Traditional marking involves individual teachers evaluating a piece of writing by asking whether or not it meets the criteria for a particular grade or mark, while Comparative Judgement involves lining up two scripts against each other and deciding which you prefer. Research has shown that people are not very good at absolute judgement, which is the type of evaluation involved when writing is marked against a rubric. Indeed, teachers not only do not often agree with each other, but sometimes do not even agree with themselves from one day to the next. By contrast, Comparative Judgement has been shown to be quicker and more reliable than marking that relies on absolute judgement or evaluation against a rubric.

Comparative Judgement, by No More Marking (a UK-based company), has been used in the UK for 7-8 years, in the US for 5 years, in Australia for 2 years, and has just completed its first project with schools in New Zealand. This approach can be used both to mark student scripts and to inform the teaching of writing. It works by placing students on a consistent scaled score. Scaled scores underpin most forms of assessment, and are valuable because they provide more nuanced information than broad categories such as ‘emerging’ or ‘exceeding’. National scales for different countries are developed once a certain number of scripts (about 5-10,000) have been entered into the online platform and evaluated using Comparative Judgement. No More Marking also calculates students’ writing age using the average score for each birth month of the participating students (so if the average score for students aged 11 years 0 months is 53, a student who scores 53 will be assigned a writing age of 11 years 0 months). It also uses categories from the local system (such as NAPLAN in Australia) to assign students a level. As there is no equivalent local system for establishing grade boundaries in New Zealand, participating teachers in New Zealand projects will probably be invited to define the levels.

In terms of using Comparative Judgement data to inform the teaching of writing, data shows that there are more similarities than differences in terms of students’ writing ability throughout the world. For example, the data shows that all good writers are good spellers, so effective teaching to support accurate spelling is a high-leverage way to enable students to become skilled writers. Students also often regard length as a proxy for sentence accuracy, and like to use conjunctions like ‘because’, ‘but’, and ‘so’. However, the data show that they tend not to understand that these words work differently, nor do they know how to use them in sophisticated ways. Similarly, the data shows that students who do not understand the difference between a fragment and a sentence, a run-on and a sentence, and a comma splice and sentence, are less likely to score highly with Comparative Judgement. This indicates that these are also high leverage skills to teach when teaching at the sentence level.