[DDI-SRG] Map 2.1 to 3.0

[Wendy Thomas]

Pascal

If A B C etc is a true text response and not a coding scheme then this has 
broader application than I considered. I would like to point out that 
other than rendering this as a discrete/continuous option intellectual 
judgement will be required.  Once again note that this attribute had a 
default setting which means that the content may not necessarily be 
indicative of anything more than a users putting nothing in this 
attribute.

If this is truely a valid attribute for multiple types of representations 
then it should be added to the abstract type Representation....but as 
what??? If it means different things for different representations then 
these need to be spelled out. I think that we have more detailed 
information in DDI 3.0 than was available in 2.1.

So, to inform further discussion, please provide an example of discrete 
and continuous for Numeric, Text, DataTime, and CodeScheme. I think 
Numeric and CodeScheme already handle this in a more detailed way, but as 
I said, if you have a use case for valid use in the other represenations 
and you can define what a) discrete / continuous means in each case, and 
b) it is consistant for all representation types, then we should add it to 
Representation as an optional attribute.

We are committed to carrying forward valid and accurate information from 
2.1, but I think we have an obligation to modify fields that provided 
erroneous or misleading information because they were designed for one 
type of data and ended up getting used for everything (accruately or not).

Wendy




On Fri, 18 Jan 2008, Pascal Heus wrote:

> Wendy:
> This may be worth further discussion. See my last note in bug
> http://mantis.ddialliance.org/view.php?id=106
> *P
>
> Wendy Thomas wrote:
>> Interval
>>
>> Interval as used in DDI 2.1 was an indication of whether the variable was
>> discrete or continuous. This could be in relationship to a numeric value
>> or a coded value. In DDI 3.O the NumericRepresentation provides much more
>> detailed NumericCodeType that indicates that a the value is a an integer
>> (thus removing the continuous intervening values) and optionally provides
>> an incremental value so that one could provide a step increase of greater
>> than one for an integer NumericDataType or a .5 or other value for a real
>> number. For CodeScheme values the description of Level provides a
>> relationship type (nominal, ordinal, interval) and an optional Interval to
>> provide the actual interval between values of the categories.
>>
>> Therefore the proper mapping of 2.1 to 3.0 is dependent upon whether the
>> variable has a numeric or codescheme representation.
>>
>> IF NUMERIC:
>>
>> intvl maps to Variable/NumericRepresentation/NumericDataType where
>> discrete = integer and continuous = real with the option of providing the
>> exact interval step in interval
>>
>> IF CODESCHEME:
>>
>> intvl maps to CodeScheme/Level/RelationshipType = interval and a
>> description of the interval between the values of the categories (however
>> the codeScheme provides the only valid values for the code)
>>
>>
>> The issue with DDI 2.1 is that intvrl has a default value of discrete
>> whether or not it made sense (in other words an open text variable would
>> have an interval value of discrete).
>>
>> The mapping will will be changed to reflect the complexity of this
>> translation.
>>
>> Wendy
>>
>> Wendy L. Thomas                          Phone: +1 612.624.4389
>> Data Access Core Director		 Fax:   +1 612.626.8375
>> Minnesota Population Center              Email: wlt at pop.umn.edu
>> University of Minnesota
>> 50 Willey Hall
>> 225 19th Avenue South
>> Minneapolis, MN 55455
>> _______________________________________________
>> DDI-SRG mailing list
>> DDI-SRG at icpsr.umich.edu
>> http://www.icpsr.umich.edu/mailman/listinfo/ddi-srg
>>
>>
>
> _______________________________________________
> DDI-SRG mailing list
> DDI-SRG at icpsr.umich.edu
> http://www.icpsr.umich.edu/mailman/listinfo/ddi-srg
>

Wendy L. Thomas                          Phone: +1 612.624.4389
Data Access Core Director		 Fax:   +1 612.626.8375
Minnesota Population Center              Email: wlt at pop.umn.edu
University of Minnesota
50 Willey Hall
225 19th Avenue South
Minneapolis, MN 55455