RE: Re: Flawed Surveys [was: VPN endpoints]

["Tina Bird" <tbird () precision-guesswork com>]

> However, surveys can be used to test an hypothesis, which is  what I
> think Marcus and Paul are going back and forth on.
I am being quite precise, cos poorly defined and used vocabulary is part of
the problem.  I'm willing to agree that surveys can be used to collect
information about a question.  But that question isn't a "hypothesis" -- a
prediction derived from a theory.

I suppose your theory might be that having your network off line for two
weeks because of Blaster is an effective goad to increasing your budget in
security, but it's rarely put exactly that way (and as theories go that's
kinda weak)...

> I mention this for two reasons.  First is there a better question ever
> to ask someone than if they agree to the following:  "I am a secret
> agent of God."?     :-)
Which God?  And aren't we all????

> Second, if one allows this method as science, then there is a common
> ground for Paul and Marcus. As long as there is an agreed on validation
> method.  Isn't the lack of that the rub here?

It's not science, but I'm not sure that matters.  What I'm hearing is:

- "people" are curious about "other people's" attitudes toward security
(where "people" and "other people" are deliberately vaguely defined)
- "people" think that asking questions and collecting answers is a good way
to collect information about the question

--> so it comes down to, what is the question we're investigating, and do we
agree that collecting the answers to the question from a self-selected (and
difficult to externally validate) set of respondents is a good way to
investigate?  It's not science, although it shows glimmers of being rational
:-)

Although I think I am with Marcus on this one -- after all, is asking one's
partner "Do you love me?" a good way to answer the question?  Or do you get
more reliable data by collecting it in other ways?  All of the data you
collect is interesting, but it is more or less useful, depending...

cheers -- tbird


_______________________________________________
firewall-wizards mailing list
firewall-wizards () honor icsalabs com
http://honor.icsalabs.com/mailman/listinfo/firewall-wizards