Dear Mr. Dad: With all the talk of bullies and school violence, isn't there a way to predict which kids will grow up to be criminals and which won't?

Answer: Sounds like an interesting idea, but if you ever saw the Tom Cruise movie "Minority Report," you know the potential pitfalls of punishing people for crimes they haven't committed yet. What if the predictions are wrong? "Minority Report" was set in 2054, but I just read two different studies that make it look like the future is already here. Both purport to be able to ferret out those bad seeds before they turn aggressive, violent or commit crimes. Personally, I'm scared. And you should be too.

In the first study, Drew Barzman, a child psychiatrist at Cincinnati Children's Hospital, and his team were looking for a way to predict violence among pediatric psychiatric patients. They developed a simple test that measures levels of cortisol, testosterone, and dehydroepiandrosterone (DHEA) in boys' saliva. The test proved to be quite accurate.

"We believe salivary hormone testing has the potential to help doctors monitor which treatments are working best for their patients," Barzman said. "Eventually, we hope this testing might also provide a tool to help improve safety in schools."

I'm OK with the part about helping doctors, but how can they possibly use the saliva test to improve safety in schools? I can just imagine every little boy being required to spit on a card. And based on the saliva analysis, some will be expelled. Or maybe even sent to jail for some future crime that hasn't happened yet - or might not happen at all.

In the second study, a team of researchers led by Dr. Sheilagh Hodgins of Universite de Montreal, were trying to find out whether childhood behavior could predict future crimes. So they compared teacher-rated conduct problems (CP) and hurtful and uncaring behaviors (HUB) at age 6 and 10 with criminal convictions up to age 24. And, sure enough, they found a connection.

Six-year-old boys who were rated by their teachers as having the highest degree of CP (things like blaming other, bullying, destroying property, disobedience, fighting, lack of empathy, lying, and stealing) and HUB were four times more likely to be convicted of violent crimes and five times more likely to be convicted of nonviolent crimes than boys with lower ratings.

Pretty much the same was true for the girls: 6-year-old females with the highest CP and HUB ratings were five times more likely than girls with lower ratings to have a conviction for nonviolent crimes by age 24.

According to Hodgins, the goal of the study was to better understand how to prevent future crime and "reduce the human and economic costs associated with criminal activity."

Quite noble, I think. But there's more.

"If their conduct problems could be identified and reduced early in life, this would potentially allow these children to alter their developmental trajectories, live healthy and happy lives, and to make positive rather than negative contributions to our society."

Whoa there. Alter their trajectories? How? Are we going to lock up a bunch of first graders? Does anyone else think that 6 years old is a little early to be labeling kids as violent, or criminal, or even likely to misbehave? Especially when basing "interventions" on the assessment of classroom teachers, most of whom aren't trained child psychologists?

Sure, being able to predict kids' future bad behavior sounds kind of cool, but the potential for abuse makes it not worth the risk.

(Read Armin's blog at DadSoup.com, send email to armin@mrdad.com, and follow him on Twitter @mrdad.)

• Read more articles by ARMIN BROTT

Order Reprint