How to deal with scientists who cheat

 
What do you do when a scientists (PI, post-doc, graduate student) is caught falsifying data? Should they be expelled from the community, fired from their job, or given a slap on the wrist and rehabilitated?

This isn't an easy question as Janet Stemwedel demonstrates in Tempering justice with mercy: the question of youthful offenders in the tribe of science. I hate it when she does that. It would be so easy to conclude that cheating scientists should be drummed out of the profession but then along comes Janet to confuse me.

She's right, of course. There ought to be a range of punishments that fit the wide range of crimes and motives.

---