Interests and Attitudes

1
Interests and Attitudes

• Chapter 14

2
Major Names in Career Interest Testing

• Edward Strong, Jr. first to do work on
  voacational interests. 1st published in 1927
• G. Fredic Kuder -- First published in 1939
• John Holland has own test and interpretive
  scheme for vocational interests

3
Approaches to measuring vocational interests

• Origin of scales empirical, criterion-keying
  approach (Strong) vs. scale developed to
  correspond with broad areas of interest (Kuder)
• Interpretation of scores Absolute (Strong) vs.
  relative (Kuder)

4
Holland and RIASEC Codes

• Realistic mechanic, aircraft controller,
  surveyor, farmer, electrician
• Investigative biologist, chemist, physicist,
  anthropologist, geologist
• Artistic composer, musician, stage director,
  writer, interior decorator, actor
• Social teacher, religious worker, counselor,
  clinical psychologist, speech therapist
• Enterprising salesperson, manager, television
  producer, sports promoter, buyer
• Conventional bookkeeper, financial analyst
  banker, cost estimator, tax expert

5
The Holland Hexagon
Realistic
Investigative
Conventional
Artistic
Social
Enterprising
6
Strong Interest Inventory (SII)

• Most recent edition, Form T317, has 317 items and
  eight categories
• Three-option, multiple-choice format (like,
  indifferent, dislike)
• Take 35-40 minutes to complete

7
Types of Scores on the SII

• General Occupational Themes (GOTs) 6 scores
  RIASEC
• Basic Interest Scales (BISs) 25 scores based on
  factor analysis
• Occupational Scales (OSs) 211 scores
  criterion-keyed to occupational groups
• Personal Style Scales 4 scores
• Administrative Indexes 3 summaries

8
Norms for SII

• Norms based on 211 occupational groups (most of
  which are separate for men and women)
• To be in norms group had to volunteer, be at
  least 25 years old, employed in field for 3 or
  more years, perform typical work for that
  occupation and express satisfaction with job
• All norms for scales are standard scores with
  M50 and SD10. GOT, BIS, PS based on male and
  female combined. OS has separate scores by
  gender.

9
Reliability of the SII

• Extensive reliability data for all scales except
  Administrative Indexes.
• Median internal consistency in upper .80s
• Test-retest reliabilities in .80s over 1-6 month
  period
• Apparent that career interests are remarkably
  stable

10
Validity of SII

• First method is to show results differentiate
  between existing occupational groups
• Second method is to show scores are predictive of
  occupation or college major that people
  ultimately select (hit rates)
• Manual reports variety of studies using both
  methods

11
Kuder Occupational Interest Survey (KOIS)

• 100 forced-choice triads.
• For each triad, person marks the activity
  preferred most and preferred least.
• 30 minutes to complete
• Four sections with 6 types of scores

12
Types of Scores on the KOIS

• Dependability is validity statement
• Vocational Interest Estimates (VIEs) percentile
  scores in 10 broad interest areas
• Occupational Scales (OSs) comparisons to 76
  occupational groups with gender-specific norms
• College Major scales comparisons to 26 college
  majors, most with separate norms by gender
• Experimental scales only counselor sees
• RIASEC scores based on averaging percentiles of
  VIEs

13
Reliability and Validity of KOIS

• Internal consistency of VIEs range from .47-.85
  with median of .66
• Median stability estimate over 2 weeks was .80
  for VIEs, .90 for OSs.
• Validity based on hit rates (scores matched
  occupation) and factor analyses.

14
Self-Directed Search (SDS)

• Inventory is self-administered, self-scoring, and
  self-interpreted
• First appeared in 1971 last edition is 1994
• Originates with Hollands RIASEC scheme
• 228 items 30 minutes to complete
• four major parts w/in each part, items grouped
  by RIASEC
• Yields 6 scores, one for each RIASEC area
• RIASEC 3-letter code used with The Occupations
  Finder

15
SDS Norms and Scoring

• Summary score scales are raw scores translated
  into RIASEC codes
• Top three scores become their 3-letter RIASEC
  code to use in Occupations booklet
• Norms are not an important part of interpretive
  scheme, though they are provided

16
Reliability and Validity of SDS

• Internal consistency in low .90s
• More information needed on temporal stability
• Validity determined by
• independence of scales
• hit rates
• consequential validitydid SDS make a difference
  in peoples lives

17
Thorny issues of Career Assessment

• Relative vs. absolute scores
• Gender differences
• Interest-ability relationship
• Long term research when test keeps changing
• Huge number of scores
• Directiveness in helping client make decision

18
Attitude Measures

• Usually defined with respect to its object
• Has three components cognitive, emotional, and
  behavioral
• Usually differentiated by method of scale
  construction Likert, Thurstone, Guttman

19
Likert Scales Method of summated ratings

• Start with large number of items
• Ask people to respond on five-point scale
  (strongly disagree to strongly agree)
• Select items that correlate most highly with
  total score
• Recommended that directionality of items be
  switched to control for response sets
• Assumes one attitudinal dimension underlies the
  items, though factor analysis often conducted

20
Thurstone scale Method of equal appearing
intervals

• Write large number of items that cover all shades
  of opinion on topic
• Response format agree/disagree
• Judges sort items into 11 (equidistant)
  categories ranging from most to least favorable
• Find M and SD for category placement of each item
• Eliminate statements with large SDs
• Select the best statements to represent
  attitudinal positions along continuum from least
  to most favorable

21
Guttman Scales Scalogram Analysis

• Stringent technical requirements
• Get set of items ordered with complete internal
  consistency
• Locate persons change-point position within
  ordered set of items
• If scale meets Guttman criteria, then know how
  person will respond to every item
• Want 90 replicability extent that responses are
  consistent