Employee Selection: References and Testing Study Notes

Introduction to Employee Selection and Selection Techniques

• Prevalence of Selection Methods: While interviews and résumés are the most commonly used methods to screen and select employees, industrial psychologists prefer other techniques discussed in this chapter for their superior predictive capabilities.

• Core Psychological Principle: In psychology, the common belief is that the best predictor of future performance is past performance. Organizations seek applicants with a history of success in similar roles.

• Verification Challenges: Verifying previous employment dates is simple, but verifying the quality of past performance is difficult. Unlike professional football teams that can use "game films" to observe college players, most employers must rely on references and recommendations.

Predicting Performance Using References and Letters of Recommendation

• Key Definitions:   

- Reference Check: The process of confirming the accuracy of information (such as employment dates and titles) provided by an applicant.    

- Reference: The expression of an opinion, either orally or via a written checklist, regarding an applicant’s ability, work habits, character, or potential. The person or organization asking for the reference determines the content and format.     

- Letter of Recommendation: A letter expressing an opinion regarding an applicant’s ability, work habits, and potential. The letter writer determines the content and format.

• Reasons for Use:     

- Confirming Résumé Details: Used to combat résumé fraud. Approximately $96\%$ of employers perform reference checks to verify truthfulness.         

- Examples of Résumé Fraud: George O’Leary (2001): Resigned as Notre Dame head football coach after lying about his academic and athletic credentials.          

- Checking for Discipline Problems: Aims to discover histories of poor attendance, sexual harassment, or violence to avoid charges of Negligent Hiring.         

     Negligent Hiring: If an organization fails to check references and an employee later commits a crime on the job, the employer may be held liable based on the duty to protect customers and staff.         

    Case Study (Virginia): A grocery store was found not guilty of negligent hiring when an employee raped a customer because the store had checked references and found no reason for concern.     

- Discovering New Information: References provide insights into character and skills. However, reference checkers should seek consensus from several sources to avoid being misled by one person’s subjective difficulty with an applicant.    

Problems with References and Letters of Recommendation

1. Predictive Validity: A meta-analysis found the average uncorrected validity coefficient to be only $0.18$, with a corrected validity of $0.29$.

2. Four Primary Problems:

        1. Leniency: Most letters are positive; fewer than $1\%$ of references rate applicants as below average. Applicants choose their own references, ensuring a biased pool.         

                - Confidentiality: Research by Ceci and Peters (1984) shows writers are less lenient when applicants waive their right to see the letter.         

                - Legal Ramifications: Writers fear defamation suits (slander for oral, libel for written). However, Conditional Privilege provides legal protection if the writer believes the statement is true and has reasonable grounds to believe it.     

        2. Knowledge of the Applicant: Professors or supervisors may not know the applicant well or may have only seen them in limited contexts. Only a fraction of an applicant's actual behavior is recalled by the writer.     

        3. Reliability: Reference reliability is only $0.22$. There is often more agreement between two letters written by the same person for two different applicants than between two different people for the same applicant.     

        4. Extraneous Factors: Longer letters are perceived more positively ($Loher et al., 1997$). Letters with specific examples are rated higher than those with generalities.

Ethical Guidelines for Providing References

• Relationship Disclosure: Explicitly state the relationship (boss, coworker, friend). This prevents bias, such as the example of an applicant being hired based on a glowing letter from a coworker who was secretly his daughter.

• Honesty in Details: Referees have ethical and legal obligations to provide relevant information. One must balance the duty to the seeker with fairness to the applicant.

• Applicant Review: Allow the applicant to see the reference before sending it, giving them the chance to decline its use.

Predicting Performance Using Applicant Training and Education

• Education Requirements: While common, their validity is inconsistent. Hunter and Hunter (1984) found educational validity to be only $0.10$.

• Police Performance Meta-analysis: Found education predicted academy performance ($r = 0.26, \rho = 0.38$) and job performance ($r = 0.17, \rho = 0.28$).

• Grade Point Average (GPA): Predicts job performance, training performance, and salary. It is most predictive in the first few years after graduation. High use of GPA can lead to high levels of adverse impact.

Predicting Performance Using Applicant Knowledge and Ability

• Job Knowledge Tests: Designed to measure mastery of specific information (e.g., how to make a drink or conduct a job analysis). These have excellent content and criterion validity but can result in adverse impact.

• Ability Tests: Use when applicants are not expected to know the job upon hire but will be trained.     

- Cognitive Ability: Includes dimensions like reasoning, mathematical facility, and memorization. Wonderlic Personnel Test: A popular 12-minute test for group settings. Siena Reasoning Test (SRT): Uses nonsense words to reduce racial score differences found in traditional tests.   

- Perceptual Ability: Includes vision, color discrimination, and hearing sensitivity (useful for machinists/tool makers).     

- Psychomotor Ability: Manual dexterity, reaction time, and arm-hand steadiness (useful for carpenters/truck drivers).     

- Physical Ability: Measured via job simulations or agility tests (e.g., push-ups, sit-ups).         - Nine Basic Physical Abilities: Dynamic strength, trunk strength, explosive strength, static strength, dynamic flexibility, extent flexibility, gross body equilibrium, gross body coordination, and stamina.         

        - Setting Passing Scores: Can be Relative (compared to others in a group; illegal under Civil Rights Act 1991 if based on protected classes) or Absolute (minimum needed for the job).

Predicting Performance Using Applicant Skill

• Work Samples: Applicants perform actual job tasks (e.g., fixing a fan belt). These have excellent content, criterion, and face validity and result in lower racial differences than written tests.

• Assessment Centers: Use multiple assessment methods and multiple trained assessors. Exercises include:     

    - In-basket Technique: Simulates desk paperwork requiring a response.     

    - Simulations: Role-playing job scenarios (e.g., a B-PAD for police).     

    - Leaderless Group Discussions: Groups solve problems without a designated leader; used to rate leadership and cooperation.     

    - Business Games: Demonstrate creativity and decision-making (e.g., Tinker Toy manufacturing game).

Predicting Performance Using Prior Experience and Biodata

• Experience Ratings: Past behavior predicts future behavior. Meta-analysis shows a significant relationship ($r = 0.27$).

• Biodata: A selection method considering life, school, and work history.     

• Development: Uses the File Approach (archival records) or Questionnaire Approach.     - Scoring: Uses the Vertical Percentage Method to weight responses based on their ability to differentiate between high and low performance groups (criterion groups).     - Standards for Items (Gandy and Dye): Items must be under a person’s control, job-related, verifiable, and non-invasive of privacy.     - Faking Biodata: Detected using bogus items (e.g., "Conducting an isometric analysis"). Bright applicants fake less often but do it better when they do.

Personality, Interest, and Character Testing

• The Big Five Model: Openness to experience, Conscientiousness, Extraversion, Agreeableness, and Emotional Stability.

• Validity: Conscientiousness is generally the best predictor across most occupations.

• Tests of Psychopathology: Measure abnormal personality (e.g., MMPI-2, Rorschach Ink Blot). These are considered medical exams under the ADA.

• Integrity (Honesty) Tests:    

        - Overt Integrity Tests: Ask directly about attitudes toward theft and previous history.         - Personality-based Integrity Tests: Measure general traits related to counterproductive behavior.       

        - Polygraph: General use made illegal for most employers by the Employee Polygraph Protection Act of 1988.

• Conditional Reasoning Tests: Measure aggressive biases (hostile attribution, potency, retribution, victimization, derogation of target, social discounting) that lead to counterproductive behavior.

• Graphology: Handwriting analysis. Research shows it is not a useful technique and is no more accurate than untrained judgments.

Medical and Psychological Testing Limitations

• Drug Testing: $8.2\%$ of employees admit to past-month drug use.     - Costs: Substance-abusing employees cost employers $\$7,000$ annually.     - Process: Initial screening ($EMIT/RIA$) followed by confirmation test ($GC/MS$).     - Detection Times: Marijuana can be detected for causal users for $8\,days$ and frequent users for up to $60\,days$.

• Psychological and Medical Exams: Can only be administered after a conditional offer of employment is made.

Comparison of Selection Techniques

• Top Validity (Performance):     - Structured Interview ($0.57$)     - Cognitive Ability ($0.51$)     - Job Knowledge ($0.48$)     - Work Samples (Verbal) ($0.48$)

• Lowest Validity (Performance):     - Unstructured Interview ($0.20$)     - Education ($0.10$)     - Interests ($0.10$)

• Highest Adverse Impact (White-Black):     - Cognitive Ability ($d = 1.10$)     - GPA ($d = 0.78$)

• Lowest Adverse Impact:     - Integrity tests ($d = 0.07$)     - Personality inventories (ranges from $d = -0.07$ to $0.16$)

Rejecting Applicants

• Organizational Impact: Proper rejection maintains customer loyalty and future application potential.

• Guidelines for Rejection Letters:     

    - Send them (though most organizations do not).     

    - Delay the mailing slightly (instant rejection feels impersonal/insulting).     

    - Be specific and personable.     

    - do not include a contact person (decreases probability of reapplying).

Ques

tions & Discussion

• Applied Case Study (City of New London Police): Robert J. Jordan was rejected for scoring a $33$ on the Wonderlic (max was $27$) because he was deemed "too bright." The city argued high intelligence leads to boredom and turnover. Jordan lost his lawsuit; the judge ruled the disqualification might be unwise but was not unconstitutional.

• Discussion Questions:     - Should organizations provide reference information? (Referees must balance the fear of slander suits with the risk of negligent reference liability).     - Are integrity tests fair? (They predict theft and performance but have higher failure rates for males and younger people).     - Ethics of Personality Tests: Critics argue they violate privacy and may be impacted by temporary moods; supporters view them as assessing organizational fit.