They add: "Feedback for students and their parents, as well as information for other teachers and educational decision-makers, is also based on (informal) teacher judgement."
The researchers, therefore, set out to measure how accurate this informal judgement is. The staff in the study were asked to rate pupils' ability in reading comprehension and maths.
First, teachers were asked to make general assessments of pupils' academic ability. The academics found these judgements were often free-floating verdicts, based on nothing more than teachers' memories of a particular pupil's ability.
"Students who were more similar to their teachers were judged more positively than students who were more dissimilar," they write. "Thus, personality similarity.systematically affects the (inaccurate) global judgement."
In particular, when teachers were called on to deliver judgements on the ability of more than one pupil, they based them on "pieces of information about the student.that are easier to access than information about relevant achievement".
For example, the sex of the pupil was conflated with stereotypes about boys' and girls' abilities. Teachers also found it easy to recall pupils' backgrounds, as well as whether or not they found individual pupils particularly likeable. Often, the more similar a pupil's background to their own, the more likeable they found that pupil.
But the findings were criticised by Amy Newsham, an English teacher at Cardinal Allen Catholic High School in Fleetwood, Lancashire. "I don't think teachers make assumptions about students," she said.
"Every child is different. As human beings, if somebody has similar attributes or a similar personality to you, or similar interests, you assume that they are like you in some way.
"Humans try to forge connections with other people. That's how you formulate relationships. And part of teaching is to forge those relationships.
"But I don't think teachers make assumptions about students. Professionally, you wouldn't."
When the researchers asked teachers to judge pupils on their performance in specific tasks, the findings were slightly different. Here, teachers had to estimate how students would perform. To provide a judgement, they therefore needed to assess the difficulty of the questions and balance that with knowledge of each pupil's ability.
These predictions ended up being more accurate, as any bias towards individual students was tempered by teachers' impartial assessment of the difficulty of the questions being asked.
Nonetheless, the academics conclude: "Teacher judgements can be considered inaccurate and far from being perfect."
A separate UK study by academics from the University of Bristol in 2009, reveals that teachers tend to underestimate the ability of pupils from ethnic minority backgrounds.
Using data taken from the key stage 2 tests of all 11-year-olds in England, academics examined the difference between pupils' written test scores and teachers' assessment of their ability.
Among white pupils, 12.4 per cent scored higher in their written tests than in their teacher assessments, compared with 20.2 per cent of Pakistani and 18.3 per cent of Black African pupils.
The researchers added that this effect was particularly pronounced in schools with a low proportion of ethnic minority pupils: in these cases, stereotypes tended to take the place of experience.