HR Trade Publication inclusion and quotes for client PreVisor
:
Predict Performance, Improve Efficiency With Assessments for Selection
by Paul Connolly, Talent Management Magazine, October 2010
There are good tests and bad tests out there and good and bad uses of them," Lahti said. "You can have a test that measures pretty well, [say] an entry-level customer service test, which may be effective for predicting performance and retention in those roles, but it might not be a good test for measuring leadership potential or predicting success in a director-level or manager-level role."
What if talent managers could hire or promote someone and predict whether or not the individual would work out?
Michael Blair said it's possible. Blair is strategic staffing leader for CenturyLink in Overland, Kan., which operates in 33 states and employs some 20,000 people. He said in his experience that those who do well on personality assessments tend to outperform those who don't.
...
...(page 2)Ken Lahti, vice president of strategy and content for PreVisor, said the easiest questions to fake are those that require a response ranging from "strongly agree" to "strongly disagree" or those that ask for a ranking on a scale from 1 to 5.
Tests that use forced choices, or those that involve the tester issuing a warning about faking, tend to show less faking. Lahti also said faking can be minimized through the use of computer-adaptive testing, which automatically refines questions based on responses.
The nature of good assessment instruments is to extract a set of scientifically objective metrics that filter out most efforts to game the system. Anyone who has taken a personality test will relate to the seemingly repetitious questions. That's part of the filtering process, but there is a good deal more built into assessments that can assure an organization receives accurate and reliable information.
Lahti said the key to good assessment instruments is the word "good." It's best to ask first what kind of research is backing them up.




