TELECOM Moderator <telecom@eecs.nwu.edu> (06/10/90)
A couple weeks ago, I posted a survey asking questions about possible
illegal telecommunications/computer activities by readers. It looks
like the survey was flawed, due to an error in the way I set it up.
There have been several messages which pointed out that unless I was
aware of the dietary and banking habits of the readers, it would be
impossible to accurately give any meaning to the results.
To avoid embarassment or possible legal ramifications, readers were
asked to answer one set of questions or another set, based on the flip
of a coin. The two question sets were:
1) 1. Have you made one or more phraud calls in the past six months?
2. Have you broken into a computer or gained unlawful access to
a computer in the past six months?
2) 1. Have you eaten a hamburger for lunch in the past two weeks?
2. Have you gone inside the bank you usually do business with in
the past two weeks?
We know that as the number of coin tosses increases, the likelyhood is
that there will be an even number of heads/tails come up. So, we can
take the number of answers received, *assume that half were answering
the relevant questions and disgard half the results, evenly from all
possible answer groups*, getting some idea of how many of you are
naughty, and how many are nice.
But several of you wrote to point out that without knowing if a
preponderance of the readers here were vegetarians, or misers who did
not trust banks, the results would be difficult or impossible to
interpret. If no one eats hamburgers or visits banks, then a large
number of 'no-no' answers would appear. And, this is in fact what
happened when the results were tallied -- more on this later.
A better way of handling the survey, aiming for the highest possible
number of accurate answers while still allowing a relative anomynity
in posting would have been to ask but the first set of questions --
the relevant ones -- with the condition that if the coin toss was
heads, answer the questions accurately. If the coin toss was tails,
then flip the coin twice more: (1) heads/tails = yes/no on phraud calls;
then (2) heads/tails = yes/no on computer cracking. In either event,
do not reveal the coin toss -- simply send along your answers.
Here are the results from the first time, although flawed. They are
presented for your amusement, and I hope you will answer the survey a
second time, using the more accurate collection techniques.
Total respondents: 636
36 (5.7%) answered yes to both cracking and phreaking in the recent past.
78 (12.3%) answered no to phreaking and yes to cracking.
66 (10.4%) answered yes to phreaking and no to cracking.
456 (71.7%) answered no to both questions. No cracking or phreaking.
Percentages slightly over 100% due to rounding.
Of course, if we assume half the respondents were talking about their
lunch and financial matters, then the figures would decrease, but the
percentages would stay the same.
Still, as pointed out above, the results are suspect, so let's do it
over again:
Flip a coin. Heads, answer these questions honestly.
1) Have you made one or more phraud phone calls in the past six
months?
2) Have you broken into a computer or gained unlawful access to
someone else's account in the past six months?
If tails, flip the coin twice more:
On the first flip, answer the first question: heads = yes / tails = no.
On the second flip, answer the second question: heads = yes / tails = no.
Then, mail your results, with the subject header 'survey' to:
telecom-request@eecs.nwu.edu
Your answers should take one of these forms:
1) Yes Yes
2) Yes No
3) No Yes
4) No No
Avoid the header 'questions' since some late responses to the first
survey are using this header. Do not reveal the coin toss(es) and do
not make other comments for which a reply is expected. Results to
this hopefully more accurate version will appear in a couple weeks.
Patrick Townson
TELECOM Moderator
PS: And remember, Bob Dobbs explained it thus: " I don't practice what
I preach because I am not the kind of person I am preaching to!" :)