Developing Selection Systems

1
Developing Selection Systems
2
The Selection Process

• Conduct job analysis
• Develop selection tests
• Validate test
• Recruit employees
• Screen applicants
• Test applicants who have passed screening
• Select qualified applicants
• Hire and/or reject applicants

3
Good Selection Systems

• Are valid
• Based on job analysis
• Predict work-related behavior
• Reduce the chance of a legal challenge
• Face valid
• Dont invade privacy
• Dont intentionally discriminate
• Minimize adverse impact
• Are cost effective
• Weigh costs associated with test
  purchase/creation, administration, scoring
  against benefits associated with improving the
  quality of the workforce

4
Common Selection Tools

• The Interview
• Unstructured Structured
• Intelligence and Ability Tests
• Personality/Interest Inventories
• Work Samples/Situational Exercises
• Biodata
• Letters of Recommendation
• Handwriting Analysis
• Polygraph Tests

5
The Interview

• Most popular method of selection
• Very subjective, however
• Dynamic process interviewee ? interviewer
• Chances for discrimination higher than for PP
  tests
• General format
• gather relevant data (e.g. resume)
• weed out unqualified, then interview rest

6
Unstructured Interviews

• Unreliable
• .15-.80 inter-interviewer reliability
• .80-.90 intra-interviewer reliability
• Not valid
• .00 - .20
• Legally problematic
• Very subjective
• Not job-related
• Rely on intuition and amateur psychology
• Suffer from many common rating problems

7
10 Factors that Influence Interview Outcomes (or,
How to get a job)

• Leniency-Severity Effect
• Halo Errors
• Central Tendency Errors
• Negative - Positive Nature of the Information
• Temporal Placement of Information
• Primacy/Recency effects
• Contrast Effects
• Who is being interviewed before you?
• Similarity of Gender with Gender-Orientation of
  Job
• Physical Attractiveness
• Similar-to-Me Effect
• Experience of Interviewer
• Interviewers Drink Together

8
Unstructured Interviews (cont.)

• If unstructured interviews have such low
  validity, why do we still do them?
• Few companies will hire someone without seeing
  them in the flesh
• Illusion of validity
• Selling job

9
Structured Interviews

• Goals
• Understanding the applicant
• Clarify information from resume or application
• Confirm resume information
• Obtain new information
• Predicting job performance
• Ask questions focused on past behavior
• Ask questions focused on skills
• Ask situational questions focused on future
  behavior
• Predicting organizational fit
• Use several interviewers
• Combine interview scores with test scores
• Selling the organization
• Realistic job preview
• Answer applicant questions

• Characteristics
• Reliable
• Valid (r .57)
• Not as prone to legal challenges
• Face valid
• No illegal questions
• Based on job analysis
• Ask same questions of each applicant
• Have standardized scoring procedure

10
Creating the Structured Interview

• Start with a thorough job analysis
• Tasks performed KSAs needed
• Conditions under which tasks are performed
• Get critical incidents of poor and excellent
  performance
• Determine best way to measure an applicants
  ability to perform each task
• Interview questions
• past focus
• skill focus
• future focus
• Psychological tests
• Personality
• cognitive ability
• integrity, etc.
• References or background checks

11
Creating the Structured Interview

• Construct questions
• Clarifiers
• Skill level determiners
• Past focus
• Future focus
• Determine rating anchors for each questions
• Typical answer approach
• Key issues approach

12
Conducting the Interview

• Build rapport
• Set the agenda
• Ask job-related questions
• Score each answer after it has been given
• Provide information about the job and
  organization
• Salary benefits
• Duties, opportunities for advancement, etc.
• Organizational history, climate, and future
• Answer questions
• End interview on pleasant note
• Compliment the interviewee
• Tell the interviewee when you will have a decision

13
General IntelligenceTests

• Based on belief that intelligence must correlate
  with job performance
• 3-D space, vocabulary, arithmetic reasoning
• 38142129?
• Shovel ditch digger as scalpel _______
• Knife, sharp, butcher, surgeon
• median r of about .53
• Problems
• is there such a thing as g?
• can one be too bright for a job?
• does g predict typical or maximum performance?
• Mean differences across racial groups on
  cognitive ability tests (not biased, but affects
  selection)

14
Personality/Interest Inventories

• Traditionally, research shows interest
  inventories are good for predicting membership in
  an occupation, but not success in the job!
• Classic Personality Test was MMPI
• but, MMPI developed for use with clinical
  population, not general population
• asks very personal questions invasion of privacy
  issues

15
Rebirth of Personality as a Predictor

• The Big Five Personality Dimensions
• Conscientiousness (reliable, dependable)
• r .23
• Extroversion (outgoing, friendly)
• r .10 - .26
• Emotional Stability (anxious, tense)
• r .07
• Agreeableness (team player, works well with
  others)
• r .06
• Openness to Experience (inquisitive)
• r 0 (performance)
• r .25 (training proficiency)

16
Work Samples

• Procedure
• Job Analysis
• Design sample to include critical tasks
• Rate performance of task according to
  appropriateness of process and outcome
• i.e., shortcuts might risk safety
• Validity
• excellent .40 - .60 (also high face validity)
• Limitations
• best for blue-collar workers only
• assesses maximum performance, not typical
• very expensive and time consuming

17
Biodata

• Consistently valid predictor
• .28 - .36
• predicts turnover, salary earnings, productivity,
  absenteeism
• Empirical Development (good for prediction)
• choose criterion
• sample current employees
• divide into highly productive and not productive
• ask many biographical questions
• select those questions that discriminate between
  the two groups of employees

18
Letters of Recommendation

• 0.0 predictive validity!
• Why?
• Poor reliability (references dont agree)
• Leniency
• Limited opportunity to view behavior

19
Trait Method to Evaluate Letters

• Underline traits in each letter place into
  following categories
• mental agility
• vigor
• urbanity
• cooperation-consideration
• dependability
• total the of traits per category
• Mental agility correlated with future GPA
• Urbanity correlated with future teaching evals

20
Handwriting Analysis

• 0.0 correlation between handwriting and
  success in the United States
• Mysteriously, moderate validities in European
  countries

21
Polygraphs/Lie detection

• Predictive validity is 0.0
• As of 1988, illegal to use in private sector
  preemployment screening
• but, used in federal govt and law enforcement
• can only distinguish fear and rage, not
  guilt
• many physical responses can be intentionally
  manipulated

22
Summary of Selection Techniques

• Best (r .4)
• Work samples
• Structured interviews
• Cognitive ability tests
• Job knowledge tests
• Moderate (r .2-.4)
• Biodata
• Assessment centers
• Personality tests
• Experience
• Integrity tests
• College Grades

• Poor (r lt .2)
• Vocational interest tests
• Application blank
• Handwriting analysis
• Projective personality tests
• Unstructured interviews
• Letters of reference