choices

Specifically, we suggest focusing on a prescriptive
approach, and designing systematic procedures
that can help individuals make better career
decisions (not necessarily rational ones!)
6
The first stage in helping clients is
needs assessment:
The 3 components of needs assessment are:

the individual’s stage in the cdm process
(“where”)

the focuses of the individual’s cdm difficulties
(“what”)

the individual’s cdm style (“who”)
7
I-
Stages in the career
decision-making process
The PIC model (Gati & Asher, 2001)
separates the career decision-making
process into 3 distinct stages:
- Prescreening
- In-depth exploration
- Choice
8
Prescreening


Goal: Locating a small set (about 7) of promising
alternatives that deserve further, in-depth
exploration
Method: Sequential Elimination





Locate and prioritize relevant aspects or factors
Explicate within-aspect preferences
Eliminate incompatible alternatives
Check list of promising alternatives
Outcome: A list of verified promising
alternatives worth further, in-depth exploration
9
A Schematic Presentation of the
Sequential Elimination Process
(within-aspects, across-alternatives)
Potential Alternatives
Aspects
a
b
1
2
3
4
.
.
.
.
N
(most
important)
(second in
importance)
c
.
n
Promising
Alternatives
10
Final step - Sensitivity Analysis
The Goal:
Verifying the adequacy of the promising list
The Method:





An alternative (compensatory-model-based)
search
“why not”
“almost compatible”
“what if”
“similar alternatives”
11
In-depth exploration

Goal: Locating alternatives that are not only promising but
indeed suitable for the individual

Method: collecting additional information, focusing on one
promising alternative at a time:



Is the occupation INDEED suitable for me?
 verifying compatibility with one’s preferences in the most
important aspects
 considering compatibility within the less important aspects
Am I suitable for the occupation?
 probability of actualization: previous studies, grades,
achievements
 fit with the core aspects of the occupation
Outcome: A few most suitable alternatives (about 3-4)
12
A Schematic Presentation of the
In-depth Exploration Stage
(within-alternative, across aspects)
Promising Alternatives
1
2
2
3
4
5
4
5
6
Suitable Alternatives
13
Choice


Goal: Choosing the most suitable alternative, and rankordering additional, second-best alternatives
Method:


comparing and evaluating the suitable alternatives
pinpointing the most suitable one
 Am I likely to activate it?



if not - selecting second-best alternative(s)
if yes - Am I confident in my choice?
 if not: Return to In-depth exploration stage
 if yes: Done!
Outcome: The best alternative or a rank-order of the
best alternatives
14
II Career
Decision-Making Difficulties

One of the first steps in helping individuals make
a career decision is locating the focuses of the
difficulties they face in the decision-making
process

Relying on decision theory, Gati, Krausz, and
Osipow (1996) proposed a taxonomy for
describing career decision-making difficulties
15
Possible Focuses of
Career Decision-Making Difficulties
(Gati, Krausz, & Osipow, 1996)
During the Process
Prior to Engaging
in the Process
Lack of Readiness
due to
Lack of
Indecimotivation siveness
Lack of Information
about
Dysfunc- Cdm Self Occupations
tional process
beliefs
Ways of
obtaining
info.
Inconsistent
Information due to
Unreliable Internal
Info.
conflicts
External
conflicts
16
The Career Decision-making
Difficulties Questionnaire (CDDQ)

The Career Decision-making Difficulties
Questionnaire (CDDQ) was developed to test this
taxonomy and serve as a means for assessing
individuals’ career decision-making difficulties

Cronbach Alpha internal consistency estimate of the
total CDDQ score is high (above .90)

The proposed structure was empirically supported
(N=10,000)

For additional information – see www.cddq.org
--- the CDDQ is offered free of charge --17
www.cddq.org
18
19
20
The Four Stages of Interpretation
1.
Ascertaining Credibility, using validity items and the
time required to fill out the questionnaire
2.
Estimating Differentiation based on the standard
deviation of the 10 difficulty-scale scores
3.
Locating the salient, moderate, or negligible
difficulties, based on the individual's absolute and
relative scale scores
4.
Determining the confidence in the feedback
and the need to add reservations to it
(based on doubtful credibility, partial differentiation, or low
informativeness)
21
The 4 Stages of Interpretation
1
Doubtful
2
Credible
Estimating
Differentiation
Questionable
3
Aggregate
Reasons to Add
Reservation (RAR)
B/W < 1
RAR = 3
RAR ≤ 2
4
Add Reservation
to Feedback
Not Credible
Evaluating
Credibility
Low
High
Locate Salient
Difficulties
Compute
Informativeness
(B /W )
B/W > 1
Receives
Feedback
No
Feedback
22
Four Studies for validating the proposed interpretation
Method
Participants: 15-30 career counselors and 25-80
graduate counseling students
 Questionnaires – including CDDQ responses:
- in Study 1 and 4 – all possible responses;
- in Studies 2 and 3 – responses of 16 actual clients


Results:

High similarity within-groups as well as between
counselors’ and students’ judgments High similarity
between the experts’ judgments and the proposed
algorithm at each stage
23
The distribution of types of feedback
in the four groups
100%
90%
80%
feedback
add reservation
70%
60%
50%
no feedback
40%
30%
20%
10%
0%
P&P
Internet
Hebrew
P&P
Internet
English
24
Conclusions

The incorporation of an intermediate
level of discrimination increases the
usefulness of the feedback and decreases
the chances and implications of potential
errors

Adding reservations when appropriate is
essential for providing a meaningful
feedback and decreasing the chances of
misleading conclusions
25
III –
Career Decision-Making
Styles

Diagnosing the client’s career decision-making
style is important in order to “tailor” the careercounseling intervention to his or her unique
characteristics

Previous research often did not take into
consideration the complexity and variety of
aspects related to the decision process, and
classified decision-styles based only on a
single, most dominant characteristic (e.g.,
rational vs. intuitive)
26
Goals

Developing a multidimensional model for
describing career decision-making styles

Developing the Career Decision-making Styles
Questionnaire (CDSQ) for testing the model and
enabling a more accurate assessment of individuals’
career decision-making styles

Empirically deriving a typology of the CDSQ
profiles from a large sample of individuals
27
Derivation of the 11 Dimensions

Comparing the most common 12 prototypes deduced
from previous research to uncover the various
characteristics differentiating among them

From this list we derived 11 basic dimensions relevant
for characterizing individuals' cdm styles.
On each dimension, individuals can be characterized
along a continuum of a bipolar scale:
e.g., on the dimension pattern of information processing
individuals can be characterized from "analytical" to
"holistic"; desire to please others – "high" to "low"
28
The 11 Proposed Dimensions











Information processing (analytic vs. holistic)
Information gathering (much vs. little)
Amount of effort invested in the process (much vs. little)
Consultation with others (frequent vs. rare)
Aspiration for an "ideal occupation" (high vs. low)
Willingness to compromise (high vs. low)
Locus of control (internal vs. external)
Procrastination in entering the process (high vs. low)
Speed of making the final decision (fast vs. slow)
Dependence on others (high vs. low)
Desire to please others (high vs. low)
29
The Career-Decision-making Style
Questionnaire (CDSQ)

44 statements (4 items x 11 dimensions)

Response scale: 1 – Strongly disagree to
7 – Strongly agree

The CDSQ is embedded in career-related selfhelp Internet sites Future Directions (Hebrew),
CDDQ.ORG (English)

3 Development samples (N=230, 404, 411)

Fourth sample - 479 subjects
30
Results – (Items)
Scale Reliabilities:
 median - .80, range .73 – .85
Factor analysis:
 10 factors
 Accounted-for Variance = .65
 2 dimensions were included in one factor
(Speed of making the final decision; Procrastination)
 Two items loaded higher on a “neighbor factor” (Informationprocessing; effort invested)
Cluster analysis:
 Accounted-for Variance = .81
 Items of 7 dimension clustered perfectly (4/4)
4 dimension – 3/4 items
31
Conclusions & Implications

The proposed and tested 11 dimensions can be
used to characterize individuals' career decisionmaking styles

Using the CDSQ, homogeneous groups of clients
with similar career decision-making styles can be
empirically identified

The CDSQ allows a more accurate assessment of
the counselees' career decision-making styles, thus
better “tailoring” the intervention to the individual

The CDSQ allows individuals to learn about their
career decision-making style, and thus to consider
adopting more desirable strategies
32
So far, I reviewed
3 components of client’s needs assessment:

The individual’s stage in the cdm process (“Where”)

The focuses of the individual’s cdm difficulties (“What”)

The individual’s cdm style (“Who”)
So, what’s next?

Some demonstrations of how can the decision-making
approach be implemented in order to actually facilitate
clients’ cdm
33

Specifically,
if career decision-making requires collecting
a vast amount of information, and
if complex information-processing is needed,

we must then utilize the best available
resource:
Career counselors’ expert knowledge, that can
be elicited and transformed into Information and
Communication Technology-based systems

Indeed,
- The computer-assisted career guidance
systems, based on a decision-theory model,
can help overcome human’s cognitive limitations
- There are several computer-assisted career
guidance systems available today on the Internet
34
MBCD
Making Better Career Decisions
MBCD is an Internet-based career planning
system that is a unique combination of

a career-information system
 a decision-making support system
 an expert system
Based on the rationale of the PIC model,
MBCD is designed to help deliberating
individuals make better career decisions
35
Making Better Career Decisions
http://mbcd.intocareers.org
36
37
38
39
40
41
However,
Although Internet-based, career-related self-help
sites are flourishing, these sites vary greatly in
quality
Therefore,
it is very important to investigate the utility and
validity of these self-help programs
42
So,
Making
Better
Career
Decisions
Does it really work?
43
Criteria for Testing the Benefits of
Making Better Career Decisions
•
Examine users' perceptions of MBCD
•
Examine changes in user’s decision status
•
Examine perceived benefits
•
Locate factors that contribute to these variables
44
MBCD’s Effect (Cohen’s d) on Reducing
Career Decision-Making Difficulties
(Gati, Saka, & Krausz, 2003)
0.8
0.72
0.65
d
0.7
0.6
0.5
0.4
0.31
0.3
0.2
0.11
0.1
0
Lack of
Readiness
Lack of
Information
Inconsistent
Information
Total CDDQ
45
Decision Status
Before and After the “Dialogue” with MBCD
Before the dialogue
After the dialogue
1
2
3
4
5
1- no direction
34
7
6
7
0
2 - only a general
direction
41
66
15
9
5
3 - considering a
few specific alternatives
27
58
84
30
6
4 - would like to examine
additional alternatives
23
51
35
54
6
5 - would like to collect
information about a
specific occupation
646- sure which
occupation to choose
9
20
21
41
28
3
0
1
9
16
Perceived Suitability of the "Promising Alternatives" List
(N=693)
100%
90%
too long
80%
70%
60%
50%
suitable
40%
30%
too short
20%
10%
0%
2
3-4
5
6
7
8-10 11-15 16-25
26+
(n=23) (n=74) (n=71) (n=121) (n=236) (n=45) (n=40) (n=46) (n=37)
Number of Alternatives
(n - of users)
47
Predictive Validity of MBCD
(Gati, Gadassi, & Shemesh, 2006)

Design:
Comparing the Occupational Choice Satisfaction
(OCS) of two groups six years after using MBCD
and getting a list of occupations recommended for
further exploration:

those whose present occupation was
included in MBCD’s recommended list (44%)

those whose present occupation was not
included in MBCD’s recommended list (56%)
48
Method

Participants
 The original sample included 123 clients who
used MBCD in 1997, as part of their counseling
at the Hadassah Career-Counseling Institute
 Out of the 73 that were located after six+
years, 70 agreed to participate in the follow-up:
44 women (64%) and 26 men (36%),
aged 23 to 51 (mean = 28.4, SD = 5.03)
49
Frequencies of Occupational Choice Satisfaction
by “Acceptance” and “Rejection” of MBCD's
Recommendations (Gati, Gadassi, & Shemesh, 2006)
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
16%
18%
low satisfaction
44%
medium satisfaction
high satisfaction
84%
38%
accepted
did not accept
recommendations
recommendations
50
Gender Differences in Directly Elicited and
Indirectly Derived Preferred Occupations
(279 Women + 79 Men, Mean Age=23; Gadassi & Gati, 2008)
Means of the Gender Dominance Ratings
According to Type-of-List and Gender
3.18
3.13
2.96
Men
Women
2.71
Indirectly Derived
3.3
3.2
3.1
3
2.9
2.8
2.7
2.6
2.5
2.4
Directly Elicited
51
Summary of Major Findings

PIC is compatible with people’s intuitive ways of
making decisions (Gati & Tikotzki, 1989)

Most users report progress in the career decisionmaking process (Gati, Kleiman, Saka, & Zakai, 2003)
 Satisfaction
was also reported among those who did not
progress in the process
 Users are “goal-directed” – the closer they are to making a
decision, the more satisfied they are with MBCD

The list of “recommended” occupations are less
influenced by gender stereotypes (Gadassi & Gati, 2008)
52
In Conclusion –
Features of our Approach

Prescreening is essential when the number of
potential alternatives is large

Instead of focusing on occupations (alternatives)
we suggest to focus on aspects

Instead of a “snap-shot” assessments of vocational
interests (e.g., the 3-highest RIASEC Holland’s
code), use for prescreening a wide range of factors
elicited by a dynamic, interactive process
53
In Conclusion –
Features of our Approach (cont.)

From the viewpoint of the individual, this enables:
- Differentiating between relative importance of
factors, the optimal level, and the willingness to
compromise
- Assessing the individual’s preference crystallization
(does s/he knows what s/he is looking for?)

With respect to occupations, this enables:
- Characterizing occupations in terms of a range of
levels, representing the within-occupation variance
- Highlighting the essence of the occupation (using
the core aspects)
54
We believe that . . .

Computers can and should be used not only for
scoring, but also for monitoring a dynamic interaction,
and providing flexible interpretations

Experts’ knowledge can and should be elicited and
transformed to design and improve interpretive
feedbacks on assessments

Career choices are the outcome of decision-making
processes; therefore, career counseling is, in fact,
decision counseling

The goal should be promoting a systematic decision
making process – not a rational one
55
Finally, we also believe that . . .

Career-related assessments can be transformed
into user-friendly Internet-based systems, which
can also be incorporated into counseling
interventions

Interpretive feedback is important but has to be
“tailored” and validated

Theory-based interventions should always be
tested for empirically validity as well as practical
effectiveness
56
www.cddq.org
itamar.gati@huji.ac.il
57
end
--
58

------sfsfsf------------
59
Previous Research
1. 39 labels used for describing decisionmaking styles were located
2. In light of the high resemblance among some of them
(e.g., logical [Arroba, 1977], rational [Harren, 1979],
active-planning [Jepsen, 1974], systematic [Johnson,
1978]), these 39 types were narrowed down
to 12 prototypes :
rational, perfectionist, procrastinator,
searching for tools, satisfying, hesitant,
impulsive, fatalist, intuitive, dependent,
rebellious, and pleasing.
60
Alternative Explanations [to MH] –
were not supported

Differences in the lengths of the lists
No difference was found in the OCS between clients
whose list included 15 or fewer occupations and clients
whose list included more than 15 occupations

Clients who accepted MBCD’s
recommendations are more compliant,
and therefore more inclined to report a high
level of satisfaction
However, following the compensatory-model-based
recommendations did not contribute to the OCS
61
Results - Typology
Group G1
G2
G3
G4
G5
G6
G7
DIMENSION
n=38
n=77
n=79 n=65 n=47
n=87
n=64
Information-processing
3.65
4.91
4.76
5.17
3.45
5.35
5.09
Information gathering
3.94
5.21
5.34
5.83
3.43
5.94
5.79
Amount of effort invested
4.82
6.06
6.05
5.41
4.62
6.32
5.72
Consultation with others
5.64
5.58
6.26
4.02
5.33
5.64
5.83
"ideal occupation"
4.12
2.01
3.12
2.25
3.39
2.56
2.14
Willingness to compromise
3.08
4.81
4.91
3.33
5.12
3.60
3.03
Locus of control
5.59
5.24
4.91
6.10
5.26
5.61
5.58
Speed of making decision
3.73
3.24
2.25
5.42
3.02
2.47
4.38
Procrastination
4.21
5.10
2.66
5.78
2.58
3.53
4.97
Dependence
6.16
5.60
3.89
6.47
5.76
6.19
6.40
Desire to please others
5.88
4.72
4.26
5.69
5.79
5.94
5.99
62
Steps in Sequential Elimination
Locating and prioritizing aspects or factors
Explicating within-factor preferences in the most
important factor not yet considered
Eliminating incompatible alternatives
Too many promising alternatives?
yes
no
This is the recommended list of occupations
worth further, in-depth exploration
63