2023-10-27 Langmann 1 Response to Stan Kutcher On October 25, 2023, Stan Kutcher posed a question to me during the SECD Senate of Canada Hearing on C-21. The question appeared purposely designed to take up all his allotted time and not allow me to respond. Stan Kutcher had not reached out to me prior to this time to discuss any of the issues he raised. In order to prevent this from appearing to be a form of academic cowardice, I have asked him to reach out to me. At this time, I have not heard from him or seen any document he purports to have.

In the following document I will address some of his questions as obtained from the Senate transcript. The transcript has been included below in italics. Some of the points he raised are intelligent but the last comments about the figures in my 2020 paper are to put it politely as possible, lacking in any intelligence or understanding of concepts of time series analysis – a common method in the social sciences (Langmann 2020). From the transcript: Senator Kutcher: This question is to Dr. Langmann. I apologize that it’s quite long and complicated. I’m going to read it because it’s about precision and research.

As you are aware, a recent Ontario study has shown that two thirds of firearm- related deaths are from suicide and that 70% of all gun-related deaths Canada are suicides. Clearly, this is not from criminal activity. From 2016 to 2020, about 2,700 men in Canada died from firearm suicide. A prospective U.S. study of 26 million people showed that males who owned handguns were eight times more likely to die by suicide; females, 35 times more likely. Public health axiom is having access to lethal means as the primary cause for death by suicide. Harvard University School of Public Health, in their

summary of evidence, disagrees with your testimony: States with more key firearm laws had fewer firearm homicides and fewer firearms suicides, after controlling for poverty, unemployment, education, race and non-firearm violence-related deaths. That’s like hanging, for example.

They noted that, “Some gun policy evaluations are designed to ensure that no effect will be found.” They also noted that the "results of a flawed study should not affect policy.” The Rand Institute report What Science Tells Us About the Effects of Gun Policies of January 2023 noted that different types of legislation had different types of effects. They stated that: The absence of evidence about a law can result from the law not having been studied or studied well. . . They further noted that poor methodology results in erroneous conclusions.

In your brief, you stated that your conclusions were “based on sound statistical analysis and information.”

2023-10-27 Langmann 2 I read your PLOS ONE paper, "Effect of firearms legislation on suicide and homicide in Canada from 1981 to 2016," which your testimony is based upon and, because I didn’t want to review it myself for the committee, I asked experts to review it for me. This is what they said in part, namely, that there were numerous problems with the methodology. I’ll just quote some of them. The application of the statistical difference in difference methodology is improperly applied. Another was that the parallel trends assumption that you use is not met in the dataset.

Another was that the stable unit treatment value assumption is also violated. They do point out — and senators will see this because I will circulate this information to them — in figure one in your paper, that you show a clear decrease in suicide by firearms for both males and females from 1980 to 2016. Also in your paper, in figure 2, you show a clear decrease in homicide by firearms for both females and males. They say: The conclusion that increased firearm legislation does not decrease suicide nor homicide is not supported by the evidence presented by the author.

They then say: The trends don’t lie, but the author engaged in complex analyses and questionable assumptions to build a model to arrive at his desired conclusion. Could you respond to these concerns? Question 1, Risk of Firearms in Homes The first key point to respond to is the question of risk. Do homes with firearms put people at more risk or less? There are some studies that examine for this, but they are generally cross sectional studies, which are the weakest epidemiological studies.

No doubt one has heard of studies that show eating a certain food is associated with less cancer, and then several years later an alternative study presents the exact opposite result. The reason for this is due to confounders and false positive and negative results. Confounders are other variables that might cause an effect that a researcher did not or was not able to control, or didn’t even know existed.

For instance, comparing different US states or countries in cross sectional studies about guns makes these studies very susceptible to false results because its hard to control for all the differences in each US state or country. Sometimes data is missing, and sometimes data is collected differently, and there are some things that are impossible to control for such as culture and social norms. Hence most social scientists value quasi experimental studies much more than cross sectional studies.

In health sciences one can study a drug and control for confounders by randomly putting patients into a group where they get the drug and a group where they get a placebo. This way all the other factors such as obesity and prior health conditions are averaged out randomly between each group and only the effects of the study drug are found. This is called a randomized controlled study or RCT. In social sciences it is very difficult to divide people into groups and keep them in a group. So instead, quasi experimental studies are the closest thing to a RCT available.

My studies are all quasi experimental studies and are the best form of study one can do in this area as noted by a recent Canadian Medical Association Journal review (Bennett, Karkada et al. 2022).

2023-10-27 Langmann 3 The question at hand is not whether homes with more firearms put people at risk, even though some studies show they don’t (Siegel, Pahn et al. 2019). The question at hand is will Bill C-21 reduce homicides or suicides? Bill C-21 has been covered elsewhere, but simply will prohibition of certain firearms, semi auto rifles and handguns, and red flag laws result in less deaths by firearms?

In the 1980s and 1990s a large number of automatic, semi automatic rifles, and handguns were banned, 550,000 in the 1990s, and as I have covered in my studies, and as other Canadian studies show, these laws will not have the desired effects of homicide or suicide reduction (Rich, Young et al. 1990, Mauser and Holmes 1992, Caron 2004, Cheung and Dewa 2005, Caron, Julien and Huang 2008, Langmann 2012, McPhedran and Mauser 2013, Langmann 2020, Langmann 2021, Langmann 2022, Langmann 2023). I have covered the Canadian evidence in depth in this

article including how the vast number of studies agree with my own when one extracts the best methods from those papers (Langmann 2023): Canada, Firearms Legislation, Suicide and Homicide. A Review of the Evidence by Caillin Langmann :: SSRN While a public health axiom is indeed, reducing access reduces harm, there does appear to be a substitution effect occurring in Canada where people unable to access firearms switch to alternative methods such as hanging. No one has published a study asking the question of whether the presence of rope in a home increases the risk of suicide.

Suffocation is the number one cause of death by suicide in Canada, but controlling rope is naturally impossible. With limited dollars available, and with the above evidence demonstrating no real benefit from these firearms laws, it is time to concentrate on other areas of harm reduction such as intervention in youth engaged in early criminal behaviour and mental health treatment and suicide prevention. Studies in other countries also demonstrate similar results.

Australia severely restricted semi automatic rifles and handguns and has seen no benefits in homicide or suicide reduction (Lee and Saurdi 2008, McPhedran and Baker 2008, Chapman, Alpers and Jones 2016, Gilmour, Wattanakamolkul and Sugai 2018). Recently a quasi experimental study in California showed that there was no associated reduction in violence (Pear, Wintemute et al. 2022). Question 2, Problems with Langmann 2020 Study First, some relevant background, my 2020 paper was peer reviewed by three academics and a statistical expert.

Not one was concerned about the “improper application of the difference in differences method”. Difference in differences is a quasi experimental method where a researcher constructs a treatment group and a control group from available data. In my case the treatment group is gun homicides and suicides and the control group is homicide or suicide from other methods. The theory is that gun laws should only have an affect on gun deaths while deaths from the other methods should not be affected.

Then one examines what is happening with the rates of death in the two groups before the gun laws and after the gun laws come into effect. The deaths by other methods should change at the same rate that has been occurring before the laws but the rates of gun deaths should decrease after the laws come into effect.

Of course other things that can affect suicide or homicide can be occurring at the same time such as socioeconomic factors like unemployment or policing or suicide prevention interventions or even new drugs, and the theory is that these would affect both gun deaths and non gun deaths the same way and thus the non gun death group is used as a control to conclude what the effect of the gun legislation was as it’s the only difference between the two groups.

The question of parallel trend assumption is a complex statistical condition, but this is answered in my study itself and it appears whoever reviewed it did not read the methods

section carefully. A generalized difference in difference model was constructed to relax the parallel trends assumption. In fact, for homicide the parallel trends assumption holds true, but for suicide the

2023-10-27 Langmann 4 generalized model was required. This is a similar model as used by previous studies and the details can be accessed for those who are inclined by reading my paper and also a paper by Gilmour et al. 2018 (Gilmour, Wattanakamolkul and Sugai 2018). The next complaint involves the stable unit treatment value assumption (SUTVA). This is actually dealt with in my paper, and is another difficult concept, however I will explain it and we can use some logic to see how it may bias results. In causal studies, SUTVA is the assumption that a treatment group will have no effect on a non treatment group.

In my studies it would apply in these two ways. In the homicide study, the assumption is that the gun laws stopping people from homicide by firearm won’t spill over into also stopping people from homicide by knives or other methods. In terms of the suicide study that the gun laws stopping people from suicide by firearm won’t spill over into stopping people from suicide by hanging or other methods. Alternatively, the gun laws could also spill over and cause more homicides or suicides by other methods. Any of these effects could bias the results.

Dealing with this with the limited data available from Statistics Canada is very hard. However, we can logically go through the potential biased results and how that would affect the studies findings. Thinking about the homicide study first, what my 2020 study showed was that homicide rates by firearm and non firearm before the gun laws was declining at the same rate. After the laws were passed the rates of homicide by firearm and non firearm declined at the similar rates and changes and therefore the conclusion is that gun laws did not have an effect.

However, what if the gun laws caused the rates of homicide by firearm to decrease and then spilled over and caused the rates of homicide by non firearm to also decrease by the same amount? The result would look like there was no difference between the two groups but in reality gun laws are saving lives in both groups! First think about it logically, homicides tend to be three different scenarios, someone killing their partner/family, someone killing for revenge, and then gang related violence.

Would someone not having a gun make them less likely to use another weapon to kill their family or someone out of revenge? Doesn’t seem plausible. In terms of gang violence this is where perhaps less guns would make the gangs less violent and then less likely to kill using other methods because they feel less threatened. Its possible, though not likely. Regardless, if such an effect was happening we should see less violence overall and a steeper overall rate of decline in homicide, but that’s not what we see.

Instead, we see after the legislation in the 1990 and late 2000s, the rates of homicide tend to flatten, in other words the decline in homicide prior to the legislation turns into less of a decline, not more – see figure 2 in my 2020 paper. It is unlikely that there is any spillover in this case. That being said, it is important to perform a number of different analyses to check a conclusion. My paper in 2012 used three other methods, time series, ARIMA, and Joinpoint, to check for an association between firearms laws and homicide and also found no association (Langmann 2012).

Considering suicide, my study showed that gun laws was associated with a decrease in gun suicide but an increase in non gun mediated suicide. Violation of SUTVA would mean that the gun laws that prevented people from using guns to kill themselves, also prevented them from using other methods. If this was occurring, we would see the other methods also decrease. We do not. In fact we see the other methods such as hanging increasing.

Question 3: Figure 1 and 2 in Langmann 2020 Study This is the part of the critique where it makes we wonder if the person critiquing this study actually understands anything about social sciences. This is quite literally, the most ludicrous part of the critique by Stan Kutcher and it is surprising because its something a lay person

2023-10-27 Langmann 5 would state. Figure 1 and 2 of my study are included below for reference. Once again this is Stan Kutcher’s statement: They do point out — and senators will see this because I will circulate this information to them — in figure one in your paper, that you show a clear decrease in suicide by firearms for both males and females from 1980 to 2016. Also in your paper, in figure 2, you show a clear decrease in homicide by firearms for both females and males.

They say: The conclusion that increased firearm legislation does not decrease suicide nor homicide is not supported by the evidence presented by the author. They then say: The trends don’t lie, but the author engaged in complex analyses and questionable assumptions to build a model to arrive at his desired conclusion. The first thing one notices when examining figure 1 is that its pretty clear the person reading my study had no idea what was happening and has no idea about the history of Canadian firearms legislation. As stated in my study, gun laws were implemented in the mid to late 1990s and early 2000s.

Prior to the gun laws the rates of suicide using firearms in males and females are dropping at a certain rate, or trend, but after the rates of suicide drop a lot less and seem to flatten out in the 2000s. Just looking at the graph one would visually say that the gun laws made things worse, because if the gun laws prevented suicide we should see a much steeper decline in suicide by firearm after the laws are implemented!

Regardless, using more complex statistical methods than visually eyeballing graphs, I actually show some associated reductions in firearms suicide in some years and age groups as discussed in the paper, but this is negated by increases in hanging. A switch to other methods. The person critiquing this

section seems to completely miss or ignore the dramatic increase in hanging, especially in females. Figure 2 is regarding homicide by firearm and non firearm. Once again, the person critiquing this

section seems to have no clue about the history of Canadian firearms laws, and how social scientists use time series analysis. Prior to the laws in the mid 1990s and early 2000s we can see in both graphs homicide rates falling at a certain trend and then post the implementation of firearms legislation the rates of homicide seem to fall at the same amount or even appear to increase slightly. A similar trend is seen with non firearm homicides. This suggests that gun laws are not associated with a reduction in homicide rates. I would also say the trends do not lie, but the person reviewing my study has lied to themselves.

2023-10-27 Langmann 6

2023-10-27 Langmann 7

2023-10-27 Langmann 8 References: Bennett, N., M. Karkada, M. Erdogan, R. S. Green and N. S. R. P . Heal (2022). "The effect of legislation on firearm-related deaths in Canada: a systematic review." CMAJ Open 10(2): E500-E507. Caron, J. (2004). "Gun control and suicide: possible impact of Canadian legislation to ensure safe storage of firearms." Arch Suicide Res 8(4): 361-374. Caron, J., M. Julien and J. H. Huang (2008). "Changes in suicide methods in Quebec between 1987 and 2000: the possible impact of bill C-17 requiring safe storage of firearms." Suicide Life Threat Behav 38(2): 195-208. Chapman, S., P .

Alpers and M. Jones (2016). "Association Between Gun Law Reforms and Intentional Firearm Deaths in Australia, 1979-2013." JAMA 316(3): 291-299. Cheung, A. H. and C. S. Dewa (2005). "Current trends in youth suicide and firearms regulations." Can J Public Health 96(2): 131-135. Gilmour, S., K. Wattanakamolkul and M. K. Sugai (2018). "The Effect of the Australian National Firearms Agreement on Suicide and Homicide Mortality, 1978-2015." Am J Public Health 108(11): 1511-1516. Langmann, C. (2012). "Canadian Firearms Legislation and Effects on Homicide 1974 to 2008." Journal of Interpersonal Violence 27(12): 2303-2321.

Langmann, C. (2020). "Effect of firearms legislation on suicide and homicide in Canada from 1981 to 2016." PLoS One 15(6): e0234457. Langmann, C. (2021). "Suicide, firearms, and legislation: A review of the Canadian evidence." Prev Med 152(Pt 1): 106471. Langmann, C. (2022). "Mass Homicide by Firearm in Canada: Effects of Legislation." medRxiv: 2022.2003.2024.22272877. Langmann, C. (2023). "Canada, Firearms Legislation, Suicide and Homicide. A Review of the Evidence." SSRN. Langmann, C. (2023). "Mass homicide by firearm in Canada: Effects of legislation." PLoS One 18(2): e0266579. Lee, W. and S. Saurdi (2008).

"The Australian Firearms Buyback and its Effect on Gun Deaths." Contemporary Economic Policy 28(1): 65-79. Mauser, G. and R. Holmes (1992). "An Evaluation of the 1977 Canadian Firearms Legislation." Evaluation Review 16(6): 603-617. McPhedran, S. and J. Baker (2008). "Australian homicide: no significant impact of gun laws." Med Sci Law 48(3): 270. McPhedran, S. and G. Mauser (2013). "Lethal Firearm-Related Violence Against Canadian Women: Did Tightening Gun Laws Have an Impact on Women’s Health and Safety?" Violence and Victims 28(5): 875- 883. Pear, V. A., G. J. Wintemute, N. P . Jewell and J. Ahern (2022).

"Firearm Violence Following the Implementation of California's Gun Violence Restraining Order Law." JAMA Netw Open 5(4): e224216. Rich, C. L., J. G. Young, R. C. Fowler, J. Wagner and N. A. Black (1990). "Guns and suicide: possible effects of some specific legislation." Am J Psychiatry 147(3): 342-346. Siegel, M., M. Pahn, Z. Xuan, E. Fleegler and D. Hemenway (2019). "The Impact of State Firearm Laws on Homicide and Suicide Deaths in the USA, 1991-2016: a Panel Study." J Gen Intern Med 34(10): 2021- 2028.