We’ve been around on this before, and all it does is impress me with
the predilection of some pro-gun folks for self-delusion on this
topic. (I’m sorry if that sounds harsh, but it seems to me that many
people suspend their powers of reason on this issue.)

Here are the NSW homicide rates from 1910-1920:
2.6 2.7 2.4 2.1 2.4 2.6 2.1 1.9 1.4 2.7 2.7
Care to demonstrate a decline?

Geoff Miller said:

What can be demonstrated very easily is that you are supporting your own
position with some very carefully chosen figures.

Untrue. Frank Crary claimed that there was a decade long decline,
ending after 1920. If the rate was declining before the 1920 gun law,
then that should be evident using just data from before 1920 (and
after 1910, since he said it was a decade long decline.)

You are the one guilty of creative data selection as we will see below.

Let’s look back at the ASCII graph of the NSW data you posted some
time ago:

3      *  ***
2.8
2.6          **   *   **
2.4            * *
2.2  *   *
2.0             *  *                    *
1.8     *           *   *      *      *                     **
1.6                      *   ** ***        *   *  *       **  **   ** **
1.4**                *              ** *      *    **   *       ***  *
1.2                                *     **  *  **    ** *              *
1.0   *                   * *                        *
0.8                        *                *
   0         1         2         3         4         5         6         7
   01234567890123456789012345678901234567890123456789012345678901234567890

From 1907 to 1918 there is a clear declining trend, an aberrant increase
in 1919 and 1920 almost certainly related to social and other disruption
accompanying the end of the war and the return of a large number of
servicemen … Firearms
controls introduced in 1920 were associated (at least chronologically)
a reduction in the murder rate to about the level in 1924 which it would
have reached had the trend of 1907-18 been continued.

Now, we can argue till doomsday about what caused the reduction from 1920-24,
and also the subsequent rise to a relatively stable level around the 1.4
mark. Presumably the gun-control legislation contributed a short-term
effect (which, historically, is all that gun-control legislation does seem
to achieve – the Canadian data show this quite clearly), but this would
have been associated with a reduction to be expected as many of the
ex-servicemen adjusted back to a lifestyle in which violence was not
acceptable.

I would also accept the point, if you care to make it, that the reduction
from 1914-18 was no doubt influenced to some degree by the fact that many
of the young and potentially violence-prone were overseas. However, I
would expect this effect to be much smaller than the increase associated
with the return of the servicemen in 1919-20.

To show that there was a declining trend, you throw out the data that
does not fit. 1919-20 does not fit your trend, so the homicide rate
in those years was affected by WWI and you discard this data. 1914-18
does fit, so the effect of WWI on these years was “much smaller”, so
you keep this data. 1921-24 fits, so by then the diggers must have
adjusted to civilian life. 1900-1906 does not fit, so you start your
trend in 1907.

Now, if I wanted to play the same game, I could exclude 1915-1923 (to
avoid all effects of WWI) and include ALL the pre-war data. Looking
at 1900-1914, we note a definite UPWARD trend, which if continued
would have resulted in a rate of 3.5 in 1924. Gun control
obviously reduced this to 0.8 :-)

It is more interesting to investigate the possible effects of WWI. It
is possible to roughly control for the effects of many young males
being overseas. 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

(It is chilling to see that 5% of males 18-44 died in the war. “Lest
we forget”.)

According to “Homicide: the Social Reality”, 75% of homicides in NSW
are committed by males 18-44, so we would expect a 38*.75% reduction
in the homicide rate in 1918 because of fewer killers being around.
(This probably underestimates the effect, since proportionately more
killers are younger, and more likely to enlist.)
We can compute adjusted homicide rates for the war years, controlling
for this effect. Here is 1910-20, using the adjusted rates

 10  11  12  13  14  15  16  17  18  19  20
2.6 2.7 2.4 2.1 2.4 2.9 2.6 2.6 2.0 2.7 2.7

and here is 1900-30 in an ascii graph

3      *  
2.8             
2.6
2.4 * 2.2 * 2.0 * 1.8 * * 1.6 * 1.4 1.2 1.0 * * 0.8
0 1 2 3 0123456789012345678901234567890

It does not appear that that the 1919-20 values are particularly
different from the preceding years, but let’s investigate further the
possibility that there was a homicide increase associated with
returning servicemen that lasted just two years.

Such a phenomenon would presumably also occur in other countries that
participated in WWI. Here are homicide figures for all the relevant
countries from “Violence and Crime in Cross National Perspective”
(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 into these two countries in 1920. Hmmm.

I don’t think I’m particularly prone to self-delusion or suspension of
the reasoning powers, but nor am I prone to using a carefully-selected
subset of my data to support an otherwise indefensible position.

It seems to me that you did just that.

Please note that my comments about self-delusion were not intended as
a flame but an attempt to communicate the way all this talk about the
rate declining before 1920 makes me feel.