Almost every job opening asks applicants to provide their CV and/or fill in an application form that largely contains information that would be included in a CV, such as education and past experience. This is typically the first step in the process. Due to the large volume of applicants, mechanisms are used to screen applicants and reduce the size of the applicant pool so that resources can be invested into applicants that demonstrate more promise. While this makes intuitive sense, evidence shows us that CVs are bad predictors of future job performance. CVs provide a snapshot of past experience, but leave out talent, soft skills, potential or role and culture fit. Screening out based on CV risks screening out talent prematurely.
CVs are a poor predictor of job performance
61% of organisations use CVs as the first step in the hiring process, but did you know that CVs are among the worst predictors of performance?
First, human recruiters only spend around 7 seconds looking at each CV, focusing on current title, time spent in each role, and education, as they try to keep pace and sift through all of the applications they have received for a given role. Only spending mere seconds examining each application considerably limits their value and the information that can be inferred about a candidate’s abilities from an already limited snapshot.
Second, CVs essentially provide details about a candidate’s years of experience and the duties they have completed during their career. However, as can be seen in the graph below, job experience is among the weakest predictors of performance compared to interviews, aptitude tests and soft skills.
Third, humans are biased. We use cognitive shortcuts to help us make decisions and these shortcuts can be harmful; research indicates that a candidate’s name can impact their chances of being hired. There is also limited evidence that unconscious bias training is effective, and even if details such as name are removed from CVs, race/ethnicity and socioeconomic status could still be inferred from other details that are not redacted. For example, the place of education, grades, and where a candidate lives might all be used to make inferences about a candidate that could unjustifiably influence their success. In other words, CVs can contain a considerable amount of personal information that could intentionally or unintentionally influence perceptions about a candidate and decisions made about their hirability. This could see you missing out on talent.
Finally, some organisations have made the move to automated CV screeners to reduce the influence of human biases on decisions and help to increase the rate at which CVs can be screened. These tools can bring problems of their own. As an automated tool, the screeners cannot make inferences or identify transferrable skills in the same way as a human recruiter can. This can lead to quality applicants incorrectly being rejected. Likewise, since matches between a candidate’s experience and the job requirements are typically made on keyword matches, the context the keywords are used in could lead to some false positives, leading to low quality applicants being moved to the next step in the funnel.
Overall, CVs only contain very limited information about candidates that often do not portray a true picture of what an individual is capable of or how they would perform in role. Often, there is too much reliance on the insights that can be derived from CVs, which can impact the quality of your talent pool. CVs are a great starting point to guide the conversation during interviews, but using them as a mechanism to filter the applicant pool right at the start of the recruitment funnel is likely harming your candidate search – you need to supplement it with something else.
Work sample tests measure performance instead of predicting it
Unlike CVs and other selection methods that aim to predict what a candidate could be capable of if they were to get the job, work sample tests measure what a candidate can do. Also known as job try-outs, work sample tests provide candidates with a task that they would realistically encounter in the job if they were to get it and evaluate their performance. Of course, a candidate’s performance won’t be perfect on their first try, but job try-outs provide a real picture of what a candidate can already do, meaning it is much easier to see their potential with some onboarding and training.
Work sample tests are particularly useful for entry-level positions and graduate programs, where candidates typically do not have much (if any) job experience, making their CVs even less useful. They can also be well-suited to emerging fields that rely on transferrable skills that candidates might naturally have or have developed during their previous employment, as well as jobs that rely on soft skills such as hospitality roles. They can even be useful for leadership roles to see what executives really can do with the resources that would be available to them.
Work sample tests are more resistant to generative AI and faking
CVs, cover letters, and interviews are particularly vulnerable to candidates cheating using generative AI. This can lead to candidates incorrectly being accused of using generative AI, potentially ruining the chances of landing top talent, or could lead to underqualified candidates that have a gleaming but inaccurate CV advancing through the funnel.
Luckily, work sample tests are naturally less susceptible to cheating using generative AI as candidates are given a task to complete within a limited period of time, so cannot go away and come back with something AI generated later. If completed online, like with SeeTalent’s AI-powered work sample tests, additional steps could also be taken such as blocking the ability to copy text presented during the try-out and paste text in as a response. There are also a number of other ways to detect cheating such as whether a candidate leaves the tab and the time taken to start typing. The interactive nature of the task means that even if a candidate is cheating using genAI, they will need to edit and select responses and engage with the work sample.
Make the switch to SeeTalent’s AI-powered work sample tests
SeeTalent’s AI-powered work sample tests present candidates with a realistic and job-relevant scenario where our generative AI poses as a client, customer, supplier, partner, or another role that makes sense for that position to test how they perform. Our science-backed algorithms developed by subject matter experts then fairly and accurately evaluate performance. What’s more, our job try-outs can measure performance, personality, and values in a single step, reducing the number of assessments you need and consequently time to hire.
To find out more, schedule a free demo here.
Author: Airlie Hilliard
