A few days ago I mentioned the UNT MARC study of field usage in MARC records. It seems a similar study was done in 1997 in Germany. They found 33 elements that appeared in more than 1% of the records. Only 5 fields were used in 100% of the records (245, 260, 300, 050, 008). The study was of over 4 million records from LC, the UNT study is using a much larger sample and breaking it down to the subfields. Still using the Walt Crawford, study in 1988, the German study in 1997 and the UNT study will give us a view over time. Someone strong in statistics could combine the studies to do a meta-analysis. The German study is in German, naturally, but since MARC uses numeric tags, even those of us not able to read German can grasp some of the results.