Thanks very much to Kay Shimizu and Mike Smitka for their illuminating responses. Here's a follow-up
question: for those of us who are not specialists in this topic but find it important for both our own research and for teaching to get a 'ballpark figure'
for the proportion of employees employed at large firms, what is the most reliable way to discover such a figure? I personally tend to use the Roodooryoku choosa for this, because it seems to me to be reliable as far as I can tell, has been used by some researchers I hold in regard in the past, and allows historical comparison of change over time, but is there a better approach?
(It is possible to look at literature on the topic too, but different authors may use different databases - often without explaining or justifying their choice or making reference to the work of other authors who make different choices - and of course such literature goes out of date quite fast.)

I realise, of course, that there can be various definitions of what is considered a 'large firm', and that the 500+ employees figure used by the Roodooryoku choosa is on the high side (300+ often being used instead, for instance).

Peter

Peter Cave
Lecturer in Japanese Studies
SALC, University of Manchester