An Archives Perspective to DDI 3 Mari Kleemola

An Archive’s Perspective to DDI 3 Mari Kleemola Information Services Manager Finnish Social Science Data Archive CESSDA Expert Seminar Ljubljana, 9. 11. 2009

My Starting Points • Need to evaluate DDI 3 for FSD’s purposes – plans to change our metadata system – need to produce various kinds of metadata • • DDI 2 for long-term preservation (probably) DDI 3 for Cessda (but no decisions made yet) ? for KDK (The National Digital Library project) … • CESSDA PPP WP 8 task 8. 2. 2

Things to look at… • • • Preservation Multilinguality Groups Reusables, referencing Comparison FSD’s current practises/processes

Test case: ISSP 2006 Finnish data • FSD the (primary) financier of data collection and involved in instrument translation etc. • Collected by Statistics Finland • Two questionnaires: fi and sv, and documentation in Finnish and in English • A single study that belongs to several groups/series

ISSP multinational series Finnish ISSP studies FSD 2410 FSD 2330 ISSP 2006 multinational file ISSP GER 2006 ISSP SWE 2006 ISSP FIN 2006 FSD 2248 Additional questions FSD 2133 FSD 2039 Q. in Finnish Q. in Swedish

Subset: variables documented Variables/Questions (see PPP document DR 8. 2. 2 Evaluation of DDI 3. 0) • Archive and ID variables - Technical variables – FSD specific • Substantial variables - types – Simple question: Variable V 4 : Q 1: Obey laws without exception – Question with items (little battery): • • V 5 Q 2 a: Public protest meetings V 6 Q 2 b: Protest demonstrations V 7 Q 2 c: National anti-government strike ISSP demographics and interview char. - types – SEX: R: Sex • Country specific Question / Variables – Ctry. specific: degree – Country specific education: Finland • In addition the following (for FSD’s test purposes): – (K 2) Year of birth (integer) - the ISSP variable AGE is calculated from this variable – (K 20) WRKSUP : R: Supervises others at work (a filter variable) – (K 21) Is R responsible for… (a battery with three subitems and a different universe due to the filter K 20, not in the basic ISSP 2006 questionnaire) – (K 22) WRKTYPE : R: Workg f priv. , pub sector, selfempl. • Total: 18 variables

Tools I tried… • De. Xtris • CESSDA Core • DDI 3 Editor-Lite All gave valuable insight to DDI 3… … but decided to start from scratch – with the help of Oxygen, Dagstuhl 2008 materials, DDI Alliance website (Best Practices, ”table”, Cessda core, …), DDI 3 Help Center, SPSSOMS 12 to. DDI 30, PPP WP 8 paper, and our existing DDI 2. 1 file

Main problems • Almost zero knowledge of XML schemas • Lack of time • Lack of supporting resources (ie. technical staff) • The complexity of DDI 3

Working with ”Simple questionnaire example”

Results of Exercise • DDI 3. 0 XML file – It’s valid, it’s got 3771 lines, but what to do with it? • Managed to include all our DDI 2. 1 metadata • Need tools to produce DDI 3 metadata • DDI 3 is very flexible, so think carefully – what is reusable (create resource packages) – what can be inherited (use groups) • Note: the way your’e going to use the metadata affects the way you group • Probably best to use DDI 2. 1 -type documentation for preservation • FSD need DDI 3 functionalities, but it probably won’t pay off to DDI 3 all our studies • More questions than answers (many answered at Dagstuhl)

To Do • To research more carefully – Grouping – Inheritance – Comparison module – Language issues • Input from FSD technical staff needed • Plan FSD actions to move from DDI 2 to DDI 3 • European co-operation (CESSDA)

Thank you!