5.2 Assessment Tasks to Build into a Lesson or Series

Choosing the task when you write the plan

Once you know why you need evidence (Section 5.1), Module 2 asks which task you will actually put in the procedure or the scheme of work. Three filters sit on the same planning line:

1. Aim / construct. If the stage aim is gist reading, the check must make learners show they got the overall message. If the aim is oral fluency in agreeing and disagreeing, the check must make them talk at length, not tick boxes about discourse-marker definitions.
2. Learners. Age, literacy, class size, L1, time of day, and whether they have seen this text type before all change what is fair. A 12-item cloze that works with literate adults may be the wrong load for a tired evening class of 14 who met the topic today.
3. Informal or formal. The same role-play can be a two-minute listen-in (informal) or a recorded, rubric-marked mid-term speaking sample (formal). Choose the marking load when you choose the task, not after you run out of register columns.

OpenExamPrep treats these filters as independent study material for TKT Module 2 planning items. The chapter does not claim that any live paper uses these classroom stories.

Selected-response tasks you might build in

Selected-response tasks give a closed set of answers. They are fast to mark, easy to use formally, and easy to misuse if you let the format choose the construct.

• Multiple choice. Useful for gist or detail in reading or listening when the options are plausible, or for checking whether learners recognise a function. Poor for proving they can produce the function in speech. A 20-item MCQ on I agree because does not test oral fluency.
• Matching. Strong for headings to paragraphs, words to definitions, or functions to exponents. A matching-headings task can be a fair gist or main-idea check. Matching verb forms to gaps is a form check, not a gist check, even if the text is a restaurant review.
• True/false. Fast informal gist or detail check after a short text or audio. Plan an “evidence word” rule so guessing is visible: learners underline the phrase that made the sentence true or false.
• Gap-fill. Typical for a taught form or lexical set (a/an/some/any after a food unit). Informal if self-checked; formal if scored. A tense gap-fill on a reading text does not test gist reading of that text.
• Cloze. A longer text with systematic deletions (often every nth word, or selected content/function words). Can sample vocabulary, grammar, and cohesion together. Still an indirect, mostly written, selected-response-like task — not a speaking test and not automatically an achievement test of a whole course.
• Error identification. Learners underline and correct a taught error type (articles, past simple, word order). Fair only if that form was taught. Do not drop in relative-clause errors after a lesson that only presented polite requests.
• Reordering. Sentence or paragraph reconstruction. Can check organisation or a taught pattern (request → reason → thanks). Weak evidence of spontaneous speaking; strong as a short written or cut-up informal check.

Selected-response work is often the right in-lesson check because it is quick. It is the wrong later formal task when the aim was a performance skill you never sampled.

Performance tasks, then the tools that make them fair

Performance tasks ask learners to do the skill: speak, interact, or write connected text.

• Interviews. Teacher–learner or paired. Good for spoken accuracy, interaction, or a diagnostic picture of range. Expensive in time: 28 individual interviews will not fit a 45-minute lesson.
• Role-plays. Fit functional aims (hotel desk, restaurant, complaint). Informal: monitor two or three pairs. Formal: record a sample and mark a grid.
• Information-gap. Each learner has information the partner needs. Strong for interaction and asking questions. If one sheet is a script to read aloud, you have accidentally planned reading aloud, not an information-gap speaking assessment.
• Guided writing. A short email, review, or paragraph with prompts, word bank, or paragraph frames. Matches writing aims and can be formal. If the aim was organisation, the rubric must score organisation, not only past-simple accuracy.
• Presentations. Extended speaking and organisation. Plan the time (including questions), the criteria, and whether peers will use the same grid.

These tasks only become assessable in a Module 2 sense when you also plan how evidence will be judged:

• Rubrics. Band descriptors for the construct you named (fluency, interaction, task completion, pronunciation of key items, organisation). The same rubric can support teacher marking, peer review, and a later formal grade.
• Self-checklists. Tick lists taken from the lesson aim (I greeted the guest; I used could or would; I asked a follow-up question). Best as informal or formative tools at the end of a stage.
• Peer review. Learners use the same criteria. Without a grid, comments collapse into “good pronunciation” and you have no evidence of the aim.
• Portfolios. A planned sequence of performance pieces (drafts, checklists, redrafts, one recorded speaking sample). The portfolio is not a different skill; it is a container for tasks you already chose.

Match the task to the construct

Construct here means the ability named in the aim. Module 2 items often hang on a mismatch:

• Gist reading is not tested by a tense gap-fill on the same article.
• Specific-information listening is not tested by a grammar MCQ about tense names after the audio.
• Oral fluency is not tested by a 20-item MCQ on discourse markers, and it is not tested by reading a dialogue aloud for weak forms.
• Written organisation is not tested by error identification of articles alone.

If the stem says the teacher wants evidence of X, the correct activity is the one that elicits X, at the right level of formality, for these learners, in the time available.

Planning constraints: practicality, fairness, time, washback

Practicality is whether you can administer, supervise, and mark the task with the people, rooms, and minutes you actually have. A beautiful interview protocol that needs two trained markers and a quiet room is not a plan for one teacher with 28 teenagers and a thin wall.

Fairness includes whether every learner had a chance to learn what you are sampling, whether instructions are clear, and whether one pair’s noise or one learner’s missing glasses should not decide the grade. Informal checks can be fair if they sample widely; they become unfair as formal grades if you only overheard the two pairs nearest the desk.

Time is both lesson minutes and marking minutes. Budget the check inside the stage. An exit ticket needs four minutes plus a two-minute sort after class. A formal writing sample needs drafting time and a realistic marking evening.

Washback is the effect of the assessment on teaching and learning. If the only formal speaking sample is reading a paragraph aloud, classes will practise reading aloud. If the formal task is a paired information-gap marked for interaction, classes will practise interaction. Plan the formal task you are willing to live with for three weeks of lessons.

Illustrative classroom minutes a teacher might budget (not an official timing table):

Worked examples: stage aim → in-lesson check → later formal task

Example A — gist reading. Aim: learners can identify the writer’s overall view in a short restaurant review. In-lesson informal check: two true/false statements (The writer would go back; The writer liked the dessert more than the service) plus underline the sentence that decided each answer. Later formal task in the unit test: matching four headings to four short reviews. Rejected task: a 10-item past-simple gap-fill on review 2. That samples narrative tense, which was not this stage aim.

Example B — oral fluency. Aim: learners can keep a conversation going when agreeing and disagreeing about weekend plans. In-lesson: information-gap diaries (A is free Saturday morning; B is free Sunday) with the teacher listening in and noting who asks follow-up questions. Later formal task: a three-minute paired discussion on a new prompt, marked on a fluency/interaction rubric. Rejected task: a 20-item MCQ on the meaning of on the other hand. That samples recognition of markers, not fluency.

Example C — written accuracy of a taught form. Aim: learners can use past simple in a short anecdote about a meal that went wrong. In-lesson: error identification on six sentences containing yesterday/last night time markers. Later formal task: guided writing (80 words) with a rubric that includes a language-accuracy criterion and task completion. If you only score ideas, you never assessed the aim; if you only score verbs, you ignored whether they wrote an anecdote.

Example D — listening for specific information. Aim: learners can pick price, time, and platform from a short station announcement. In-lesson: a table to complete (train 1 / train 2). Later formal: a similar announcement with a new table. Rejected task: multiple-choice questions on the grammar label “present continuous for future arrangements” after the audio. The recording was a vehicle for detail listening, not a grammar presentation.

Comparison table: task, formality, construct, caution

Traps

• Testing what was not taught. Relative-clause error identification after a polite-request lesson is not a valid progress check. Sample the aim you wrote.
• A “speaking test” that is only reading aloud. Pronunciation of a written dialogue can be a pronunciation or reading-aloud check. It is not oral fluency or interactive speaking unless learners have to invent turns.
• Peer assessment with no criteria. “Tell your partner two things you liked” may be nice classroom climate. It is not peer-assessment of the aim unless they use a grid that names the construct.
• Letting the worksheet choose the skill. If the coursebook page is a gap-fill, you may still need a different check for a gist or speaking aim — or you must change the aim to match the task, on purpose.
• Ignoring practicality. Formal individual orals for a large class belong in a series (rotation, continuous sampling), not in a single procedure line that cannot happen.

When you face a Module 2 matching item about assessment activities, name the aim in one phrase, name informal or formal, then keep only the task that could produce that evidence in that classroom. That is the whole skill this section trains.

The bar chart is a planning aid, not a published exam specification. Use it to sanity-check a procedure: if the lesson is 45 minutes and you have already used 30 on presentation and controlled practice, a 25-minute presentation-plus-peer-grid does not fit as an “in-lesson check.” Move the performance sample to a later lesson in the series, or choose a lighter informal check today and keep the formal rubric task for the scheme-of-work slot you can actually staff and mark.