In article <411270$p5f@nntp.Stanford.EDU> fkk@leland.stanford.edu writes:
>>Actually, my analysis method was:
>>Mean homicide rate in years 1910-1920: 2.3
> This is a quote for a posting by Tim
>>Mean homicide rate in years 1921-1930: 1.5
>>Difference = -0.8 t statistic is -4.9 (19df) p = 0.00009
>>I.e. the mean homicide rate 1921-30 is significantly less than that
>>for 1910-20.
> I am asking that he send me the data - not just the numbers so I
>may look at this myself. Tim, you post numbers without units -
>what are you measuring? What was the sampling
>period?
I've appended a whole bunch of data to the posting so anyone can play
with it. It tells you the units and sources for the data.
>> If one with the drop
>>after 1915 fits the data better than the one with the drop after 1920,
>>then it is impossible that this drop was caused by gun control.
> From this statement I conclude you analysis shows the drop occurs
>before the gun control.
No, you have misunderstood. That's a conditional statement above. I
went on to point out that in fact the drop seems to occur following
gun control. (That is, the step model that best fits has the drop
after 1920.)
Of course, just because a drop in homicide rates follows a gun control
law, it does not follow that the gun control law caused the drop. It
is possible that some other factor caused the drop. In the case in
question the most plausible other factor is the Great War. By killing
lots of hotheaded young males it could have depressed homicide rates
for a decade or more. However, homicide rates in the neighbouring
states of Victoria and Queensland do not show similar drops after
1920, so it does not seem likely that war caused the drop since it
should have affected Victoria and Queensland as well.
Tim
Here are the Australian states for which I have homicide data. (Data
for WA and Tas are incomplete in ways that make them useless for
evaluating gun laws.)
Qld and SA figures come from "Source Book of Australian Criminal & Social
Statistics 1900-1980" Mukherjee. Victoria from "Victorian Year Book"
volumes for 1901,1902,1903,1904, etc. NSW from "Homicide:The Social
Reality" Wallace. -1 means no data available. All are rates per 100,000
pop. NSW, Vic are for homicides, SA for homicides including manslaughter by
driving, Qld are for charges for murder and manslaughter.
NSW Qld SA Vic
1920 1927 1929 1921 <- gun laws introduced
1900 1.5 7.0 -1.0 -1.0
1901 1.4 9.0 -1.0 2.6
1902 2.3 7.5 -1.0 2.2
1903 1.0 7.1 -1.0 1.2
1904 3.0 7.5 -1.0 1.9
1905 1.8 6.6 -1.0 3.3
1906 2.1 6.1 -1.0 1.6
1907 3.0 7.4 -1.0 1.7
1908 3.0 6.8 -1.0 1.5
1909 3.1 4.4 -1.0 1.2
1910 2.6 4.6 -1.0 3.1
1911 2.7 4.0 -1.0 1.8
1912 2.4 3.1 -1.0 2.1
1913 2.1 4.4 -1.0 1.8
1914 2.4 4.0 -1.0 1.6
1915 2.6 3.9 -1.0 1.7
1916 2.1 4.2 -1.0 1.4
1917 1.9 4.9 -1.0 1.3
1918 1.4 5.3 -1.0 1.3
1919 2.7 3.8 -1.0 1.8
1920 2.7 3.5 -1.0 1.2
1921 1.8 3.8 3.4 1.4
1922 1.6 2.9 3.0 1.5
1923 1.0 3.6 2.0 1.3
1924 0.9 5.1 5.1 1.1
1925 1.1 3.7 3.7 1.3
1926 1.6 5.1 4.0 2.0
1927 1.6 6.1 3.0 1.8
1928 1.9 3.8 2.8 1.8
1929 1.7 3.7 3.9 2.0
1930 1.7 2.6 1.2 1.1
1931 1.6 2.4 2.4 2.3
1932 1.3 3.3 1.6 0.9
1933 1.5 2.5 1.1 1.6
1934 1.4 2.5 1.8 1.4
1935 1.8 2.4 0.9 1.3
1936 1.4 1.5 1.9 1.0
1937 2.0 2.5 2.5 1.0
1938 1.2 3.3 1.5 1.3
1939 1.3 3.3 0.6 0.8
1940 1.7 3.6 0.8 0.6
>From "Australia During the War" I computed the
percentage of males 18-44 in NSW who had enlisted by the middle of
each war year:
1914 0
1915 13
1916 27
1917 34
1918 38
Here are homicide figures for all the countries from "Violence and
Crime in Cross National Perspective" that participated in WW1 (figures
are convictions in some countries, so are not comparable across
countries)
England Germany Japan Scotland
Australia Italy Portugal USA
1911 0.81 0.74 0.52 8.88 3.09 2.05 0.68 6.6
1912 0.86 1.08 0.55 9.83 3.17 2.16 0.85 6.5
1913 0.91 1.12 0.55 9.52 3.82 2.79 0.66 7.2
1914 0.74 1.20 0.63 8.36 3.88 2.97 0.65 7.3
1915 0.74 0.82 0.53 8.21 3.47 2.44 0.94 6.9
1916 0.70 0.71 0.28 6.69 3.53 2.48 0.92 7.1
1917 0.61 0.59 0.27 6.05 3.59 2.43 0.48 7.6
1918 0.55 0.41 0.24 5.41 3.42 2.49 0.48 6.8
1919 0.80 0.91 0.29 8.42 3.52 2.81 0.78 7.5
1920 0.84 0.71 0.70 13.6 3.34 3.50 0.84 7.1
1921 0.66 0.84 0.98 15.37 3.31 3.50 0.72 8.5
1922 0.64 0.82 1.01 16.69 3.29 3.54 0.68 8.4
1923 0.67 0.53 0.85 14.44 4.30 3.01 0.66 8.1
1924 0.67 0.58 0.67 11.14 3.97 3.04 0.68 8.4
1925 0.75 0.69 0.96 11.11 4.04 2.97 0.86 8.6
1926 0.72 0.61 0.90 8.99 4.36 3.13 0.72 8.8
1927 0.71 0.73 0.91 7.43 4.01 2.66 1.05 8.7
1928 0.65 0.67 0.79 5.89 3.70 2.84 0.85 8.8
1929 0.73 0.75 0.63 5.28 3.22 3.77 0.95 8.5
1930 0.72 0.45 0.65 5.28 3.58 3.15 0.66 9
Some places had no increase after the war (Japan, Portugal), other
places did have an increase that lasted at least into the mid 20s
(Germany, Italy, USA). Only England and Scotland followed a similar
pattern to NSW, of a drop after 1920. However, gun control was
introduced in England and and Scotland in 1920.
I also have a breakdown of the homicide rate for NSW, gleaned from the
NSW statistical register.
NSW, rates per 100,000 population.
Column 2 firearm homicides
Column 3 cutting or piercing instrument homicides
Column 4 Other means homicides
Infanticide and justifiable homicides excluded
(except that 1928-30 includes justifiable homicides)
1910 0.6 0.2 1.2
1911 0.4 0.1 1.3
1912 0.7 0.1 0.8
1913 0.8 0.1 0.8
1914 0.7 0.3 0.9
1915 0.9 0.3 0.8
1916 0.5 0.4 0.8
1917 0.4 0.1 0.6
1918 0.3 0.1 0.7
1919 1.0 0.3 1.0
1920 1.1 0.3 1.0
1921 0.6 0.1 0.8
1922 0.1 0.0 0.9
1923 0.1 0.1 0.6
1924 0.5 0.4 1.1
1925 0.2 0.2 0.5
1926 0.3 0.3 0.5
1927 0.5 0.2 0.8
1928 0.5 0.3 1.0
1929 0.6 0.4 0.7
1930 0.7 0.1 0.8
t tests comparing 1910-20 with 1921-1930 show significant decrease in
firearms rate, overall rate (excluding infanticide and justifiable), and
firearms percentage. Changes in sharp instrument, and other means rates
were not significant.