Grammar and Vocabulary Tests Online: Do They Predict Workplace English?

What a grammar or vocabulary test online actually measures

A grammar or vocabulary test online is one of the easiest assessments a hiring or L&D team can run. It is quick to sit, quick to score, and the results feel objective: a candidate either knows that "since" takes a present perfect, or recognises what "reconcile an invoice" means, or they do not. Multiple-choice items can be auto-marked in seconds, and the numbers line up neatly on a spreadsheet.

Used well, these tests measure something real. Grammar and vocabulary sit at the core of receptive knowledge — the ability to recognise and understand language you read or hear. A strong grammar test online tells you whether someone can parse a complex sentence, follow a written policy, or make sense of a technical manual. Vocabulary breadth correlates reasonably well with reading comprehension and with the ability to follow fast, idiomatic speech. For roles where English is mostly consumed rather than produced, that signal has genuine value.

The problem starts when teams treat that signal as a proxy for everything a new hire will do in English — including the parts the test never looked at.

The validity gap: knowing the rules vs. using the language

There is a well-worn distinction in language assessment between receptive skills (reading, listening) and productive skills (speaking, writing). Grammar and vocabulary tests measure receptive knowledge directly and productive ability only by inference. That inference is where predictions go wrong.

Consider two candidates who score identically on a 40-item grammar and vocabulary screen:

• One writes clear, correctly punctuated client emails and speaks fluently on a call, self-correcting small slips as they go.
• The other freezes when asked to produce language in real time, mislabels the register, and takes three drafts to write an email a client will not misread.

The test cannot tell them apart, because it never asked either of them to produce anything. Recognising the correct option in a multiple-choice question uses a different cognitive process from generating accurate, appropriate language under time pressure, with a real audience, on a topic you did not choose. Someone can "know" the past conditional and still not deploy it correctly mid-sentence on a customer call.

This is the validity gap: the distance between "knows the rules" and "can hold a client call or write a clear email." For most office roles, the second is what you are actually hiring for.

What receptive-only tests miss in the workplace

Map a typical knowledge worker's day against what a grammar-and-vocabulary screen covers:

Two of the three highest-stakes communication tasks — writing and speaking — fall entirely outside the test. Yet emails, calls and meetings are where miscommunication actually costs money: a mishandled support call, an ambiguous instruction to a supplier, an email that reads as rude because the register is wrong. A screen that ignores productive skills is silent on precisely the risks that matter most.

We have written more about this trade-off in tests that skip speaking and writing, and the pattern is consistent: the cheaper, faster the receptive-only test, the more tempting it is to over-read its results.

Why "separate modules" quietly becomes "never tested"

Several established platforms take a modular approach. TrackTest, for example, offers CEFR-based online English testing with subscription tiers and a core focus on grammar, vocabulary, reading and listening, with speaking and writing offered as separate modules. On paper, that looks flexible — buy the productive skills when you need them. In practice, teams that want a full picture from a single sitting often look for a TrackTest alternative that scores all four skills in one flow rather than as paid add-ons.

In practice, separation creates friction, and friction becomes omission. When speaking and writing are extra steps with extra cost and extra scheduling, busy teams under a hiring deadline skip them. The result is that the two skills most predictive of workplace communication are the two most likely to go untested — not by decision, but by default. The structure of the product nudges you toward the weaker screen.

Making speaking and writing first-class parts of the same test, rather than optional bolt-ons, removes that temptation. If every candidate produces spoken and written language as a matter of course, you always have the signal — and every result is comparable, because everyone did the same thing.

Where receptive-only is genuinely enough

None of this means grammar and vocabulary tests are useless. The honest answer is that it depends on the role.

Receptive-only can be a fair screen when:

• The job mostly involves reading English — following documentation, processing written forms, reading dashboards.
• Spoken interaction is minimal or in the local language.
• The English output required is templated or reviewed before it goes out.
• You are sifting a very large top-of-funnel and need a fast first filter before a deeper stage.

You need productive skills tested when:

• The role is client- or customer-facing by phone, video or in person.
• The person writes unsupervised English that reaches customers, partners or the public.
• Register, tone and clarity carry commercial or reputational weight.
• You are making a final decision, not just an initial cut.

A sensible programme often uses both: a quick receptive filter early, then a full four-skill assessment before anyone reaches a shortlist or a job offer.

How to design a screen that predicts on-the-job English

If the goal is to predict real workplace performance, a few principles hold regardless of vendor:

1. Test the skills the job uses. Start from the actual tasks — calls, emails, reports — and work back to the skills. If the role produces language, the screen must ask candidates to produce language.
2. Include speaking and writing by default. Treat them as core, not as a paid extra you might skip. The moment they become optional, they become skipped.
3. Anchor results to a shared scale. A raw grammar percentage means little across roles and reviewers. A CEFR level (A1–C2) travels: everyone understands what B2 implies for an email or a meeting.
4. Keep scoring consistent and fast. Human panels are slow and drift over time. Automated and AI scoring applies the same standard to every candidate and returns a level in minutes, so hiring speed is not the excuse for cutting the productive skills.
5. Make results comparable. One flow, one report, one scale beats a patchwork of separate module scores that no two reviewers weight the same way.

The International English Test was built around exactly this. It assesses all four skills — listening, reading, speaking and writing — in a single flow, with speaking and writing AI-scored inside the test rather than sold as separate modules. Results map to CEFR A1–C2 and come back in minutes, and pricing is credit-based (roughly £8.99–£11.99 per test depending on volume) with no contracts. As an ALTE Associate Member, the test follows recognised standards for quality and fairness. If you want the detail behind the scoring, how we score walks through it.

The bottom line

A grammar or vocabulary test online is a useful instrument for what it measures: receptive knowledge, scored fast and cheaply. What it cannot do on its own is tell you whether a candidate will write a clean client email or hold a confident call — the productive skills that carry most workplace communication. The gap between recognising correct English and producing it is where receptive-only screens quietly overpromise.

The fix is not to abandon grammar and vocabulary testing, but to stop treating speaking and writing as optional. Test the skills the role actually uses, keep them in one comparable flow, and anchor everything to CEFR. That is how a screen stops merely measuring what someone knows and starts predicting what they can do.

English assessment tests for companies let you screen all four skills in one CEFR-aligned flow, so your results predict on-the-job English rather than just recognition of the rules.