The elephant in the record: On the multiplicity of data recording work