Rubric design - Edith Cowan University

advertisement
Edith Cowan University
Centre for Learning and Development
Rubric design
There are two broad types of rubrics: analytic and holistic.
Holistic
Holistic rubrics are most often used as general descriptors of the overall quality of student achievement. As
an example, the Facione and Facione Holistic Scoring Rubric (1994) for Critical Thinking is copied below and
is available free, with a page of instructions.
Holistic rubrics such as the one below are useful as an overall guide for students to aspire to. However,
these aspirations may be unevenly realised. For example, in this example a student has presented several
qualities generally attributed to excellent students, but also some at a lower level. If this rubric were to be
used for grading it would be difficult to decide what the final grade should be – perhaps an average? But
what if the poor achievement - related to interpretation of evidence and to reasoned judgement - were
actually the most important desired quality?
4 Consistently does all or almost all of the following:
Accurately interprets evidence, statements, graphics, questions, etc.
Identifies the salient arguments (reasons and claims) pro and con.
Thoughtfully analyzes and evaluates major alternative points of view.
Draws warranted, judicious, non-fallacious, conclusions.
Justifies key results and procedures, explains assumptions and reasons.
Fair-mindedly follows where evidence and reasons lead.
3 Does most or many of the following:
Accurately interprets evidence, statements, graphics, questions, etc.
Identifies relevant arguments (reasons and claims) pro and con.
Offers analyses and evaluations of obvious alternative points of view.
Justifies some results or procedures, explains reasons.
Fair-mindedly follows where evidence and reasons lead.
2 Does most or many of the following:
Misinterprets evidence, statements, graphics, questions, etc.
Fails to identify strong, relevant counter-arguments.
Ignores or superficially evaluates obvious alternative points of view.
Justifies few results or procedures, seldom explains reasons.
Regardless of the evidence or reasons maintains or defends views based on self-interest or preconceptions.
1 Consistently does all or almost all of the following:
Offers biased interpretations of evidence, statements, graphics, questions, information, or the points of view of
others.
Fails to identify or hastily dismisses strong, relevant counter-arguments.
Ignores or superficially evaluates obvious alternative points of view.
Argues using fallacious or irrelevant reasons, and unwarranted claims.
Regardless of the evidence or reasons, maintains or defends views based on self-interest or preconceptions.
Exhibits close-mindedness or hostility to reason.
Document1
1
Edith Cowan University
Centre for Learning and Development
Analytic
Most of the rubrics you've come across are probably analytic rubrics that are used as scoring, or grading
guides, for staff and students. Analytic rubrics are generally represented in table format with a list of
criteria down the left side and gradations of quality (of a performance or a product) across the top.
For example, Mid-South Community College (West Memphis, AR) has an Analytical rubric for critical
thinking.
Component
Identifies and
summarises the
problem/question
at issue
4
Accurately identifies the
problem/question and
provides a welldeveloped summary.
3
Accurately identifies
the
problem/question
and provides a brief
summary.
Identifies and
assesses the
quality of
supporting
data/evidence
Provides well-developed
examination of the
evidence and questions
its accuracy, relevance,
and completeness.
Clearly distinguishes
between fact and
opinion.
Accurately identifies and
provides a welldeveloped explanation
of contextual issues with
a clear sense of scope.
Examines evidence
and questions the
quality. Distinguishes
between fact and
opinion.
Accurately identifies the
author’s meaning
and/or potential bias
and provides a welldeveloped explanation.
Accurately identifies
meaning and/or bias
and provides a brief
explanation.
Accurately identifies
conclusions,
implications, and
consequences with a
well-developed
explanation. Provides an
objective reflection of
own assertions.
Accurately identifies
conclusions,
implications, and
consequences with a
brief evaluative
summary.
Identifies and
considers the
influence of the
context* on the
issue
Demonstrates
higher level
thinking by
interpreting the
author’s meaning
or the potential
bias
Identifies and
evaluates
conclusions,
implications, and
consequences
Document1
Accurately identifies
and provides an
explanation of
potential contextual
issues.
2
Identifies the
problem/question
and provides a poor
summary or
identifies an
inappropriate
problem/question.
Merely repeats
information
provided. Does not
justify position or
distinguish between
fact and opinion.
1
Does not identify or
summarise the
problem/question
accurately if at all.
Does not explain
contextual issues;
provides inaccurate
information; or
merely provides a
list.
Does not explain,
provides inaccurate
information, or
merely lists potential
bias or inferred
meanings.
Does not identify or
consider any
contextual issues.
Does not explain,
provides inaccurate
information, or
merely provides a list
of ideas; or only
discusses one area.
Does not identify or
evaluate any
conclusions,
implications or
consequences.
Does not identify or
assess the quality
of supporting
evidence.
2
Edith Cowan University
Centre for Learning and Development
Rubrics can also be arranged the other way, with criteria across the top and gradations of quality down the
left, such as this IELTS rubric (follow link to download full rubric).
If used for scoring, an analytic rubric will provide a separate score for each criterion, although some
examples will combine a number of criteria together. This results in a rubric somewhere between an
analytic and a holistic one. These scores may be weighted before being combined to produce a composite
score for overall performance.
Why would you not use a grid design?
Using a grid design, all criteria need to have the same number of gradations. Each level should represent a
quality that is clearly different from the one before or the one after. Unfortunately, in the real world, not all
criteria have the same number of clearly discernible qualitative differences.
If a rubric is to effectively support validity, reliability and transparency, the gradations should only describe
clearly discernible qualitative differences.
Here is an example of how Cleveland State University got around the problem of criteria with an unequal
number of levels in their Discussion Rubric (follow link to download rubric).
Document1
3
Edith Cowan University
Centre for Learning and Development
How many criteria should there be?
The number of criteria identified in your rubric will depend on the number of discrete learning outcomes
anticipated from the task or unit. Holistic rubrics only have one criterion, with gradations of quality
described across multiple criteria. Grading rubrics are quite often holistic.
How many levels should there be?
There is no specific number of levels a rubric should or should not possess. More levels result in more finegrained distinction between performances – but only if the performances have clear qualitative differences.
While some criteria may only have two levels (done or not done) others may have four or five. If you wish
to assure reliability (consistency of marking by different markers, or by the same marker at different times)
the most important factor is whether each level of performance described can be clearly and consistently
differentiated from other levels when evaluating student work.
Document1
4
Download