Sandy Hook: A Quantitative Criminological Inquiry
If the inordinate secrecy, multiple inconsistencies, and baffling residual mysteries associated with the Establishment’s narrative regarding the Sandy Hook event are any indication, the inner nature of the event is likely to be very dark indeed.
[Image credit: NPR]
The complexity of the event means in part that analyzing it rationally demands the application of the talents and experiences of a broad array of individuals united in their dedication to the truth and the right to know. Happily, much progress has already been made in this regard; much quality work has been produced.
It is hoped that the present effort, rendered from the perspective of a criminologist, constitutes a small contribution to scrutiny of the event from a quantitative vantage point.
To be sure, quantitative methodologies are no more ultimately dispositive than qualitative methods of inquiry. Quantitative and qualitative methods complement, and typically augment one another. Thus, quantitative scrutiny of the Sandy Hook event might tend to bolster claims emanating from careful analysis of statements, reports, video footage, and the like.
What follows is a quantitative analysis of “Active Shooter” data gathered by the NYPD’s counterterrorism unit (Anyone can access the data by clicking here.) The Department of Homeland Security (DHS) defines an active shooter as “an individual actively engaged in killing or attempting to kill people in a confined and populated area; in most cases active shooters use firearms(s) and there is no pattern or method to their selection of victims.”
It will be apparent that since there is no victim specification requirement (in terms of amount) associated with the “active shooter” designation, while active shootings may amount to mass murders, not all of them do. Similarly, in part because there are no location and weapon requirements connected with the definition of mass murder, not all mass murders are active shootings.
The NYPD’s counterterrorism unit narrowed DHS’s definition by excluding gang-related shootings, shootings that occurred solely in domestic settings, robberies, drive-by shootings, attacks that did not involve a firearm, attacks categorized primarily as hostage-taking incidents, and attacks that did not “spill beyond intended victims” (Kelly, 2012; click data link above).
The data analyzed range from 1966 to the Newtown shooting. The data were decomposed such that only cases of shooting mass murders in schools were included, since that is the type of case Sandy Hook ostensibly presents us with.
The question broached and rigorously tested with the NYPD data was, to the best of my knowledge, initially brought to light by Dr. James Tracy. The question concerns the likelihood of having observed, in an active shooting that was also a mass murder that transpired at a school, figures on the order of 26 dead (not including Adam and Nancy Lanza) and only 2 wounded (keeping in mind that, to the best of my knowledge, the second wounded victim has still not been identified).
Thus, the idea behind the analysis was to determine whether a statistically significant relationship exists for the 34 above-described cases (Sandy Hook is not included in the dataset, since its validity is precisely what is in question), such that as the total number of victims increases, the likelihood that a greater number of them will end up wounded rises. Intuitively, as the total number of victims rises, chance alone would dictate that we would expect more of the victims to end up categorized as wounded, since there would be an increasing number of potential victims to assign to that category. Clearly, there is also a surfeit of non-chance reasons why we might expect this relationship to hold, but in the interest of space elaboration of those reasons is omitted.
To answer this question, logistic regression was employed. With logistic regression, the dependent variable is dichotomized so that it has two levels. Here, if a case had more than 2 wounded victims, the case received a score of “1” for the dependent variable, while if it had 2 or fewer wounded victims, the case received a score of “0” for the dependent variable. The independent variable was the total number of casualties in the active shooting (dead plus wounded).
Essentially, logistic regression was used to model the probability, given the total number of victims in a case, of observing 3 or more wounded victims as opposed to two or less (and since the two possibilities are mutually exclusive, one probability manifestly implies the other).
The analysis was conducted in STATA 12 and produced the following results:
The key result of course pertains to the casualties variable. The odds ratio of around 1.34 (itself significant at p <.05) indicates that for each additional casualty (whether dead or wounded), the probability that the wounded count will be greater than two rises by 34%. Clearly, since Sandy Hook supposedly involved a total of 28 casualties, there are grounds to be suspicious that only 2 were wounded (the schooldi variable is interesting in its insignificance, and shows that whether the active shooting mass murder took place at a high school as opposed to an elementary or middle school was irrelevant with respect to whether more than 2 were wounded).
Furthermore, with this result in hand we can estimate the probability that with 28 total victims 3 or more of them would end up wounded rather than dead. Here is the output in connection with this issue:
The key figure is .9976669, which signifies that the probability is around 2 in 1000, or 1 in 500, that fewer than 3 would be wounded when the victim count in an active shooting mass murder at a school is 28.
While the 1 in 500 figure already indicates that odds are very low indeed that an active shooting mass murder at a school with 28 total victims will yield fewer than 3 wounded, more can and should be added. First, as was alluded to above, the second wounded victim has still not been identified. The statistical results, put together with the reasoning in support of them, suggest that the odds of fewer than 2 wounded victims (that is, 1 or 0) when there are 28 victims in total would be lower than 1 in 500; they might even be much lower. Furthermore, even if we go with the 1 in 500 figure, given the rarity of active shooting mass murders in schools, we might well expect to wait hundreds of years before such an event genuinely arises.
We shouldn’t leave this particular issue without noting that, from a statistical and criminological point of view, while the distinction between one and two wounded victims may seem inconsequential to the untrained eye, its statistical import might well be huge, especially when more data is compiled. This suggests that Sandy Hook researchers may want to redouble efforts with respect to investigating all available evidence concerning who this second purported wounded victim was and so forth.
Finally, the NYPD data supplies more statistical trouble for the Establishment’s Sandy Hook narrative. Unless I’ve overlooked something (and I don’t believe I have), according to the NYPD there are no other instances period of active shooter mass murderers who killed in schools and also committed matricide beforehand.
In sum, traversal of the byzantine, highly detailed investigative highways, byways, and “lieways” of Sandy Hook leaves dedicated researchers with the sense that all is far from right regarding Newtown. It is believed that the current offering provides some reason for concluding that the numbers don’t add up either.
Jason Kissner, Ph.D., J.D., is associate professor of criminology at California State University, Fresno. You can reach him at crimprof2010(at)hotmail.com.
No comments:
Post a Comment