"An Armed Society is A Polite Society"--A Testable Hypothesis?
Aleksandra de Accipitre

Everyone in the SCA has heard the title quote; many recite it as an
article of faith.  I have recently moved from approaching this as a
skeptic (Where is your evidence?) to approaching it as a social
scientist:  How could you test that statement?

(A brief digression for those who think "social science" is an
oxymoron: While not as precise as the physical sciences, it can make
probabilistic predictions not unlike the life sciences.  Medical
research cannot tell us whether an individual will get cancer, but it
can accurately predict that a group of smokers will have more cases of
lung cancer than a comparable group of non-smokers.  Similarly, I can't
tell you how a given individual will vote, but I can predict for 1996
that if all the major presidential and vice-presidential candidates are
white, the Democrats will get 70-plus percent of the black vote.
$50.00....any takers?)

Back to the issue at hand:  we have a hypothetical correlation between
two variables ("armed" and "polite"), both of which are attributes of
"societies."  We also have an implied causal relationship:  being armed
causes a society to be polite.  We can begin to build a theory from
this by suggesting a causal mechanism:  people are more likely to be
polite if they know that being impolite is likely to get them killed
(or even maimed).  So far, so good; looks like social science to me.

Now, how can we design a research project that would test this
relationship?  A comparative approach seems appropriate:  if a number
of societies differ in the degree to which they are "armed," we should
find that the more "armed" ones are also more "polite."  Sound simple?
OK, let's start by defining our terms:  "armed," "polite," and
"society."  Of course we all intuitively understand what we mean by
these terms, but to make a rigorous test, we must carefully specify, in
advance of collecting our data, just what it is we are going to
compare.

How do we decide which of two societies is more "armed"?  Should we
compare societies with different types of weapons (e.g., clubs, swords,
muskets, Uzis, ICBMs)?  Or is number of weapons per capita a more
appropriate measure?  If so, should we include military and police
weapons, or not?  Should we consider only the average number of weapons
per person, or do we need to know how evenly they are distributed?
(This may not be as straightforward as it first looked.)  Once we
decide on a definition, we still have to figure out how to measure it
consistently across different societies,  and decide how many levels we
should use:  armed/unarmed, slightly armed/moderately armed/extremely
armed, or something else.

Next, "polite."  Should we measure this by using the right fork and
writing thank-you notes?  By elaborateness of formal etiquette?  Should
we define politeness as "being considerate" (and how would we measure
that)?  We could define "polite" societies as those having lower crime
rates, but since societies define "crime" differently, our measures
might not be comparable.  Besides, if we really mean "an armed society
is a low-crime society," we should say so in the first place, and not
muddy the discussion with vague terms like "polite."

Assuming we have come up with acceptable measures for our independent
and dependent variables ("armed" and "polite"), we must then define our
unit of analysis  ("societies"), and select cases for observation.
Obviously, a society is a group of people who interact according to
norms and patterns of behavior.  But what are its boundaries?  Should
we compare countries, cities, or neighborhoods?  How about tribes, or
religious "communities" (Christendom vs. Islam, for example)?  Should
we do our comparison across time, and how will this affect the
measurement of our variables?

Ideally, we should compare at allof these levels, but resource
constraints may force us to be selective.  What we cannot do is
meaningfully compare across different levels (Japan vs. Bensonhurst, or
Apaches vs. the Holy Roman Empire).  Additionally, we must be sure our
cases are comparable--differently armed, but as similar as possible in
other respects.  (Whether they are also differently polite is what
we're trying to discover.)

The final thing a research design must consider is:  what evidence, if
found, would disprove our hypothesis?  A test is valid only if failure
and success are both possible.  We must squarely face the possibility
that our theory is wrong, and specify in advance how we will know if
this is so.  (If we don't know what will prove us wrong, how can we
know what will prove us right?)

These are the bases of a research design--the details are left as an
exercise for the student.  Maybe someone could get a government grant
to carry out such a study.  Alternatively, apply to the NRA.  I'm sure
they'd love to fund a study to prove this theory (unless, of course,
they're afraid of it being disproved).  Happy Research, and do keep me
posted--I'm still waiting for the evidence.



© copyright Sondra Venable 1996.  Not to be reproduced without the author's permission.




"If it can't be expressed in figures, it is not science; it is
opinion."

"What are the facts?  Again and again and again--what are the facts?
Shun wishful thinking, ignore divine revelation, forget what "the stars
foretell," avoid opinion, [emphasis added, and see above] care not what
the neighbors think, never mind the unguessable "verdict of
history"--what are the facts, and to how many decimal places?"

--Robert A. Heinlein,  The Notebooks of Lazarus Long


From the June 1996 Seahorse