Interview screening test or call: which goes first

Picture a support role with 2,000 applicants and two ways to start. Option one: send everyone a 30-minute online test and talk only to the people who pass. Option two: put everyone who clears the resume read into a short screening conversation and test only the people you like. Both get called an interview screening test, both are sold as the efficient choice, and they produce different shortlists from the same pool.

The order matters more than most hiring plans admit, because every step costs you two things: your time on each candidate, and the candidates who never finish the step. Below is what each step actually measures, what it costs, and a role-by-role rule for which one goes first.

What is an interview screening test?

An interview screening test is any first-round check that filters applicants before a full interview. It takes one of two forms: a short screening interview (a recruiter call or an AI voice interview that checks must-haves and how someone talks about their work), or a pre-employment test (an online assessment of reasoning, job knowledge, a work sample or typing).

Google’s own answer mixes the two, which is why recruiters searching for one keep finding the other. The difference matters. A test measures what a person can do with a task in front of them. A screening conversation measures what they can explain, what they really did in past jobs, and whether the job’s basics (shift, location, notice period, pay) fit. Our guide to the phone screen interview covers the conversational version in detail.

What each step actually measures

Here is the evidence, briefly. The most-cited comparison of hiring methods was revised in 2022 by Paul Sackett and colleagues in the Journal of Applied Psychology, who found that older estimates had been inflated by statistical corrections. In the revised ranking, as summarised by SIOP, structured interviews came out on top at .42, ahead of job knowledge tests at .40, work sample tests at .33 and cognitive ability tests at .31.

Two practical readings of those numbers:

The point is not that interviews beat tests. It is that a structured interview is not the soft option it is often treated as, and that the right test depends entirely on what the job is.

The cost nobody puts in the plan: drop-off

Every step you add loses some candidates who would have passed it. A long test is a big ask of someone with five other applications open, and the applicants with the most options are the ones least obliged to say yes. The effect can be large: in campus hiring, recorded video submissions lose around half of applicants on their own.

So the question for each step is not only “how well does it predict?” but “how many good candidates does it cost us before we learn anything?” A 45-minute test sent to 2,000 people may give you a clean ranking of the 900 who finished it, and tell you nothing about the 1,100 who did not.

Two rules keep drop-off down. Put the shortest step first, so candidates invest a little before they invest a lot. And tell them up front exactly how long each step takes, so nobody starts a test they cannot finish.

Test first or talk first: a rule by role

Put first the step that best samples the core of the job, at the lowest cost per candidate for both sides.

For customer-facing roles, telecallers especially, the conversation is the test; our telecaller interview questions include a three-minute mock call for exactly that reason.

How to combine them without doubling the drop-off

When a role needs both, do not send two separate links on two separate days. Each invitation is another chance for a candidate to drift away.

• One sitting, one link. Knock-outs, a few reasoning questions and the role questions can run back to back.
• Knock-outs first, always. Whoever cannot work the shift should find out in two minutes, not after a test.
• Do not test what the resume already shows. If the role needs a degree, the resume says whether they have one.
• Score each part separately. A strong conversation should not hide a failed accuracy test, or the other way round.

Unproctored tests have one more problem: they are easier to game than they used to be. We wrote about how assessment cheating is rising and why a live conversation keeps more of it visible, because a follow-up question is harder to outsource than a multiple-choice answer.

Where HireQwik fits in the order

HireQwik runs the conversation step, and for most volume roles we think that belongs near the front. Resumes are scored against the role first. Candidates who clear it book a 10 to 20 minute AI voice interview at a time that suits them. Knock-outs come first and end the call early when one fails; then the role questions, with follow-ups when an answer is vague. Recruiters choose the interview type per job: an HR screen for most volume roles, a Technical round, or an Aptitude round that screens reasoning for campus and other high-volume funnels.

Every answer is scored against the job’s rubric with the candidate’s words quoted, and switching away from the interview tab is flagged for the recruiter to look at, without changing the score. The recruiter makes the call.

If you need a timed coding or typing test, keep a dedicated tool for it, and put it after the conversation for experienced hires or before it for data-entry roles, following the table above. To hear what the conversation step sounds like from the candidate’s side, take the 3-minute live AI interview.