http://www.ined.fr/fichier/t_telechargement/6575/telechargement_fichier_fr_classeur1.xls

Together with poll data:

Age     SR  NS   share of population in %

18-24   53  47   11.2

25-34   54  46   16.3

35-49   56  44   26.8

50-64   51  49   23.9

65-+    25  75   21.8

NS wins only the 65+ part of the population, quite impressive.

Remultiplying share of population and SR vs NS choice gives a total of 47.385 vs 52.615 which is coherent with the overall result, may be there's not much difference in participation by age (to be checked, could be cancellation too).

I note that from 22Apr2007 to 27Apr2007, SR did +1.5 (to 47.5), NS -1.5 (to 52.5) and no opinion went from 11% to 9%.

If this is trend and not noise at 5% no opinion SR wins 50.5-49.5 :).