SLE Written Expression Test: Grammar, Format and Examples
Table of Contents11 sections
The SLE written expression test does not ask you to write a single sentence of your own. It is a multiple-choice test of French grammar, vocabulary and usage in workplace texts, and your score gives you the written expression letter of your linguistic profile.
Below you will find both formats, score bands, grammar by level, common anglicisms and five worked examples.
Format: 65 questions in 90 minutes supervised (55 scored), or 30 fill-in questions in 45 minutes unsupervised.
Level B: 31 to 42 out of 55 supervised, 17 to 23 out of 30 unsupervised. An E is only possible on the supervised test.
What the SLE writing test measures
The Public Service Commission of Canada (PSC) runs the Second Language Evaluation (SLE) for bilingual federal positions. Written expression is one of three separate tests, with reading comprehension and the oral language assessment, and it supplies the middle letter of a profile such as CBC.
Despite the name, there is no essay and nothing to compose. Every question is multiple choice: the test checks whether you can recognize correct, appropriate written French. In the Treasury Board's level descriptions, B means short factual texts on concrete topics with some errors that do not generally hinder understanding, and C means detailed, well-structured texts on complex or abstract topics.
SLE written test format: supervised vs unsupervised
You cannot book the test yourself. An HR team invites you after you apply to a bilingual position, or your manager arranges it, and you take it in the PSC's Candidate Assessment Tool (CAT), either supervised or as an unsupervised internet test (UIT).
Feature | Supervised | Unsupervised (UIT) |
|---|---|---|
Questions | 65 (55 scored, 10 pilot) | 30 |
Time | 90 minutes | 45 minutes |
Item types | Fill in the blank and error identification | Single-blank fill-in only |
Highest result | E (exemption) | C |
The pilot items are not identified, so treat every question as if it counts. Supervised results appear in CAT 48 hours later, online tests cannot be rescored, and you must wait 30 days before retesting.
How the two item types work
Fill in the blank: you choose the best word or group of words out of four options. The key usually turns on a small detail: a gender, a number, the preposition a verb requires or a mood triggered earlier in the sentence.
Error identification: three sections of a sentence are in bold, numbered (1) to (3), and you pick the one that contains one or more errors. Option (4), Aucun des choix offerts, means none of them does. Do not treat (4) as a trap: sometimes the sentence really is correct, and the item checks whether you can accept a structure that only looks suspicious.
Score bands: what you need for A, B and C
Result | Supervised (out of 55) | Unsupervised (out of 30) |
|---|---|---|
X | 0 to 19 | 0 to 10 |
A | 20 to 30 | 11 to 16 |
B | 31 to 42 | 17 to 23 |
C | 43 to 51 | 24 to 30 |
E | 52 to 55 | Not available |
Level B on the supervised test needs 31 correct answers, about 56%. Level C starts at 43 (about 78%), and the E band leaves room for only three wrong answers. An E exempts you from retesting indefinitely.
SLE French grammar by level
Levels are cumulative: a C item still assumes everything at A and B. Each row pairs a correct form with an incorrect one.
Level | What it tests | Example contrast |
|---|---|---|
A | Articles and gender | la réunion, not le réunion |
A | Basic agreement | les dossiers sont prêts, not les dossiers est prêt |
A | Passé composé auxiliary | elle est arrivée, not elle a arrivé |
A | Prepositions with places, dates | au Québec, le 3 mars, not en Québec, sur le 3 mars |
B | Tense and mood in correspondence | Si vous avez des questions, not Si vous auriez |
B | Pronouns y, en, lui, leur | je lui ai répondu, not je l'ai répondu |
B | Relatives qui, que, dont | le dossier dont je parle, not que je parle |
B | Homophones a/à, ou/où, ces/ses | le bureau où je travaille, not ou je travaille |
C | Subjunctive after bien que, pour que, il faut que | bien qu'il soit tard, not bien qu'il est tard |
C | Concordance des temps | si nous avions su, not si nous aurions su |
C | Participle agreement with avoir | les notes que j'ai rédigées, not rédigé |
C | Precise vocabulary, register | mettre en œuvre un plan, not implémenter un plan |
C | Connectors in formal paragraphs | par conséquent for a result, en revanche for a contrast |
C | tout, tous, quelque | tous les dossiers, not tout les dossiers; quelque 200 (about) |
Anglicisms and calques common in federal French
English constantly leaks into French memos in federal offices. These calques sound normal because you hear them every day, which makes them natural distractors and natural errors, so learn the standard form.
Calque to avoid | Standard French |
|---|---|
prendre place | avoir lieu |
appliquer pour un poste | postuler / poser sa candidature |
être en charge de | être responsable de |
adresser un problème | régler / traiter un problème |
faire du sens | avoir du sens |
à date | jusqu'à présent |
céduler | planifier / fixer |
rencontrer les exigences | satisfaire aux exigences / répondre aux exigences |
sur une base régulière | régulièrement |
The Language Portal of Canada, run by the Translation Bureau, has free writing tools and quizzes covering many of these.
Five worked examples
These original items follow the test format. Try each before reading the answer.
Sample 1: fill in the blank (level B, pronoun)
Texte : Au sujet des formulaires de remboursement, nous ___ avons reçu trois ce matin et les autres arriveront demain.
- (1) les
- (2) en
- (3) y
- (4) leur
Answer: (2). A quantity (trois formulaires) calls for en. Les cannot combine with trois, y replaces a place or à + thing, and leur is an indirect object for people. Reçu stays invariable after en.
Sample 2: fill in the blank (level C, subjunctive)
Texte : Bien que le comité ___ déjà approuvé le budget, la directrice souhaite revoir certaines dépenses avant la fin du trimestre.
- (1) a
- (2) ait
- (3) avait
- (4) aura
Answer: (2). Bien que always takes the subjunctive, here the past subjunctive ait approuvé. The other options are indicative (passé composé, plus-que-parfait, futur antérieur). The same rule applies after pour que and à moins que.
Sample 3: error identification (participle agreement)
Texte : Les recommandations que nous (1) avons présenté au comité la semaine dernière (2) ont été bien accueillies par (3) l'ensemble des membres.
- (1) avons présenté au comité
- (2) ont été bien accueillies
- (3) l'ensemble des membres
- (4) Aucun des choix offerts
Answer: (1). With avoir, the participle agrees with a direct object placed before the verb. Que stands for les recommandations (feminine plural), so the correct form is avons présentées. In (2), accueillies agrees with the subject of a passive, and (3) is correct.
Sample 4: error identification (no error)
Texte : Les analystes (1) se sont rendu compte que les données (2) qu'on leur avait transmises étaient incomplètes. Ils ont donc demandé (3) qu'on leur envoie une version corrigée.
- (1) se sont rendu compte
- (2) qu'on leur avait transmises
- (3) qu'on leur envoie
- (4) Aucun des choix offerts
Answer: (4). In (1), the direct object is compte, placed after the verb, so rendu stays invariable. In (2), que stands for les données and comes before avoir, so transmises agrees. In (3), demander que takes the subjunctive, and envoie is that form. If you picked (1), you agreed the participle with the wrong word.
Sample 5: a short email with two blanks
Texte : La séance de formation sur le nouveau système de gestion des dossiers ___ (blank 1) lieu le 14 novembre, à 9 h, dans la salle 305. Les documents sont déjà ___ (blank 2) votre disposition sur l'intranet. Merci de confirmer votre présence d'ici vendredi.
Question 1 (blank 1):
- (1) prendra
- (2) aura
- (3) fera
- (4) mettra
Question 2 (blank 2):
- (1) a
- (2) à
- (3) en
- (4) dans
Answers: (2) and (2). Blank 1 needs avoir lieu: prendra lieu blends it with the English take place, and fera and mettra form no expression with lieu. Blank 2 needs the fixed expression à votre disposition; without the accent, a is the verb avoir, and en and dans do not fit.
Strategies for the SLE writing test
- Read the whole sentence first. The deciding word often comes after the blank: a plural noun, a que or a conjunction that needs the subjunctive.
- Say each option in your head. Slot every option into the sentence. With reading practice, wrong forms start to sound wrong.
- Eliminate. Cross out options that break a rule you are sure of. Two eliminations leave a 50/50 at worst.
- Check agreements backwards. For each participle or adjective in bold, go back to the word it depends on and check gender and number.
- Watch the clock. 90 minutes for 65 items is about 80 seconds each. Choose your best answer and move on.
- Answer everything. There is no penalty for a wrong answer, so never leave a question blank.
- Use the free [PSC self-assessment tests](https://www.canada.ca/en/public-service-commission/services/second-language-testing-public-service/self-assessment-tests.html). They include a French writing test with an answer key; results are unofficial and the format differs from the real test.
PSC test content is Protected, so a site selling "real SLE questions" is a red flag.
Practise the SLE written expression test on Mocko
Mocko's SLE French practice runs in your browser, with explanations in English and questions in French. The written expression quick test follows the unsupervised format: 30 single-blank items in 45 minutes, top level C. The full mock exam includes the 65-question test with both item types, and Mocko converts your result onto the 55-point scale with the published bands.
A free account, no credit card, gives you two quick tests and one full mock exam of all three tests. The free mock's review explains your first missed question on each test; Premium explains every question. Start on the SLE French test page or create your free account.
FAQ
No. The SLE written expression test is entirely multiple choice. You choose the best word for a blank or identify which bold section of a sentence contains an error.
