Chapter 14: Occupational Setting Flashcards
Purpose of Testing
- To determine potential for success in a program.
- To place individuals into programs.
This involves matching the candidates’ abilities and competencies with the requirements of specific training programs.
3.To match applicants with specific job openings. - To counsel individuals, for career advancement, or career changes, for example.
- To provide information for program planning and evaluation.
Preemployment Testing
- To elicit a candidate’s desirable and undesirable “traits.”
- To identify those characteristics of the candidate that most closely match the requirements of the job.
Test can be given before interview to screen out people
Test can be given after interview to confirm findings
Government regulations
- Before 1960s, employment discrimination was not illegal
- Civil Rights Act of 1964 was designed to provide equality
- In 1978, EEOC published the “Uniform Guidelines on Employee
Selection Procedures” - In 1987 (revised 2018), SIOP published the Principles for the
Validation and Use of Personnel Selection Procedures
Predictors
- Cognitive Ability
- Psychomotor tests
- Personality tests
- Integrity tests
- Work samples
- Assessment centers
- Biographical Information
- Interviews
The Criterion Problem
- Difficulty or complexity in measurement of performance criteria
- Performance: Criterion in which most organizations, employees,
managers, and I/O psychologists are interested
Nathan and Alexander (1988) conducted metaanalyses of validity coefficients from tests of clerical abilities for five criteria: supervisor ratings, supervisor rankings, work samples, production quantity, and production quality.
How Job Success is Measured
- Quantity and/or quality of production
- Personnel records
- Administration actions
- Performance ratings
- Job samples
Criteria are Dynamic
Ghiselli and Haire (1960) suggested that the criteria against which tests are validated are dynamic, that is, they change over time.
Criteria can change over time
The example used in the book with the taxi drivers may have had
more to do with amount of learning on the job.
You may need a criteria for when first learning a job versus a criteria
for performance after a person has been on the job for a certain
amount of time.
Level of performance desired should be determined BEFORE the
validation study
Ratings
Another meta-analysis was conducted by Conway & Huffcutt (1997) on multi-source ratings.
The found results similar to Harris & Schaubroeck (1988).
* Subordinates had the lowest mean reliability (.30) followed by peers (.37), and supervisors were the highest (.50).
* Comparing the correlations between the ratings for the different
groups:
o Subordinate ratings
.22 with supervisor and with peers
.14 with self-ratings
o Self-ratings
.22 with supervisor
.19 with peers
o Supervisor- peer correlation was .34
Rating Errors
Cognitive processing – supervisors need to recall information about employee performance. There can be errors anywhere in that process.
* Observe behavior
o Miss important behaviors or see what wants to see
* Encode information about behavior
o Label information incorrectly or not well enough
* Store information
o Don’t store information or store the wrong information
* Retrieve information
o Not able to retrieve or retrieve irrelevant information
* Integrate information
o Make a poor decision or make a biased decision
For all types of ratings errors, there can be a true component of
the rating and an invalid component
Forced distribution methods for ratings (forced distribution,
alternation-ranking, and paired comparisons) are not popular with
raters and ratees.
Ratings using behavioral anchors have more face validity with
ratees.
They also tend to give better ratings.
Rating Errors: Halo
Two types:
1. Use a global evaluation to assess the performance of an employee
2. Unwilling to distinguish levels of performance on independent
dimensions
Rating Errors: Leniency
Mean ratings is higher than mean ratings by other raters
Mean ratings are higher than the midpoint of the scale
Rating Errors: Central Tendency
Only use the midpoint of the rating scale
Rating Errors: Severity
Use only the low end of the ratings scale
Give ratings that are lower than other raters
Standardized tests
- What do you think about the used of standardized tests in employment contexts?
- Evidence suggests that general cognitive ability Accounts for a large proportion of variance in criterion
performance - Validity coefficient = .532 (or r2 = .25)
- Predicts performance similarly across countries
- Example: Wonderlic Personnel Test
Types of standardized tests
Specific Cognitive Ability Tests
Predict the likelihood that an individual will do well in a
particular job given his or her specific abilities; validity
coefficients range from .40 to .50
* Clerical tests
* Minnesota Clerical Test
* Number and name comparison
* Mechanical comprehension
* Bennett Mechanical Comprehension Test
