Overview of Mallard Question Types
This documentation describes the standard question types and their
associated parameters. Additional question types may be available for
specific courses. Also included is an overview of question syntax,
a discussion of feedback and partial credit, and an explanation of the
formats used in describing parameter arguments.
Standard Question Types:
Note: In the actual documentation, the following links lead to example
quizzes on the site. Because this is not an actual Mallard site, the
interactivity is not possible on this page. The links lead only to the
question type documentation which appears at the top of the
example quizzes - there are no example questions included here.
Arith
Blank
List
M_of_N_choice
Match
Multiple_blank
Multiple_choice
Multiple_choice_group
True_false
Overview of Question Syntax:
- All Mallard filenames, names of question types, parameter names,
etc.
must be single words; in particular, no spaces are allowed. For this
reason, an underscore ("_") is typically used to connect what would
otherwise be two words (e.g. multiple_choice and partial_credit).
- All Mallard questions have the following syntax structure:
- <input type=...>
- <param ...>
- other input information
- </input>
- <answer>
- answer
- </answer>
- The input information is given between the <input> and
</input> tags and consists of:
- a required specification of the question type
- <input type=blank>
- optional parameter specifications
- <param size=25 feedback=high
display=pulldown>
Parameters control how a question is displayed, what factors need to be
checked in evaluation of the student response, assignment of credit, and
feedback. It is not necessary to list all parameters, because each
parameter has a default value.
- input information required for question types such as
multiple_choice, m_of_n_choice, match, true_false, and
multiple_choice_group.
- <input type=multiple_choice>
- selection #1
- selection #2
- selection #3
- selection #4
- </input>
- The answer information is specified between the <answer> and
</answer> tags. Examples include:
- a correct textual answer (and allowable alternatives) for a Blank
question (e.g., Illinois,IL,Ill)
- the number corresponding to the correct response for a
Multiple_choice question
- a numerical value or expression for an Arith question
(e.g., 2*3 or 6.0).
- an expression involving previously defined
random variables for
an Arith question (e.g., <var name=x1> * <var
name=x2>)
How Parameters Are Described:
The documentation for each question type includes a list and
descriptions of allowable parameters. In order to understand this
documentation, we explain here the terminology and notation of a few
representative examples. Note that each parameter has a default value,
which is used if the parameter value is not explicitly set to something
else. Hence, if the author accidentally says "feedback=mdium" instead of
"feedback=medium", the setting is not recognized and the feedback will
revert to its default value of feedback=low.
- feedback=[low, medium, high]
- Listed in bold within the square brackets "[]" are the
allowable values for the feedback parameter:
feedback=low, feedback=medium, or feedback=high.
The first value is always the
default; i.e., if no other parameter value is specified, then the
feedback is low for this question instance.
- Example: <param feedback=high>
- choice_text=[string] (Default is blank.) (Used
only with display=pulldown
- left_header1=[string] (Default is blank.)
- The italic string means that the author supplies text
in the form of a string
some of characters that may include letters, numbers, spaces,
punctuation, or other symbols.
Even HTML tags can be used to specify what will appear as labeling text.
This string could be a word, several words, or several characters.
By default, no text is displayed.
- Example: The following would display the message "You
can do it!" at the top of a pulldown menu:
- <param choice_text='You can do it!'>
- Example: The following would display a boldface heading (Column
A) centered above the left column in a match question:
- <param left_header1="<center><b>Column
A</b></center>">
- size=[number] (Default is the browser default.)
- The italic number means that the author supplies some
number (and not the word "number"). The following would set the size of
the input blank to 16.
- Example: <param size=16>
- non_shuffle_items=[list of integers] (Default is
none.) (Used only with shuffle=yes.)
- The italic list of integers means that the author specifies
a list of integers. For instance, "2,4-6,8,9" and "2,4,5,6,8,9" and
"2,4,5,6,8-9" correspond to all answer choices that are to remain in their
given locations. By default, there are no non_shuffle_items and so all
items are shuffled.
- Example: The following would exclude answer choices 2,4,7,8,9
from being shuffled. They would remain in
their given locations.
- <param non_shuffle_items="2,4,7-9">
Feedback and Partial Credit Parameters
All question types support the feedback parameter. This determines what
feedback is given to the student when a question is submitted for
correction. In general, correct answers are redisplayed in green, incorrect in red, and
partially correct (e.g. a correct list but in the wrong order or a word
spelled correctly but with incorrect capitalization) in yellow. Feedback messages may explain anything marked
as partially
correct and may give correct answer information as well. The possible
feedback options are as follows:
- feedback=low No information is given. Low feedback
(the
default) simply tells the student whether the the entire question
has been answered correctly (totally right) or incorrectly (not totally
right.) There is no partial information and no feedback message.
- feedback=high Total information is given. High feedback
not
only
tells the student what parts were answered correctly, but also gives away
the correct answers.
- feedback=medium Medium feedback lies
between these two extremes: documentation needs to be consulted on a
question-by-question basis. In some cases, there may be more than one
medium feedback level, such as medium1 and medium2.
Partial credit parameters control the score a student receives on
each question.
The partial_credit parameter is a "gateway" to other partial credit
parameters such as guessing_penalty, wrong_grammar_penalty, and
out_of_order_penalty. If
partial_credit=no, then no partial credit is
given: the question will get full credit (100 points) or no credit (0
points). If partial_credit=yes, then different parts of questions are
given equal weight; e.g., if there are N number of parts, then each part
is worth 100/N.
If partial_credit=yes, then other partial credit parameters designate how
much credit should be deducted for various types of errors.
Wrong_grammar_penalty will deduct for case, spacing, or punctuation
errors if grammar is being checked (i.e., if case, spacing, or
punc = check). The out_of_order penalty will deduct for answers supplied
in the wrong order if ordering is being checked (i.e., if
ordered=yes). The guessing_penalty allows an author to deduct
points for incorrect (as opposed to blank) responses.
(Guessing_penalty can be used only with questions that have multiple
responses, because no question can result in an overall negative score.)
The partial_credit and feedback parameters are completely independent.
Although perhaps illogical, an author could give partial credit and yet
give absolutely no feedback to hint at which items the student got right
or wrong.
For example, if an author wants only
the answer 'Washington' to be accepted, he would set
partial_credit=yes and case=check. This means that
if
a student answered 'washington', she would receive only partial credit
(the
default partial credit is 50%). If the author selected low feedback, the
student would only see that her answer is incorrect (highlighted in red) but that her raw score is 50%. This could seem
contradictory. Hence, if an author wishes to allow partial credit, it is
more logical to use medium or high level feedback.
Comments? Questions? General harassment? Mail it to
maiko@wocket.csl.uiuc.edu