IMS Question and Test Interoperability Information Model

Attribute : identifier [1]: string

Attribute : title [1]: string
The title of an assessmentItem is intended to enable the item to be selected in situations where the full text of the itemBody is not available, for example when a candidate is browsing a set of items to determine the order in which to attempt them. Therefore, delivery engines may reveal the title to candidates at any time but are not required to do so.

Attribute : label [0..1]: string256

Attribute : lang [0..1]: language

Attribute : adaptive [1]: boolean = false
Items are classified into Adaptive Items and Non-adaptive Items.

Attribute : timeDependent [1]: boolean

Attribute : toolName [0..1]: string256
The tool name attribute allows the tool creating the item to identify itself. Other processing systems may use this information to interpret the content of application specific data, such as labels on the elements of the item's itemBody.

Attribute : toolVersion [0..1]: string256
The tool version attribute allows the tool creating the item to identify its version. This value must only be interpreted in the context of the toolName

Contains : responseDeclaration [*]

Contains : outcomeDeclaration [*]

Contains : templateDeclaration [*]

Contains : templateProcessing [0..1]

Contains : stylesheet [0..*]

Contains : itemBody [0..1]

Contains : responseProcessing [0..1]

Contains : modalFeedback [*]

itemSession is an abstract class to help illustrate the requirements on Delivery Engines when delivering to candidates items that conform to this specification.

Associated with : assessmentItem [1]
An itemSession is associated with one and only one assessmentItem.

Attribute : completionStatus [1]: identifier
Delivery Engines must maintain the value of the built-in outcome variable completionStatus as part of the session state. It starts with the reserved value "not_attempted". At the start of the first attempt it changes the to the reserved value "unknown". It remains with this value for the duration of the item session unless set to a different value by a setOutcomeValue rule in responseProcessing. There are four permitted values: completed, incomplete, not_attempted and unknown. Any one of these values may be set during response processing, for definitions of the meanings see [CMI]. If an Adaptive Item sets completionStatus to complete then the session must be placed into the closed state, however, an itemSession is not required to wait for the complete signal before terminating, it may terminate in response to a direct request from the candidate, through running out of time or through some other exceptional circumstance. Similarly, Non-adaptive Items are not required to set a value for completionStatus, however, Adaptive Items must maintain a suitable value and should set completionStatus to "complete" to indicate when the cycle of interaction, response processing and feedback must stop. Delivery Engines are encouraged to use the value of completionStatus when communicating using [CMI]. See the accompanying integration guide for more details.

Attribute : duration [0..1]: float
Systems that support Time Dependent Items must record the duration of the session. The duration is defined as being the accumulated time (in seconds) of all Candidate Sessions for all Attempts. In other words the time between the beginning and the end of the itemSession minus any time the itemSession was in the suspended state. The resolution of the duration must be at least 1s and should be 0.1s or smaller. If the resolution is denoted by epsilon then each value of duration represents the range of values [duration,duration+epsilon). In other words, duration values are truncated. For items that are not time dependent duration must not be used.

Contains : itemVariable [*]
The itemSession keeps track of the current values assigned to all itemVariables. The values of completionStatus and duration are treated as special item variables. They share the same namespace as the item variables explicitly declared through variableDeclarations.

Contains : sessionContext [1]
An itemSession is also associated with a sessionContext which provides information about the candidate, when and where the session took place and so on.

The following diagram illustrates the user-perceived states of the itemSession. Not all states will apply to every scenario, for example feedback may not be provided for an item or it may not be allowed in the context in which the item is being used. Similarly, the candidate may not be permitted to review their responses and/or examine a model solution. In practice, systems may support only a limited number of the indicated state transitions and/or support other state transitions not shown here.

For system developers, an important first step in determining which requirements apply to their system is to identify which of the user-perceived states are supported in their system and to match the state transitions indicated in the diagram to their own event model.

Lifecycle of an Item Session

A delivery system notionally creates an instance of an itemSession object when it first becomes eligible for delivery to the candidate. The itemSession's state is then maintained and updated in response to the actions of the candidate until the session is over. At this point the state of the session is turned into a session report (or thrown away). A delivery system may also allow a session report from a past session to be used to re-create the session in order to allow a candidate's responses to be seen in the context of the item itself (and possibly compared to a solution).

The initial state of an itemSession represents the state after it has been determined that the item will be delivered to the candidate but before the delivery has taken place.

In a typical non-Adaptive Test the items are selected in advance and the candidate's interaction with all items is reported at the end of the test session, regardless of whether or not the candidate actually attempted all the items. In effect, itemSessions are created in the initial state for all items at the start of the test and are maintained in parallel. In an Adaptive Test the items that are to be presented are selected during the session based on the responses and outcomes associated with the items presented so far. Items are selected from a large pool and the delivery engine only reports the candidate's interaction with items that have actually been selected.

A candidate's interaction with an item is broken into 0 or more attempts. During each attempt the candidate interacts with the item through one or more candidate sessions. At the end of a candidate session the item is placed into the suspended state ready for the next candidate session. During a candidate session the itemSession is in the interacting state. Once an attempt has ended response processing takes place, after response processing a new attempt may be started.

For non-adaptive items, response processing may only be invoked a limited number of times, typically once. For adaptive items, no such limit is required because the response processing adapts the values it assigns to the outcome variables based on the path through the item. In both cases, each invocation of response processing indicates the end of an attempt. The appearance of the item's body, and whether any modal feedback is shown, is determined by the values of the outcomeVariables.

When no more attempts are allowed the itemSession passes into the closed state. Once in the closed state the values of the response variables are fixed. A delivery system or reporting tool may still allow the item to be presented after it has reached the closed state. This type of presentation takes place in the review state, summary feedback may also be visible at this point if response processing has taken place and set a suitable outcomeVariable.

Finally, for systems that support the display of solutions, the itemSession may pass into the solution state. In this state, the candidate's responses are temporarily replaced by the correct values supplied in the corresponding responseDeclarations (or NULL if none was declared).

The details of sessionContext will be application specific and are therefore outside the scope of this document. Applications that handle result reports using [RR] should consider the limits of the data model imposed on the context element in that specification.

Variable Declarations

Item variables are declared by variable declarations. All variables must be declared except for the built-in session variables referred to below which are declared implicitly. The purpose of the declaration is to associate an identifier with the variable and to identify the runtime type of the variable's value. At runtime (i.e., during an itemSession) the value of the variable is notionally represented by a class derived from itemVariable

Attribute : identifier [1]: identifier
The identifiers of the built-in session variables are reserved. They are completionStatus and duration. All item variables declared in an item share the same namespace. Different items have different namespaces.

Attribute : cardinality [1]: cardinality
Each variable is either single valued or multi-valued. Multi-valued variables are referred to as containers and come in ordered, unordered and record types. See cardinality for more information.

Attribute : baseType [0..1]: baseType
The value space from which the variable's value can be drawn (or in the case of containers, from which the individual values are drawn) is identified with a baseType. The baseType selects one of a small set of predefined types that are considered to have atomic values within the runtime data model. Variables with record cardinality have no base-type.

Contains : defaultValue [0..1]
An optional default value for the variable. The point at which a variable is set to its default value varies depending on the type of item variable.

Associated with : variableDeclaration [1]
At runtime, item variables are created in the itemSession each corresponding to a variableDeclaration in the corresponding assessmentItem.

Attribute : identifier [1]: identifier
The purpose of an itemVariable is to associate the runtime value of the variable with the variable's identifier and declaration. At runtime the variable has the cardinality and baseType given in the associated declaration

Contains : value [*]
An itemVariable may have no value at all, in which case it is said to have the special value NULL. For example, if the candidate has not yet had an opportunity to respond to an interaction then any associated responseVariable will have a NULL value. Empty containers and empty strings are always treated as NULL values.

A class that can represent a single value of any baseType in variable declarations. The base-type is defined by the baseType attribute of the declaration except in the case of variables with record cardinality.

Attribute : fieldIdentifier [0..1]: identifier
This attribute is used for specifying the field identifier for a value that forms part of a record.

Attribute : baseType [0..1]: baseType
This attribute is used for specifying the base-type of a value that forms part of a record.

Attribute : interpretation [0..1]: string
A human readable interpretation of the default value.

Contains : value [1..*]

single

multiple

ordered

record

An expression or itemVariable can either be single-valued or multi-valued. A multi-valued expression (or variable) is called a container. A container contains a list of values, this list may be empty in which case it is treated as NULL. All the values in a multiple or ordered container are drawn from the same value set, however, containers may contain multiple occurrences of the same value. In other words, [A,B,B,C] is an acceptable value for a container. A container with cardinality multiple and value [A,B,C] is equivalent to a similar one with value [C,B,A] whereas these two values would be considered distinct for containers with cardinality ordered. When used as the value of a responseVariable this distinction is typified by the difference between selecting choices in a multi-response multi-choice task and ranking choices in an order objects task. In the language of [ISO11404] a container with multiple cardinality is a "bag-type", a container with ordered cardinality is a "sequence-type" and a container with record cardinality is a "record-type".

The record container type is a special container that contains a set of independent values each identified by its own identifier and having its own base-type. This specification does not make use of the record type directly however it is provided to enable customInteractions to manipulate more complex responses and customOperators to return more complex values.

A base-type is simply a description of a set of atomic values (atomic to this specification). Note that several of the baseTypes used to define the runtime data model have identical definitions to those of the basic data types used to define the values for attributes in the specification itself. The use of an enumeration to define the set of baseTypes used in the runtime model, as opposed to the use of classes with similar names, is designed to help distinguish between these two distinct levels of modeling.

identifier
The set of identifier values is the same as the set of values defined by the identifier class

boolean
The set of boolean values is the same as the set of values defined by the boolean class.

integer
The set of integer values is the same as the set of values defined by the integer class.

float
The set of float values is the same as the set of values defined by the float class.

string
The set of string values is the same as the set of values defined by the string class.

point
A point value represents an integer tuple corresponding to a graphic point. The two integers correspond to the horizontal (x-axis) and vertical (y-axis) positions respectively. The up/down and left/right senses of the axes are context dependent.

pair
A pair value represents a pair of identifiers corresponding to an association between two objects. The association is undirected so (A,B) and (B,A) are equivalent.

directedPair
A directedPair value represents a pair of identifiers corresponding to a directed association between two objects. The two identifiers correspond to the source and destination objects.

duration
A duration value specifies a distance (in time) between two time points. In other words, a time period as defined by [ISO8601]. Durations are measured in seconds and may have a fractional part.

file
A file value is any sequence of octets (bytes) qualified by a content-type and an optional filename given to the file (for example, by the candidate when uploading it as part of an interaction). The content type of the file is one of the MIME types defined by [RFC2045].

uri
A URI value is a Uniform Resource Identifier as defined by [URI].

A special class used to create a mapping from a source set of any baseType to a single float. When mapping containers the result is the sum of the mapped values from the target set. See mapResponse for details.

Attribute : lowerBound [0..1]: float
The lower bound for the result of mapping a container. If unspecified there is no lower-bound.

Attribute : upperBound [0..1]: float
The upper bound for the result of mapping a container. If unspecified there is no upper-bound.

Attribute : defaultValue [1]: float = 0
The default value from the target set to be used when no explicit mapping for a source value is given.

Contains : mapEntry [1..*]
The map is defined by a set of mapEntries, each of which maps a single value from the source set onto a single float.

Attribute : mapKey [1]: value
The source value

Attribute : mappedValue [1]: float
The mapped value

Response variables are declared by response declarations and bound to interactions in the itemBody.

itemSession defines one built-in pre-bound response variable: duration.

Contains : correctResponse [0..1]
A response declaration may assign an optional correctResponse. This value may indicate the only possible value of the response variable to be considered correct or merely just a correct value. For responses that are being measured against a more complex scale than correct/incorrect this value should be set to the (or an) optimal value. Finally, for responses for which no such optimal value is defined the correctResponse must be omitted. If a delivery system supports the display of a solution then it should display the correct values of responses (where defined) to the candidate. When correct values are displayed they must be clearly distinguished from the candidate's own responses (which may be hidden completely if necessary).

Contains : mapping [0..1]
The mapping provides a mapping from the set of base values to a set of numeric values for the purposes of response processing. See mapResponse for information on how to use the mapping.

Contains : areaMapping [0..1]
The areaMapping, which may only be present in declarations of variables with baseType point, provides an alternative form of mapping which tests against areas of the coordinate space instead of mapping single values (i.e., single points).

Attribute : interpretation [0..1]: string
A human readable interpretation of the correct value.

Contains : value [1..*]

A special class used to create a mapping from a source set of point values to a target set of float values. When mapping containers the result is the sum of the mapped values from the target set. See mapResponsePoint for details. The attributes have the same meaning as the similarly named attributes on mapping.

Attribute : lowerBound [0..1]: float

Attribute : upperBound [0..1]: float

Attribute : defaultValue [1]: float = 0

Contains : areaMapEntry [1..*] {ordered}
The map is defined by a set of areaMapEntries, each of which maps an area of the coordinate space onto a single float. When mapping points each area is tested in turn, with those listed first taking priority in the case where areas overlap and a point falls in the intersection.

Attribute : shape [1]: shape
The shape of the area.

Attribute : coords [1]: coords
The size and position of the area, interpreted in conjunction with the shape.

Attribute : mappedValue [1]: float
The mapped value

At runtime, response variables are instantiated as part of an itemSession. Their values are always initialized to NULL (no value) regardless of whether or not a default value is given in the declaration. A response variable with a NULL value indicates that the candidate has not offered a response, either because they have not attempted the item at all or because they have attempted it and chosen not to provide a response.

If a default value has been provided for a response variable then the variable is set to this value at the start of the first attempt. If the candidate never attempts the item, in other words, the itemSession passes straight from the initial state to the closed state without going through the interacting state, then the response variable remains NULL and the default value is never used.

Implementors of Delivery Engine's should take care when implementing user interfaces for items with default response variable values. If the associated interaction is left in the default state (i.e., representing the default value) then it is important that the system is confident that the candidate intended to submit this value and has not simply failed to notice that a default has been provided. This is especially true if the candidate's attempt ended due to some external event, such as running out of time. The techniques required to distinguish between these cases are an issue for user interface design and are therefore out of scope for this specification.

Outcome variables are declared by outcome declarations. Their value is set either from a default given in the declaration itself or by a responseRule during responseProcessing.

itemSession defines one built-in outcome variable: completionStatus.

Attribute : interpretation [0..1]: string
A human interpretation of the variable's value.

Attribute : longInterpretation [0..1]: uri
An optional link to an extended interpretation of the outcome variable's value.

Attribute : normalMaximum [0..1]: float
The normalMaximum attribute optionally defines the maximum magnitude of numeric outcome variables, it must be a positive value. If given, the outcome's value can be divided by normalMaximum and then truncated (if necessary) to obtain a normalized score in the range [-1.0,1.0]. normalMaximum has no affect on responseProcessing or the values that the outcome variable itself can take.

Outcome variables are instantiated as part of an itemSession. Their values may be initialized with a default value and/or set during responseProcessing. If no default value is given in the declaration then the outcome variable is initialized to NULL unless the outcome is of a numeric type (integer or float) in which case it is initialized to 0.

For Non-adaptive Items, the values of the outcome variables are reset to their default values prior to each invocation of responseProcessing. For Adaptive Items the outcome variables retain the values that were assigned to them during the previous invocation of response processing. For more information, see Response Processing.

Contains : block [*]

The item body contains the text, graphics, media objects and interactions that describe the item's content and information about how it is structured. The body is presented by combining it with stylesheet information, either explicitly or implicitly using the default style rules of the delivery or authoring system.

The body must be presented to the candidate when the associated itemSession is in the interacting state. In this state, the candidate must be able to interact with each of the visible interactions and therefore set or update the values of the associated responseVariables. The body may be presented to the candidate when the item session is in the closed or review state. In these states, although the candidate's responses should be visible, the interactions must be disabled so as to prevent the candidate from setting or updating the values of the associated response variables. Finally, the body may be presented to the candidate in the solution state, in which case the correct values of the response variables must be visible and the associated interactions disabled.

The root class of all content objects in the item content model is the bodyElement. It defines a number of attributes that are common to all elements of the content model.

Attribute : id [0..1]: identifier
The id of a body element must be unique within the item.

Attribute : class [*]: styleclass
Classes can be assigned to individual body elements. Multiple class names can be given. These class names identify the element as being a member of the listed classes. Membership of a class can be used by authoring systems to distinguish between content objects that are not differentiated by this specification. Typically, this information is used to apply different formatting based on definitions in an associated stylesheet.

Attribute : lang [0..1]: language
The main language of the element. This attribute is optional and will usually be inherited from the enclosing element.

Attribute : label [0..1]: string256
The label attribute provides authoring systems with a mechanism for labeling elements of the content model with application specific data. If an item uses labels then values for the associated toolName and toolVersion attributes must also be provided.

Elements that can appear within an object.

Elements that behave as spans of text, such as the contents of paragraphs.

Elements that provide structure to the text, such as paragraphs, tables etc. Most elements are either inline or block elements.

Elements that can appear inside list items, table cells, etc. which includes block-type and inline-type elements.

A sub-class of inline that excludes interactions.

A sub-class of block that excludes interactions.

A sub-class of flow that excludes interactions.

Contains : inline [*]

Contains : block [*]

Contains : inline [*]

A text run is simply a run of characters. Unlike all other elements in the content model it is not a sub-class of bodyElement. To assign attributes to a run of text you must use the span element instead.

Note that the title attribute defined by XHTML is not supported.

Note that the title attribute defined by XHTML is not supported.

Attribute : cite [0..1]: uri

Contains : flow [*]

Although pre inherits from atomicBlock it must not contain, either directly or indirectly, any of the following objects: img, object, big, small, sub, sup.

Attribute : cite [0..1]: uri

Contains : dlElement [*]

Contains : inline [*]

Contains : flow [*]

Contains : li [*]

Contains : li [*]

Contains : flow [*]

Contains : objectFlow [*]

Attribute : data [1]: string
The data attribute provides a URI for locating the data associated with the object.

Attribute : type [1]: mimeType

Attribute : width [0..1]: length

Attribute : height [0..1]: length

Attribute : name [1]: string
The name of the parameter, as interpreted by the object.

Attribute : value [1]: string
The value to pass to the object for the named parameter. This value is subject to template variable expansion. If the value is the name of a template variable that was declared with the paramVariable set to true then the template variable's value is passed to the object as the value for the given parameter.

When expanding a template variable as a parameter value, types other than identifiers, strings and uris must be converted to strings. Numeric types are converted to strings using the "%i" or "%G" formats as appropriate (see printedVariable for a discussion of numeric formatting). Values of base-type boolean are expanded to one of the strings "true" or "false". Values of base-type point are expanded to two space-separated integers in the order horizontal coordinate, vertical coordinate, using "%i" format. Values of base-type pair and directedPair are converted to a string consisting of the two identifiers, space separated. Values of base-type duration are converted using "%G" format. Values of base-type file cannot be used in parameter expansion.

If the valuetype is REF the template variable must be of base-type uri.

Attribute : valuetype [1]: paramType = DATA
This specification supports the use of DATA and REF but not OBJECT.

Attribute : type [0..1]: mimeType
Used to provide a type for values valuetype REF.

DATA

REF

Contains : inline [*]

Contains : col [*]

Attribute : summary [0..1]: string

Contains : caption [0..1]

Contains : col [*]
If a table directly contains a col then it must not contain any colgroup elements.

Contains : colgroup [*]
If a table contains a colgroup it must not directly contain any col elements.

Contains : thead [0..1]

Contains : tfoot [0..1]

Contains : tbody [1..*]

In XHTML, table cells are represented by either th or td and these share the following attributes and content model:

Attribute : headers [*]: identifier

Attribute : scope [0..1]: tableCellScope

Attribute : abbr [0..1]: string

Attribute : axis [0..1]: string

Attribute : rowspan [0..1]: integer

Attribute : colspan [0..1]: integer

Contains : flow [*]

row

col

rowgroup

colgroup

Contains : tr [1..*]

Contains : th [1..*]

Contains : tr [1..*]

Contains : tableCell [1..*]

Attribute : src [1]: uri

Attribute : alt [1]: string256

Attribute : longdesc [0..1]: uri

Attribute : height [0..1]: length

Attribute : width [0..1]: length

Although a inherits from simpleInline it must not contain, either directly or indirectly, another a.

Attribute : href [1]: uri

Attribute : type [0..1]: mimeType

The math class is defined externally by the MathML specification. It can behave in the item's content model as an inline, block or flow element.

Attribute : outcomeIdentifier [1]: identifier
The identifier of an outcome variable that must have a base-type of identifier and be of either single or multiple cardinality. The visibility of the feedbackElement is controlled by assigning a value (or values) to this outcome variable during responseProcessing.

Attribute : showHide [1]: showHide = show
The showHide attribute determines how the visibility of the feedbackElement is controlled. If set to show then the feedback is hidden by default and shown only if the associated outcome variable matches, or contains, the value of the identifier attribute. If set to hide then the feedback is shown by default and hidden if the associated outcome variable matches, or contains, the value of the identifier attribute.

Attribute : identifier [1]: identifier
The identifier that determines the visibility of the feedback in conjunction with the showHide attribute.

A feedback element that forms part of a Non-adaptive Item must not contain an interaction object, either directly or indirectly.

When an interaction is contained in a hidden feedback element it must also be hidden. The candidate must not be able to set or update the value of the associated responseVariable.

show

hide

Attribute : vi