The Oral Presentation Mark That Sits Between Two Bands
An Armenian continuers class of 14 students completes its first school-assessed oral presentation. The task asks students to speak for four minutes on a topic of personal significance, then answer two impromptu questions. The teacher, Ms. Haroutyunian, has a criteria sheet with four descriptors: Content and Ideas, Coherence and Structure, Language Resources, and Interaction. Each is scored out of 5, giving a raw total out of 20 that is later scaled to the school's assessment weighting. One student, Ani, speaks fluently for 4 minutes 20 seconds about her grandmother's migration story, using rich vocabulary and complex subordinate clauses, but her two impromptu answers are one sentence each and she misreads a question about her grandmother's hometown. Ms. Haroutyunian initially writes 5/5 for Language Resources, 5/5 for Content, 4/5 for Coherence, and 2/5 for Interaction, then pauses. A colleague points out that another student, Vartan, spoke less fluently but answered every question with developed, accurate responses and scored 4/5 on Interaction. Ani's parents have emailed asking why her raw score does not match her perceived fluency.
1. Identify the assessment principle most at risk in Ms. Haroutyunian's initial marking and explain how the criteria descriptors should resolve the apparent conflict between Ani's and Vartan's scores.
Read the worked answer
The principle at risk is validity - specifically, whether the marks awarded actually measure the construct the task was designed to assess. The task explicitly includes impromptu interaction as a scored component, so a student who cannot sustain interaction has not demonstrated the full construct, however fluent the monologue.
The descriptors should be applied criterion by criterion, not holistically. Ani's Interaction score of 2/5 is defensible if the descriptor for 2 states something like 'responds briefly; requires repetition; limited repair strategies.' Vartan's 4/5 is also defensible if his descriptor says 'responds with developed answers; occasional hesitation; effective repair.' The two scores are not in conflict because they sit on different criteria: Ani's Language Resources (5/5) and Vartan's Interaction (4/5) are independent judgments.
The risk is halo effect - letting Ani's strong fluency inflate her Interaction mark, or letting Vartan's weaker fluency depress his. Ms. Haroutyunian should re-read each descriptor against the evidence notes she took during the presentation, not against her memory of overall impression. If her notes do not support a 2 for Ani's Interaction (e.g., she actually gave two developed answers and only one misread), she should adjust upward. If they do, the 2 stands.
Finally, moderation protects against idiosyncratic marking: a colleague's sample of the same recordings should reach comparable judgments on the same descriptors before scores are finalised.
2. Draft the substance of a response to Ani's parents that explains the scoring without disclosing other students' results, and identify one procedural safeguard the school should apply before releasing the mark.
Read the worked answer
Substance of the reply (paraphrase, not verbatim):
- The presentation is assessed against four published criteria, each worth 5 marks. Fluency and vocabulary are assessed under Language Resources and Content; the ability to answer questions spontaneously is assessed separately under Interaction.
- Ani performed very strongly on the first two criteria. Her Interaction mark reflects the specific evidence of the two impromptu questions - brief responses and one misheard question - not her overall ability.
- The criteria and descriptors were provided to students before the task. The mark is a judgment against those descriptors, not a ranking against classmates.
- Ani will receive written feedback identifying exactly which descriptor elements were not yet met and what would lift the Interaction mark (e.g., paraphrasing a question she did not catch, extending answers with a reason or example).
Procedural safeguard: the school should complete internal moderation - a second teacher (or panel) reviews a sample of recordings, including Ani's, against the descriptors - and confirm the marks are consistent before they are entered into the assessment record and communicated. This protects both the student and the teacher, and it means any adjustment happens before the mark is treated as final rather than after a parent query.