Good info (with the actual data) here:
http://www.guncite.com/gun_control_gcdgaga.html
The "more likely to be used against you" thing is essentially a media distortion of Arthur Kellermann and Don Reay, "Protection or peril? An analysis of firearms related deaths in the home,"
New Engl J Med 1986 (314:1557-60). Kellerman has rehashed the study in several iterations since, without significantly changing his methodology.
Contrary to media myth, Kellerman did not find ANY instances of the homeowner's gun being used to kill the homeowner by a criminal, AFAIK. I believe that in every case in which Kellerman cited a gun-owning homeowner being shot, he/she was shot by a criminal's gun brought into the house from outside (the homeowner probably being unarmed at the time; Kellerman only looked at whether a gun was present in the home, not whether the owner tried to use it defensively).
The VAST majority of the deaths Kellerman et al cite are suicides, not homicides or accidents. If you are not at risk for suicide, there goes most of the alleged risk. Accidental gun deaths are statistically insignificant, so much so that I doubt he even recorded any (current rate is ~600/year nationwide for 80+ million gun owners, and this figure does not exclude accidents by gun-owning criminals, a fact which tends to skew the perceived risk higher).
Of homicides in which the victim was the gun owner, Kellerman counted these as if the homeowner's gun were at fault, but in every instance the gun used was the criminal's gun, brought into the house by the criminal with lethal intent. Kellerman didn't bother to control for the fact that people who are in greater danger of being murdered are more likely to purchase a gun than those in the control group, nor did he control for understatement of gun ownership by the control group (which given the area and political climate almost certainly would have skewed his results).
The commonly cited "43 times more likely to kill a family member than defend against an intruder" statistic from Kellerman et al excluded ALL defensive gun uses that didn't result in the death of the criminal, which acts systemically to hugely underestimate the number of actual defensive uses. (Surveys of self-defense incidents imply that 98% or more do NOT result in the wounding of the criminal, much less his death). Using this methodology, if the intended victim pulled a gun and the criminal fled, it didn't count; if the victim fired a warning shot and the criminal fled, it didn't count; if the victim shot the criminal, halting the attack, but the criminal survived, it didn't count; and if the criminal were shot and killed, and was known to the victim, IIRC it was counted in the "shooting a friend or family member" category. Not exactly an objective study. Using this methodology, one could prove pretty much anything she/he wants.
Certainly one can quibble about the Lott et al data, and the Gary Kleck et al data on defensive gun uses, but if you exclude suicides I don't think there is any data anywhere that suggests a gun in the home is a significant danger to you or your family.
And FWIW, the National Crime Victimization Survey showed that victims who defended themselves with firearms were the least likely to be injured in the course of the crime (even less than those who offered no resistance at all), and all victims who used guns and were injured were injured BEFORE they accessed the gun. Unfortunately, the sample size was VERY small, so that data is not as robust as it might be with a larger data set.