In article none+1 stratos@crl.com Janine K. Johnson wrote:

Recently, several postings discussed the Orlando Florida phenomena of 1966/67,
in which a drop in the rape rate was noted after a much publicized program co
sponsored by the local police and the Orlando Sentinel, in which 6,000 women
were trained in the use of firearms.

Knox, and other experts who have analyzed the Orlando phenomenon, contend
that the rape rate decreased because of the media publicity, and because
women were armed and trained. (Paxton Quigley, Armed and Female, St.
Martins Press.):

Mr. van Meurs rebutted this contention by showing data that demonstrate that
when examined over a longer period of time, the drop between the 1966 and
1967 rape rate is not distinguishable from the noise of random variations in
rapes in Orlando. Based on this, Mr. van Meurs believes he has proved the
training program did not have any effect on the rape rate.

I think you have misunderstood him. The training program might have
had an effect on the reported rape rate. However, you cannot reject
the alternative explanation that the decrease was due to random
variation in the rate.

Steve D. Fischer writes:

No, Mr Van Meurs did not prove his case. He failed to show that the
previously observed drops in rape did not ALSO occur because of:

(1) Increased police vigilance due to public outrage over past increases
in the rate of rape.

(2) Media attention given to the high magnitude of rape.

One simple explanation could be called the “pissed off” model. Rape continues
to rise until it reaches a level at which the public becomes pissed off and
pressures the police to do something about it. The increase in police activity
either gets the rapists off the streets or scares them away to an area which
is receiving less heat at the moment. Viewed in that way, the gun training
program was simply one way the public showed its outrage. All the past
decreases could have occurred through very deterministic ploys aimed at getting
the rape rate down to “an acceptable level.”

Perhaps they did. However, this is an argument AGAINST the
possibility of the gun training causing the decrease. If these other
measures decreased the rate in the past then this suggests that they
decreased the rate in 67 as well, and the gun training may not have
had any effect.

The potential mistake Mr Van Meurs has made is in assuming that rape is
totally random. The only way to test my hypothesis is to go back and read
the news accounts each year in Orlando and determine if special efforts were
taken in the years that the previous decreases in rape were noted.

There’s a big problem with your hypothesis. Let me illustrate with an
example:

I’m going to roll a pair of dice to simulate a random crime rate. If
the rate gets into double figures the community is outraged and
demands that something be done to reduce it. I’ll respond by saying
“Abracadabra” before rolling the dice to try to drive the crime rate.
Here goes:

4 9 7 9 9 5 11 (Oh no! Crime has skyrocketed! Abracadabra!!) 6
(Phew! Got it back down.) 7 10 (Oh no! It wore off! Abracadabra!!) 6
6 10 (Abracadabra!!) 4 (Wow! reduced it 60%!) 7 8 etc etc.

Every time I said “Abracadabra” the rate went down. Do you think that
I made it go down?