Presentation to EATING (Education and Technology Interest Group), 17th January 2008

advertisement
Investigating the use of short
answer free-text questions in
online interactive assessment
Sally Jordan
17th January 2008
Centre for Open Learning of Mathematics, Science
Computing and Technology (COLMSCT)
My plan
•
•
•
•
•
•
History – how and why did I get involved
What we are trying to achieve
Our questions
Evaluation
How we write the questions – have a go!
Discussion
The S151 : Maths for Science
experience
• This course does not have TMAs;
• But we wanted to be able to provide students with
targeted and relatively detailed feedback on their
answers;
• And we wanted to be able to provide this feedback
rapidly;
• We wanted more than multiple choice questions;
• And this was for summative assessment.
Since S151…
• The eAssessment system has become OpenMark, being
used on several courses in the Science Faculty and
elsewhere in the University, for diagnostic, formative and
summative purposes;
• OpenMark has been incorporated within Moodle at both
the assessment and the question level;
• Around 10 COLMSCT projects are investigating ways in
which the use of this type of assessment can be
extended;
• My project is a joint one with Barbara Brockbank,
supported by Phil Butcher and Tom Mitchell (Intelligent
Assessment Technoloies Ltd.)
S104 : Exploring science
• A new course from February 2008;
• Students will spend 9 months studying the course, and
we want to keep them up to date and engaged with it;
• So we are using regular iCMAs (interactive computer
marked assignments) with feedback (alongside
conventional tutor marked assignments);
• The iCMAs will be summative (but low stakes), so that
students do them, but their primary function is to provide
pacing and feedback.
Questions for S104 : Exploring
science
• We want to be able to ask questions that require slightly
longer free text answers;
• So we are working with a commercially provided,
linguistically based authoring tool to write questions that
require answers of about a sentence in length;
• Student responses are being used to refine the
questions;
• We are providing targeted feedback on incorrect and
incomplete answers.
Evaluation 1:
IET research lab observations
• Six S103 students were observed in June 2007;
• They reacted to the free-text questions in interesting ways
e.g. because they thought we were just looking for
keywords, some tried to ‘help’ the computer by giving
answers in note form; one of the students appeared to
‘brain dump’;
• Use made of the feedback provided was also variable; one
student said he found it useful but clearly didn’t use it;
others read the feedback carefully, checked things in their
course book, and altered their response successfully.
Evaluation 2:
Human-computer marking comparison
• Computer marking (with and without ‘flagging’) compared
with that of 6 ALs and the question author;
• For most questions the computer’s marking is
indistinguishable from that of the ALs;
• Perhaps not surprisingly, the computer’s marking is closer
to that of the question author than that of some of the ALs;
• The computer is not always ‘right’, but neither are the ALs;
• The computer seems to behave better when credit is not
given for flagged answers.
Writing questions and answer
matching
• Have a go..
• We’ve written the question
for you….
In the photograph on the right,
who is the taller?
Interesting issues
• Answers that are difficult to parse include those that are
very long and those in note form
• Questions have to be quite tightly constrained e.g.
‘You are handed a rock specimen that consists of
interlocking crystals. How would you decide, from its
appearance, whether it is an igneous or a metamorphic
rock?’
has become
‘You are handed a rock specimen that consists of
interlocking crystals. How could you be sure, from its
appearance, that this was a metamorphic rock?’
More interesting issues
• Writing questions and improving the answer matching is
interesting, but time consuming;
• I found it relatively easy to get used to writing
appropriate questions and using the authoring tool, but I
am used to writing assessment questions and am quite
logical (I’m a physicist!);
• But one of our tasks is to investigate whether these sorts
of questions could be written and used more widely
across the University, and to compare with other
systems;
• And who should be writing the questions?
Short answer questions
https://students.open.ac.uk/openmark/s103-07j.blocks123world/
https://students.open.ac.uk/openmark/s103-07j.block4world/
https://students.open.ac.uk/openmark/s103-07j.block5world
https://students.open.ac.uk/openmark/s103-07b.block7v1aworld/
https://students.open.ac.uk/openmark/s103-07b.block8world/
https://students.open.ac.uk/openmark/s103-07b.block9world/
https://students.open.ac.uk/openmark/s103-07b.block10world/
https://students.open.ac.uk/openmark/s103-07b.block11world/
Workshop questions (developmental server)
http://kestrel.open.ac.uk/om-tn/iat.demo/
OpenMark examples site
http://www.open.ac.uk/openmarkexamples/index.shtml
Centre for Open Learning of Mathematics, Science
Computing and Technology (COLMSCT)
The Open University
Walton Hall
Milton Keynes
MK7 6AA
http://cetl.open.ac.uk/colmsct
s.e.jordan@open.ac.uk
Download