Building models of correlation/causality
hoe you interpret the results is important
until data archives started being formed, old-boy network got people access to data
data archives made it much more democratic, so others can reuse.
TRAC $40k
at the institution level
seals of approval are en vogue now, but not really helpful as indicators for succession, etc.
ICPSR
levels of repositories: when sthg is the reference repository, then they get to set the standards for the field
multiple points of discovery and multiple points of access
return used to be to get access to the data sets, but now return on the fee to ICPSR is the curation
paid for as an infrastructure resource, not by the library (social sci dept)