As a respiratory therapist, I only have a few comments [start snickering now...]:
1) Absolutely. Burning ANYTHING and inhaling the smoke does harm to the lungs. To my knowledge, there have been no studies comparing the use of a vaporizer to other methods of inhalation or ingestion (but I'm a little skittish about using the very powerful search tools at the hospital. I'd rather not have to explain in somebody's office why I was looking up papers on marijuana on work time.). That said, the use of a vaporizer seems to be at least theoretically MUCH better. Burning something not only vaporizes some of the the THC but creates an aerosol of smoke particles. A vaporizer (at least a good one) should only produce molecular gas. The particles are the problems, not the THC.
2) There have been some well-performed studies that do seem to show a negative effect on cognition and short-term memory for long-term, chronic users. The big problem with every study I've ever read is that -- because of the federal scheduling of marijuana -- it is very difficult to get ANY study approved. It's very hard to show a treatment effect or composite effects when your sample size is severely limited. At best, you can show individual effects. Just look at the studies required to get a drug past the FDA to market. It takes years of rigorous work with enormous sample sizes to demonstrate safety and efficacy and we STILL find drugs after-the-fact that shouldn't have gone to market. Compare that to a study of marijuana with 30 people in it and you can see why it's so hard to get anywhere with the research at the professional level.
3) Agreed that documentaries -- while useful -- are not peer-reviewed research. I think what they do provide is a different perspective from which to begin a search for more information.
My biggest complaint with people citing studies (not just on marijuana but even just at my normal job) is often they cannot tell you anything about the actual research -- what was the authors' hypothesis? What methods were used? Do the methods described actually test the hypothesis? What were the conclusions the authors drew and do the methods and results actually support that conclusion or are the authors taking some license? What were the stated and/or implied limitations of the study? Do the mathematics test what the authors state? What other statistical test could be used to check the results? Not many people have the academic wherewithal to critique original research. People who can't tell you difference between an hypothesis and a theory or what a chi-square test means and tests feel fully qualified to judge the merits of a professionally peer-reviewed journal article. At best, they read the authors' conclusion (or worse, what someone else wrote about the study in some news article) and take what like and ignore what they either don't like or don't understand. Outside of work I have met damn few people who can really read and critique a scientific paper.
Marijuana is far safer than alcohol or tobacco and any argument that applies to keeping marijuana illegal also applies to either or both of those. As was pointed out, it is impossible to ingest enough THC to overdose -- you can kill yourself the first time you drink. Any marijuana smoker who smoked the equivalent of a pack-a-day of cigarrettes would be baked out of their mind 24/7 and likely be non-functional for the vast majority of their day. It's pretty hard to maintain the kind of smoking habit in marijuana that leads to smoking-related illness from tobacco (typically, we start really expecting to see smoking-related illness at about 20 pack-years. Pack-years are the number of packs-per-day times the number of years smoked; pack a day for 20 years, half a pack for 40 years, etc... 20 pack-years of marijuana is a shit-ton of weed.).
If alcohol is legal; if tobacco is legal; if vicodin is legal; if prozac is legal; if valium is legal then there is no reasonable argument for keeping marijuana illegal.