A recent Planet Money podcast focused on the increasing use of job-screening tests that help big businesses find call center employees and collections agents in ways that go far beyond screening resumes for keywords. The companies that create the tests work with their clients to identify their best (and longest-lasting) employees, isolate the qualities that predict such performance and retention, and then design tests that gauge which applicants have these qualities. Indeed, the podcast noted that the companies using these new online job assessments are finding that they do a better job of predicting employee performance and retention than they had achieved through traditional approaches like reviewing past job experience (through a resume) or asking about applicants’ strengths (through a cover letter or interview).
Why would a faceless computer program out-perform a friendly hiring manager? Planet Money pointed out the blatantly obvious: people are flawed. Angling for a job, employees may be inclined to write cover letters and answer screening tests in ways that are more about what they think the employer wants to hear than what's really true. Meanwhile, people like to work with people like them, so employers tend to look for cues in resumes and interviews that indicate whether candidates are like them – same or similar colleges, relevant volunteer experiences, and so forth – rather than focusing purely on skills or knowledge. And then there's past experience, which we often look to as a proxy for likely success. For Xerox, past experience in call centers was found to have little bearing on success or longevity among its call center employees. Elsewhere, Google has applied this “people analytics” to much of human resources work; the company backwards-mapped the characteristics of its most effective employees and found that its most successful employees don’t necessarily have college degrees, high test scores or strong grade-point averages. Google and other employers are also considering how to measure more intangible qualities like resilience, using not only formal application data and traditional job tests but also social data like LinkedIn profiles, Facebook profiles and Klout scores.
These new screening tests are designed to overcome these human problems, conducting a more objective screening of potential employees before bringing them in for an interview, and dig into applicants’ traits in unexpected ways: interestingly, the podcast revealed that collections agents’ success correlated with their creativity more so than with their persuasiveness. These tests also use clever approaches to uncover applicants’ true strengths and limitations. For example, rather than outright asking if potential call center employees are good at time management or are self-starters, “forced-choice” questions ask applicants to choose which of two statements best applies to them. This structure minimizes the chances that a candidate will simply pick the "right" answer, and maximizes the likelihood they'll just go with their honest gut feeling.
Could this approach work in education? First, we’d need to figure out which current teachers and administrators are effective – no small task. Then, we would isolate the qualities that distinguish the most successful and long-lasting employees from the rest of the pack, determine how to gauge those qualities upfront, and design valid and reliable tests that accurately measure those qualities. For example, the Haberman Educational Foundation claims its “Star Teacher Pre-Screener” questionnaire predicts 95% of the teachers who will stay and succeed (or fail/leave) based on measuring potential teachers’ beliefs and likely behaviors. Gallup’s TeacherInsight screening tool also promises to surface the best potential teachers, although it’s unclear whether it maps to improved student achievement in teachers’ classrooms. We might also bring in some of the tools and techniques being developed for the corporate world: imagine if employers or teacher training programs used games like those developed by Knack for corporations and medical schools to assess teacher candidates’ classroom management skills or social intelligence before awarding them a degree or credential.
This approach could teach us humans a thing or two – if we're willing to have our biases reprogrammed.