Two people—HaaTa and Matt3o—are building force measuring machines. These are the only people who can reliably answer this question, because the difference is too subtle to determine solely by feel. This is really one for jacobolus, as he has an Alps switch collection and (possibly only limited) access to HaaTa's force gauge.
My feeling is that the overall cause is cost cutting, to improve competitiveness with both prestige manufacturers (e.g. Cherry and Futaba) and with the clones. There were a lot of ugly Alps switches that were probably made at lower cost in their overseas subsidiaries (Forward and Gold Star Alps) and this might have had some effect on the overall design, such as the still unknown reason for slit removal around 1993.
However, when handling a switch, it's impossible to reliably gauge change in feel, but if you can measure the force in a machine, you'll get semi-reliable graphs. Graph reliability will be improved with sample size (as Alps switches degrade badly and vary hugely between keyboards, especially white Alps), and weather variation (to address the theory that Alps switches work better in humid conditions). The humidity issue is an interesting one as I've personally noticed distinct variation in feel in my Dell AT102W with hybrid blue Alps (which I very rarely use), but as my human perception is subject to distortion, only a machine is able to give me trustworthy data.
Changes to both plastic and metal parts sounds reasonable. It might not just be cost cutting — it may have also been done to reduce manufacturing time as demands grew.