The Midyis Test - Holbrook Academy

Report
The MidYIS Test
The CEM at Durham University
• The Centre for Evaluation and Monitoring is the largest
educational research unit in a UK university.
• Works with schools, colleges, education authorities and
government agencies to provide high-quality
information through scientifically grounded research,
using evidence rather than authority or opinion as a
guide to educational practice.
• Home of a widely-used family of baseline testing
systems including ALIS (6th form), Yellis (Key Stage 4),
MidYIS (Key Stage 3) and PIPS (Key Stage 2).
The MidYIS test
• MidYIS stands for Middle Years Information
System and is currently operating in over 3000
secondary schools.
• Tests are designed to measure ability and
aptitude for learning rather than achievement.
• Not an IQ Test, as it is designed to provide a
measure of ‘typical’ performance which can be
used to give an expected level of attainment at
GCSE based on the national average results.
Computer Adaptive Testing
• Adaptive test is based on growing bank of questions.
• The adaptive nature of the test means that all pupils are
challenged and receive a bespoke test suited to their ability.
• The most able pupils don't waste time on items that are too
easy for them and lower ability pupils are not discouraged
when faced with questions they cannot answer.
• Adaptive testing is considered the most efficient method of
ascertaining a measure of pupils' abilities.
• It offers pupils questions based on their answers to
previous questions, and can therefore quickly focus in on
each pupil's ability measure without requiring them to
answer reams of unnecessary questions: questions that
may be too easy (in the case of your more able pupils) or
too hard (for lower ability pupils).
What is tested?
• Vocabulary: The CEM’s research has shown that development
of vocabulary is not fundamentally affected by teaching and is
a vital component for success in all subjects.
• Maths: pupils are presented with a variety of mathematical
questions, ranging from basic arithmetic through to algebra.
• Non-verbal: a battery of tests looking at the ability to handle
shapes in order to test visual and spatial skills.
• Skills: This section is a test of proof-reading and also
perceptual speed and accuracy. Pupils have to identify
grammatical errors in short passages, such as spelling
mistakes, incorrect punctuation and capitalisation.
Standard Score
• A pupil’s raw MidYIS scores are standardised against the results of
everyone in the country taking part in the project to allow scores to
be compared with different year groups or pupils in different
schools.
• A score of 100 represents the national average.
• Only 2% of pupils nationally will score more than 130.
The scores can be summarised as follows:
• 76-87
Well below the national average
• 88-95
Below average
• 96-105
Average
• 106-112
Above Average
• 113-124
Well above average
• ≥125Far above average
MidYIS Band
The score for each component and for the overall
test is used to place pupils in a performance band.
Each band represents 25% of the national ability
range, with A being the highest. The scores for each
band are as follows:
• <90
Band D (bottom 25%)
• 91-99
Band C
• 100-109 Band B
• ≥110
Band A (top 25%)
Stanine
• The MidYIS Band places pupils into four ability
bands with equal numbers of pupils in each band.
• The ‘stanines’ divide the results scale into nine
more or less equally-spaced divisions, with
different numbers of pupils in each stanine.
• 9 is the highest and has the fewest pupils. Most
will be in the middle – stanines 4 to 6.
• When plotted on a graph, the results appear as a
‘bell curve’. Most pupils are, inevitably, ‘average’
and so the highest point of the curve is the
average score, standardised to 100.
The diagram shows how the four ability bands and the nine
stanines are mapped.
Percentile
• Measures relative performance by showing the
percentage of pupils who achieved a lower score
nationally in a particular component.
• E.G. a score of 91 would show that only 9% of
students in the national sample performed better
in that component.
• The percentile is often helpful in showing
discrepancies in performance between different
components.
Confidence Limits
Confidence Limits
• It is important to be aware that there is a margin for error in the
tests. A pupil could have sat the same test on a different day and
have got slightly different scores in individual components and
overall.
• The graph shows the score with an upper and lower limit for
potential variation – a 95% confidence limit. The Vocabulary score
achieved was 104. The graph shows us therefore that we can be
95% confident that the maximum this could have been is about 112
and the lowest about 95.
• The confidence limit is generally much narrower for the overall
score than for any of the individual components. As a rule we can
say this means that we can be very confident about the overall
score, but must be aware that there is a possible variation of about
±5.

similar documents