Transparency
How this works
An instrument that will not explain itself should not be trusted. Here is exactly what Zymmax measures, how it arrives at a placement, and what it deliberately refuses to claim.
Three sources, kept apart
Most language assessments collapse into a single number. Zymmax keeps three kinds of evidence separate, because the interesting findings live in the differences between them.
What you believe comes from the learner profile questionnaire — confidence, anxiety, motivation, habits, preferences and interests. What you can do comes from a keyed diagnostic where answers are simply right or wrong. What you produce comes from a tutor's observation rubric, which reaches things no multiple-choice item can: hesitation, circumlocution, willingness to speak.
Placement is staged, not totalled
A single percentage is misleading. A learner can pick up scattered advanced items through cognates or luck, and a raw total would quietly promote them.
Instead every level block is scored independently against its own mastery threshold, and the placement is the highest level at which every band beneath it also holds. The thresholds fall as the levels rise, because higher-level items carry more linguistic load. Anything mastered above a gap is reported as evidence of recognition, not as attainment.
The comparison that matters
Each self-perception construct is paired with the ability construct it claims to describe, and the gap between them is reported in words rather than as a signed number.
A learner with strong grammar and low speaking confidence does not have a grammar problem; more grammar teaching will make them worse, because it confirms their belief that they are not ready yet. A learner who is fluent but inaccurate has been rewarded by being understood and needs corrective feedback rather than new material. These are opposite interventions, and no single score distinguishes them.
Errors are clustered, not counted
Every keyed item is tagged with the teachable point behind it. Six wrong answers spread across the past tenses, the pluperfect and the counterfactual conditionals are not six problems — they are one, and it has a name and a teaching sequence.
Nothing here is generated by a language model
Every sentence in a profile is derived from the learner's own answers by rules that are inspectable and stable. Two people with the same answers get the same reading; the same person re-reading their profile finds it unchanged. Phrase variation is chosen by a stable hash of the attempt, so reports are not word-for-word identical between learners, but they are never improvised.
What this is not
It is not a certification, and it does not entitle anyone to claim a CEFR level. It is a teaching instrument: a considered starting point, a plan, and a baseline to measure against in eight to twelve weeks. The second sitting is where the real information appears — what moved, and what stubbornly did not.
Every construct measured
| Construct | Measured by | What it tells you |
|---|---|---|
| Grammar | Demonstrated ability (objective, right/wrong) | Control of the structures of the language, measured on keyed items. |
| Vocabulary | Demonstrated ability (objective, right/wrong) | Breadth and precision of word knowledge, measured on keyed items. |
| Reading comprehension | Demonstrated ability (objective, right/wrong) | From retrieving explicit facts to inference and critical interpretation. |
| Communicative competence | Demonstrated ability (objective, right/wrong) | Choosing language that is not merely correct but socially appropriate. |
| Composure under pressure | Affect (anxiety, enjoyment) | The inverse of language anxiety: how settled the learner is when using the language. |
| Fluency resilience | Affect (anxiety, enjoyment) | Whether the learner keeps going when something goes wrong mid-sentence. |
| Accuracy orientation | Study behaviour and habits | How much attention the learner gives to being correct while producing language. |
| Study habits | Study behaviour and habits | How regular and self-directed independent practice is. |
| Review discipline | Study behaviour and habits | Whether the learner revisits material and corrections after first meeting them. |
| Exposure and background | Background and exposure | How much contact with the language the learner has had. |
| Interests | Interests (tallied, not scored) | Subjects the learner cares about, used to choose practice material. |
| Content preferences | Interests (tallied, not scored) | The kinds of material the learner would actually consume. |
| Motivation | Motivation and commitment | Commitment to continuing, and how much the learner wants proficiency. |
| Speaking confidence | Self-perceived ability (confidence) | How capable the learner believes they are when speaking. |
| Listening ease | Self-perceived ability (confidence) | How manageable the learner finds spoken input. |
| Reading confidence | Self-perceived ability (confidence) | How manageable the learner finds written text. |
| Writing confidence | Self-perceived ability (confidence) | How capable the learner believes they are in writing. |
| Grammar confidence | Self-perceived ability (confidence) | How secure the learner feels about structure. |
| Vocabulary confidence | Self-perceived ability (confidence) | How well the learner believes they can find the words they need. |
| Pronunciation confidence | Self-perceived ability (confidence) | How intelligible and comfortable the learner feels when speaking aloud. |
| What drives you | Learning preference (tallied, not scored) | What the learner says pulls them forward. |
| What drains you | Learning preference (tallied, not scored) | What the learner says causes them to lose motivation. |
| How you want to be corrected | Learning preference (tallied, not scored) | Feedback style the learner responds to. |
| How you want to learn | Learning preference (tallied, not scored) | Activity types the learner finds most useful. |
| How you learn words | Learning preference (tallied, not scored) | The learner's current vocabulary strategy. |
| Your priorities | Learning preference (tallied, not scored) | Skills the learner most wants to improve. |