Convicted wrongly then commit a crime in prison...

To clear up the confusion over the double jeopardy issue, the law says a person can’t be tried for the same crime twice. But if you were tried once for murdering someone in 1984 and then again for murdering that same person in 2004, you would be getting tried for two different crimes.

Hm. Now, as I understand the law, it is possible to get a new trial without there being reversible error if it can be shown based on new evidence that a reasonable jury would have delivered a “not guilty” verdict. I will freely admit this is based on some extremely cursory research done a few years ago after seeing the scenario play out on a two-part episode of The Practice so corrections and/or confirmation would be ducky.

I’m not familiar with this finding, verdict or whatever it is but wouldn’t any decent defense counsel make sure the jury knew about it if it could get their client off? Is the defense sometimes prohibited from letting a jury know about things they are allowed to take into consideration?

Jury nullification
The concept is an old one in The United States, in fact pre-dating the United States. It is fairly controversial, with people on either side of the issue becoming extremely passionate about it.

In a nutshell, a “not guilty” verdict from a jury is final because of the double jeopardy clause of the Constitution. Juries are charged to consider the evidence placed before it and the law as instructed by the judge in reaching a verdict. There is, however, no real way of stopping jurors from deciding in the course of its deliberation that even if there is no question that the defendant did the act for which he’s on trial, they don’t want to convict him for whatever reason. Maybe jurors disagree with the law, maybe they like the defendant or think he’s a good guy, maybe the potential punishment is more than they think he deserves, maybe (as I understand happened in the South) the defendant is white, the victim black, and an all-white jury acquits. Regardless, if the jury comes back with a “not guilty” then there’s nothing the judge or the prosecutor can do (absent something like, as has been discussed briefly, a prosecution in federal court after a state trial). Thus, people who think juries should be nothing other than finders of fact tend not to like jury nullification as a concept and those who think juries should be informed they have the power to nullify think it should be part of the formal jury instruction.

Personally I lean toward the “juries are finders of fact and it’s the judge’s, not the jury’s, role to rule on matters of law” camp, but as this isn’t GD I won’t go any further with that. I’m sure a quick spin through GD will turn up any number of threads on the topic.

I currently live in a common law state so I’m really interested in this aspect of trial law that I’ve never heard of before.

If a juror tells his fellows of the principle and a skeptical juror takes his concerns to the judge, can the judge declare a mistrial or have that juror dismissed?

It seems that doing anything to stop juries knowing about this right to deliver a verdict on the law rather than the accused is an infringement of his right to the best possible defense.

Well, see, that’s one of the big questions about jury nullification (perhaps the big question). Is it a “right”?

Also, defendants aren’t entitled to the “best possible defense.” They are entitled to the effective assistance of counsel.

I suggest Googling and/or searching GQ or GD for previous threads on the topic.

In U.S. law does the accused-appellant have to show that the jury would have delivered a not guilty verdict? or simply that it is not possible to say with certainty that the jury would have convicted, if they had had the new evidence?

Canadian criminal law uses the second test - the accused-appellant doesn’t have to show to a certainty that the jury would have acquitted; simply that the new evidence is signficant enough that it is not possible to say with certainty that the result would have been the same. It’s an extension of the “reasonable doubt” approach - the accused doesn’t have to prove that there would have been an acquittal with the new evidence. The onus is on the Crown to defend the conviction, even with the new evidence. Otherwise the appellate court will set the conviction aside as unsafe.

Note that if the appellate court does set aside the conviction, it doesn’t mean the accused is scot free. The court sends the matter back for a re-trial by another jury, since it’s not the function of the appellate court to judge the evidence.

Bear in mind that my knowledge of the topic is limited to a two-part episode of a TV show which has been known to be less than accurate regarding matters legal…

The scenario of the show was that a man was about to be executed for the rape and murder of his girlfriend and the murder of her mother several years earlier. The conviction was obtained on the basis of a supposed confession recounted by a jailhouse snitch, one witness statement and a blood type match on semen from a rape kit. This was before DNA testing. The defense wanted to do a DNA test on the rape kit to attempt to disprove that it was the defendant’s semen. The appeals court refused, stating that the test could only be done if the defense could show that a reasonable jury with the test results would have come to a different verdict. Not that the likelihood was high that they would, but that they would have acquited. In other words, they had to prove a jury would have found him innocent before getting a test which might prove his innocence. The defense ended up discrediting the snitch and the witness and they found and tested semen found on the nightgown of the victim which was buried with her in accordance with Jewish tradition which didn’t match the defendant and finally after all of that and in the face of continued opposition from the prosecution, the original trial judge vacated the guilty verdict and (presumably because of double jeopardy which brings the thread full circle) set the defendant free.

Very dramatic, but unknown how accurate from a legal standpoint.