He believes it shows positive feedback has no impact upon the quality and quantity of participation, negative feedback encourages users to post more frequently at lower quality.
I'm not sure that's precisely what the data shows.
I think it shows the following:
1) Downvotes (negative feedback).
Users whose posts receive downvotes post more frequently, but their posts receive a greater number of downvotes in the future. They are also more likely to give downvotes and stay in the community for a slightly longer period of time than average and positive feedback groups.
2) Upvotes (positive feedback).
Users that receive upvotes also post more frequently, but their posts do not receive a greater quantity of upvotes in the future.
3) Average users
This is the most interesting area. Users which receive neither positive or negative feedback post far less frequently and are most likely to leave the community.
The definitions of 'quality' (number of positive votes) and feedback (quantity of posts) are open to challenge. A study on the quantity and quality of comments received would be far more interesting.
For two simple action items, positive voting has no major visible negative impact upon future participation. Negative voting is likely to encourage negative behavior. Posts that receive no visible feedback are likely to drive members away.