ATA Certification common mistakes
ATA Certification · Understand the traps and how to avoid them.
Leaving target-language sentences in source-language word order, producing calques such as placing the verb or the modifier where the source puts it rather than where the target language requires it.
Under time pressure candidates translate clause-by-clause and keep their eyes on the source sentence, so the syntactic frame of the source carries over into the draft and never gets restructured.
What to do instead
Draft the meaning, then reread each target sentence alone, with the source covered, and rebuild word order around target-language syntax; mark any sentence that still mirrors the source and rewrite it before moving on.
Producing a register that is too casual or too stiff for the stated purpose and audience, for example colloquial phrasing in a legal or official text or heavy bureaucratic phrasing in a public-facing notice.
Candidates focus on lexical equivalence and treat style as decoration, and they do not consciously identify the source register and the target audience before starting to translate.
What to do instead
Before translating, write one line stating the text type, purpose, audience, and required register, then check the finished draft against that line; adjust contractions, formality of address, and sentence length to match.
Producing target text that reads fluently but omits or blurs a qualifier, negation, scope word, or condition present in the source, so the translation is smooth but not equivalent.
Candidates paraphrase to keep the target readable and drop small words such as scope particles, hedges, or negatives because they seem redundant in the target language.
What to do instead
Do a dedicated fidelity pass comparing source and target segment by segment, ticking off every negation, quantifier, modal, and condition; treat any unticked item as an error even if the sentence reads well.
Misreading ambiguous source syntax, such as attaching a modifier to the wrong noun, misidentifying the subject of a passive or impersonal construction, or missing the scope of a conjunction.
Candidates translate the first plausible parse without testing alternatives, and source-language ambiguity is resolved by guesswork rather than by grammatical analysis or context.
What to do instead
When a source clause has more than one possible parse, list the readings, test each against surrounding context and real-world plausibility, and only then translate; if the ambiguity is genuine and material, translate conservatively and flag it.
Inventing a plausible-sounding term or using a false friend instead of verifying it, so a technical or institutional term is rendered with a word that does not exist in the field or means something else.
Candidates rely on bilingual dictionary headwords or on cognate resemblance and skip the step of confirming the term in target-language usage.
What to do instead
For any term outside everyday vocabulary, confirm the target equivalent in a monolingual target-language source or a domain glossary, record where it was verified, and use the same rendering consistently throughout.
Using a permitted resource in a way the instructions do not allow, such as relying on a machine translation engine or a pre-built translation memory when only dictionaries and glossaries are permitted.
Candidates do not read the resource rules carefully or assume that anything available on the computer counts as a dictionary.
What to do instead
Read the resource rules before the exam starts, list exactly which tools are permitted, close everything else, and practise a full timed passage using only that permitted set so the workflow is habitual.
Translating a term inconsistently, using two or three different renderings for the same source concept across the passage.
Candidates translate linearly and forget earlier choices, and they have no running record of decisions to consult.
What to do instead
Keep a scratch glossary of recurring terms and their chosen renderings as you go, and run a search for each key term in the finished target text to confirm one consistent rendering.
Ignoring an explicit instruction in the prompt, such as a required format, a specified audience, a mandated term, or a note about what to do with an untranslatable item.
Candidates skim the instructions once and then concentrate entirely on the text, treating the prompt as background rather than as part of the task.
What to do instead
Read the instructions twice and convert them into a short checklist of constraints; verify each constraint against the finished translation before submitting.
Failing to adapt the text to its stated medium and purpose, for example keeping print-style layout, headings, or abbreviations in a text meant for a web page or an oral presentation.
Candidates treat the source layout as sacred and do not consider how the target text will actually be used.
What to do instead
Identify the medium and purpose first, then decide deliberately what to preserve and what to adapt, such as expanding abbreviations, converting formatting, or restructuring headings for the target context.
Submitting a first draft without a revision pass, leaving typos, agreement errors, punctuation faults, and unfinished or doubled words in the target text.
Candidates run out of time because they translate slowly and linearly, so revision is the first thing sacrificed.
What to do instead
Budget revision time from the start, reserve a fixed final block for proofreading the target alone, and practise timed passages so drafting speed leaves room for a genuine revision pass.