At AlphaPlus we develop a range of assessments for different clients. Each assessment produces indications of a learner’s performance; this article discusses two common approaches we use to help learners and teachers interpret this information.
Why do we need standardised scores?
All assessments result in raw scores. This relates to the total number of marks a learner receives in an assessment. Marks can be awarded simply one per question the learner answers correctly, or by using a mark scheme to evaluate a learner’s answer and award a score (these latter are sometimes called ‘partial credit’ or multimark questions).
These raw scores are normally easy to understand in relation to a single assessment, e.g. ‘I got 9/10 in my spelling test’. However, they are less easy to use when comparing results from one assessment with results from another. Sticking with the spelling test example, interpreting the results could depend on a few things:
- How difficult were the words tested?
In a series of spelling tests, a learner might get 9/10 in the first week, because only simple words were tested, whereas, the next week they score 5/10 because the next test focused on more complicated words.
- When was the test taken?
Three learners could take the same spelling test and the expectation for performance for each will be quite different. For example, one learner from Year 4 takes the assessment at the beginning of the year, another from Year 4 takes it at the end of the year and a learner from Year 6 takes it in the middle of the year.
Where assessments have been trialled and standardised with a large and representative group of learners, we are able to convert the raw scores from different assessments into standardised scores. This enables learners’ outcomes to be compared using a single scale that accounts for the different difficulties of questions and/or the different ages of learners so it is easier to compare performance.
What is a standardised score?
A standardised score is calculated by comparing an individual learner’s outcome with the outcomes for other learners in their year group or their defined cohort. Raw marks from one or several different assessments are converted, using a well-established statistical method, into a number within a given range that follows a particular format, for example, 950 to 1050. In this example, the average score is 1000. This means that learners who score below 1000 have performed below average for their cohort, whereas those with a score above 1000 have performed better than average. The scores take account of the fact that different learners may have taken different tests of slightly different difficulties – they are “standardised”.
It is important to clearly define the cohort of students for which the standardised score will be used – i.e. what other learners’ scores is the score being standardised against. Following the spelling test example, if the cohort is ‘whole academic year 6 population in England’ then scoring 1,000 means the performance is ‘average’. However, this would reflect an extraordinarily good outcome for a non-English first language learner who had only been in the English school system for a number of months, reflecting strong second language command for a young learner.
What is an age standardised score?
The standardised score described above allows for learners taking different assessments, but does not make allowances for the age of the learner completing the assessment within their defined cohort. In the UK most classes can learners of up to twelve months apart in age (i.e. some are born in September and some are born in the following August); whether or not learners have had this twelve months of cognitive development can be significant, particularly for younger learners.
An age standardised score accounts for this by comparing the learners’ assessment outcomes with the scores of other learners who were born in the same month and year as them. It then adjusts for this using what amounts to a slightly tweaked standardised score approach (as described above).
This means that a younger learner could have a lower raw score than an older learner in the same assessment, but receive the same or higher age standardised score. This is because these scores are referenced to the learner’s own month-specific age group: the younger learner is being compared with other learners of the same age as them, and is performing better relative to this group than the older learner is compared to their month-specific age group.
What are the benefits of using standardised scores?
Standardised and age-standardised scores are one of a number of tools that can identify the relative performance of learners in relation to their peers across multiple assessments. They can be used to implement catch up interventions for those who are underperforming and put in place additional stretch and challenge for those who are exceeding expectations.
Objective scores from externally created assessments can also provide a useful check against independent working assumptions made by teachers about learners’ abilities.