2aHawaii
General Topics => Legal and Activism => Topic started by: Tom_G on October 02, 2015, 08:25:26 AM
-
Like many things, this came to me via facebook. The linked article (http://www.motherjones.com/politics/2014/10/mass-shootings-rising-harvard) attributes the story originally to the LA TImes (http://www.latimes.com/opinion/op-ed/la-oe-follman-rise-in-mass-shootings-20141020-story.html) and the study to Harvard. Note that I haven't found the study.
-
Like many things, this came to me via facebook. The linked article (http://www.motherjones.com/politics/2014/10/mass-shootings-rising-harvard) attributes the story originally to the LA TImes (http://www.latimes.com/opinion/op-ed/la-oe-follman-rise-in-mass-shootings-20141020-story.html) and the study to Harvard. Note that I haven't found the study.
Okay, so they took a period of time from 1982-2011 (30 years) and calculated the average time between mass shootings.
Then they took a 4-year average to show a higher frequency of mass shootings are occurring.
Anytime you take a large slice of time, the period "dilutes" the sample, reducing the average. When you take a smaller slice of time, in this case one that's 13.3% as big as the 30 year slice, and if that slice happens to have a concentration of shootings, then it will look like there is an increase.
The study itself is making conclusions based on flawed comparisons. I could take any 4 years in the 34 year dataset and make it look like there were more of fewer shootings than the 30-year average being used.
An analogy would be the year we had 40 days and nights of rain. Comparing that 2 month period with the previous 36 month's average would have yielded the same kind of comparison, and the false conclusion would have been that rainy days have become more frequent than the previous 3 years.
-
Okay, so they took a period of time from 1982-2011 (30 years) and calculated the average time between mass shootings.
Then they took a 4-year average to show a higher frequency of mass shootings are occurring.
Anytime you take a large slice of time, the period "dilutes" the sample, reducing the average. When you take a smaller slice of time, in this case one that's 13.3% as big as the 30 year slice, and if that slice happens to have a concentration of shootings, then it will look like there is an increase.
The study itself is making conclusions based on flawed comparisons. I could take any 4 years in the 34 year dataset and make it look like there were more of fewer shootings than the 30-year average being used.
An analogy would be the year we had 40 days and nights of rain. Comparing that 2 month period with the previous 36 month's average would have yielded the same kind of comparison, and the false conclusion would have been that rainy days have become more frequent than the previous 3 years.
I am trying to figure out their methodology but I don't know if your analysis is accurate.
Looking at the second chart on the Mother Jones link you can see that in the previous years it was frequently the case that the shootings were occurring much more often than the average 200 days however there were those tall spikes of time in which no shootings occurred. It would seem that in the last 4 years we haven't had any of those longer periods of no mass shootings. While it is possible that we are going through a rainy period, the data set would seem to clearly show that there have been more mass shootings recently. I know some have reported that mass shootings were on the decline but this report shows otherwise. The top chart dataset pretty clearly shows an increase in the number of points compared to all the previous 30 years, even the other clumps of incidents.
This is assuming the data presented is accurate. I don't trust mother jones but I trust FBI statistics.
I do think that it is right of the FBI to study beyond their definition of mass shooting (4 people dead) because this in some ways can provide an inaccurate picture. If for example someone shoots 12 people but only 3 die then technically it wasn't a mass shooting by definition, though it should still be considered when studying this phenomenon.
I think one weakness of this is that it doesn't take account for population.
On top of that, just looking at numbers doesn't really solve anything. Whether it is happening a little more or a little less, we need to find out why these things are happening and stop them. How can the mental health system help, how can we reduce bullying, how can we get the media to stop making these people more famous, etc.
I know this is in conflict with data Michael Medved mentions on his show but perhaps the inclusion of other sets of data explain the differences.
-
Study seems reasonable but I agree with mauidog that it could just be coincidence. They should do this study again after another 10 years to see how it holds up.