Empirical researchers have long pointed out that more than five years of previous experience did not predict great job performance for new hires.[i] A new review of 81 studies looking at the correlation between prior job experience and performance in new companies showed there is no link between the two.
Many hiring managers and recruiters screen for years of experience on a job or in an industry because it is a “number,” appears objective, and is easy to attain. Years of experience, however, does not convey the job candidate’s quality of work, if the candidate’s skills are up to date, what was achieved, what obstacles were overcome, and level of success.
The recent study by Chad H. Van Iddekinge of Florida State University and his colleagues looked at 15 of the 23 job families listed by the US Labor Department’s Occupational Information Network. The study measured the performance of new hires by either supervisor evaluations or objective, quantifiable measures such as sales achievement or parts produced by machine operators.
What are better screening criteria?
It is better to focus on job knowledge, skills, and traits needed to be effective on the job, or even measures of experience at the task level. For example, for pilots or truck drivers instead of asking for how many years they worked in those jobs in initial screening questions, ask how many hours they have logged flying or driving. These metrics were better predictors of success.
To better ascertain job knowledge, skills, and traits, interviewers will do much better, according to the research, by using structured interviews that have identified the critical knowledge, skills, and tasks to be done and then ask the job candidates about these competencies as they review their resumes. The interviewer will want to look for answers from the job candidates that match the answers they would get from their current best employees. This takes a thorough review of the job and learning what makes your current top performers the best.
Companies can also use skill tests such as coding in a specific language for software engineers, manual dexterity for manual assemblers, or tests on the competence of the Microsoft office suite software for administrators.
Research also shows that work sample tests are a reliable screening method. With work sample tests, job candidates perform a hands-on simulation of some or part of the job, such as preparing and presenting a pitch to customers for a marketing employee, or hands-on assembly for factory workers.
Personality tests can be useful for selecting sales employees and leaders.
Whatever skill or personality test is used needs to be provided by a vendor who has validated their test to see that it predicts excellent job performance and does not discriminate in hiring. Don’t trust HR technology companies that claim to use “the new field of data science” unless they can show you they have validated their selection procedures and the data. In addition, the company using the test needs to try it out on a sample of its workforce to verify it can pick out the best workers. Using a test that has not been validated can leave the company open to lawsuits if the test has a bias.
Besides, companies need to track the results of their hiring decisions to see that the screening methods used resulted in selecting outstanding performers and did not adversely select out applicants based on gender or ethnic bias.
Finding and hiring new hires is tough enough in a tight talent economy. You want to be sure you are smartly screening your best candidates.
Victor Assad is the CEO of Victor Assad Strategic Human Resources Consulting. Today’s blog includes excerpts from his new book: Hack Recruiting: the Best of Empirical Research, Method and Process, and Digitization. You can buy it online at Amazon, Barnes and Noble, and Archway Publishing.
[i] Frank L. Schmidt and John E Hunter, “The Validity and Utility of Selection Methods in Personnel Psychology: Practical and Theoretical Implications of 85 Years of Research Findings,” Psychological Bulletin, 124, No. 2 (1988): 272. Copyright `998 by the American Psychological Association, Inc. 0033-2909/98/.