How to vet candidates on evidence: structured interviews (.42), job knowledge tests (.40) and work samples (.33) predict best. A 6-step vetting process.


Book a free discovery call. We map the role, set your budget, and start sourcing pre-vetted European talent.
Book a call โTo vet remote candidates well, use the methods that research shows predict job performance, in this order: a structured interview with the same questions and a scoring guide for everyone, a skills test built from the real work of the role, and a check of job knowledge. Drop the methods that predict little: unstructured "tell me about yourself" conversations and filtering on years of experience. The largest recent review of the evidence, by Sackett, Zhang, Berry and Lievens in the Journal of Applied Psychology (2022), found structured interviews had the highest average validity of any selection method (.42), while unstructured interviews scored .19 and years of experience .07.
Below are the numbers, what they mean in practice, a six-step vetting process built on them, the checks remote roles add, and how we run vetting at Euro Remote Talent.

For decades, the reference point was Schmidt and Hunter's 1998 review. In 2022, Paul Sackett and colleagues showed that it had over-corrected for a statistical problem (restriction of range), which inflated several estimates, and recalculated them. The table shows the revised operational validity of each method, with the 1998 figure for comparison, as reported by SIOP's summary of the study and Master International's comparison table.
| Selection method | Validity, Sackett et al. 2022 | Validity, Schmidt and Hunter 1998 |
|---|---|---|
| Structured interview | .42 | .51 |
| Job knowledge test | .40 | .48 |
| Empirically keyed biodata | .38 | .35 |
| Work sample test | .33 | .54 |
| Cognitive ability test | .31 | .51 |
| Integrity test | .31 | .41 |
| Assessment center | .29 | .37 |
| Interests (fit with the work) | .24 | .10 |
| Conscientiousness questionnaire | .19 | .31 |
| Unstructured interview | .19 | .38 |
| Years of job experience | .07 | .18 |
Source: Sackett, P. R., Zhang, C., Berry, C. M. and Lievens, F. (2022). Revisiting meta-analytic estimates of validity in personnel selection. Journal of Applied Psychology, 107, 2040 to 2068.
A validity is a correlation between a method's score and later job performance, from 0 (no relationship) to 1 (perfect). A .42 is a strong result by the standards of hiring research, but it is far from certainty: a structured interview ranks candidates better than chance and better than a chat, and still gets some calls wrong. That is why every serious process combines methods, and why the combination should be scored rather than discussed.
Two practical changes. First, cognitive ability tests dropped from the top (.51) to the middle (.31), so a general reasoning test is no longer the obvious first filter. Second, work sample tests dropped from .54 to .33 but remain useful, especially when they mirror the real job. Structured interviews and job knowledge tests now lead. Years of experience, already weak, fell to .07: two candidates with five and ten years in a role tell you almost nothing about which will perform better.
List the four or five things the person must do well in the first six months, with how you would know. This is the scoring guide for everything that follows. Without it, every interviewer scores a different job.
Replace "five years of experience" with the evidence that matters: the number they owned, the tools they used, the scale they worked at. The 20-minute screening call script has three questions that get to it fast.
A work sample or job knowledge test should look like the first week of the job: a short brief, the same materials the person would use, two to three hours at most, and paid. Score it against the criteria from step 1 before you read the CV again. A generic online test measures something; a role-specific task measures the job.
Same questions, same order, for every candidate. Each question tied to one criterion, each answer scored on a 1 to 4 scale with a written description of what a 1 and a 4 sound like. Score during the interview, not after. Two kinds of questions work well: past behavior ("Tell me about the last time you had to win back an account; what did you do, and what was the result?") and situations ("A customer in Germany escalates at 19:00 your time; walk me through the next hour.").
Use references to confirm what the candidate told you: the role, the scope, the number from step 2. The 2022 review could not re-estimate the validity of reference checks because the original studies lacked the data, per Master International's summary, so treat a glowing reference as confirmation, not as a score. The guide to checking references for a European hire covers what you may ask and how to tell the candidate.
Add the test and interview scores, discuss only the cases where interviewers disagree by two points or more, and decide. A debrief that starts with "what did everyone think?" undoes the structure you built.
For Hiring Done For You searches, the process follows the same evidence. After a 30-minute call, the client gets a written search plan within 48 hours, including the skills tests we will build for the role. We headhunt directly across Europe for two weeks; everyone who gets through takes the role-specific skills test, including how the candidate works with AI, then a structured interview on ability, communication and motivation. The client meets three candidates, each with an interview recording, the test result and a written note on where they are strong and where they are not, and runs the final round. We handle the references. The fee is a flat $6,500, with a shortlist of three in 21 days or the deposit back, and a 12-month replacement guarantee.
If you run the search yourself, the same steps apply: post the role for $299, then use the process above. The guide to hiring without a recruiter puts it on a timeline.

Combine a structured interview (same questions, scored on a guide) with a skills test built from the real work of the role. In the 2022 review by Sackett and colleagues, structured interviews had the highest average validity (.42), followed by job knowledge tests (.40) and work samples (.33).
Much less than they feel. Their validity in the 2022 review was .19, less than half that of structured interviews. Keep a short unscripted part for rapport if you like, but do not score it.
Barely. The 2022 estimate was .07. Ask for evidence of what the person did and achieved instead of counting years.
Add three checks to the usual process: a written deliverable in the skills test, the hours they can work in your time zone confirmed in writing, and a paid task that shows they deliver on time without supervision.
The evidence says: structured interviews, job knowledge tests and role-specific work samples predict performance; unstructured chats and years of experience barely do. Write the criteria first, test the real work, ask every candidate the same scored questions, use references to verify, and decide on the scores. For remote roles, add writing, hours and setup. If you want three candidates vetted this way in 21 days for a flat $6,500, book a call.

Julien founded Euro Remote Talent, which helps people in Europe find remote jobs and places European remote professionals with remote companies.
Euro Remote Talent LLC
30 N Gould St Ste N
Sheridan, WY 82801, USA