Please respond to lbo-talk at lists.panix.com
To: lbo-talk at lists.panix.com cc: (bcc: DANIEL DAVIES) bcc: DANIEL DAVIES Subject: Re: Mr. Byfield's C- Posts
Charles Brown wrote:
>Interrogation, which is a postmod method, I believe.
>Damn, those postmodernists are guilty of so many crimes, it's hard to
>know where to start the bill of indictment. First they overemphasize
>subjectivity, then they kill the subject. Then they sow uncertainty,
>only to turn into cops. And then the minute they act like cops, they
>become stalkers. This postmodernism thing is so confusing. Bring back
>the old days when men were men and the Kremlin was full of godless
>commies!
>Doug
Hmmmm .... it's quite clear that this issue is never going to be resolved merely by all these arty types talking about the meaning of "stalking" and what have you. We need the firm objectivity of a trained econometrician, just like we do with beauty and art and stuff. So here we have:
Mr Davies' Bloody Awful Statistical Research ----------------------------------------------
ABSTRACT:
The aim of this study was to discover whether T Byfield had been stalking Yoshie Furuhashi. In this context, stalking is to be defined as the statistical significance of a "harassment function" including persistence and interrogativeness.
METHODOLOGY:
Using the lbo-talk archive, we analysed all 41 of T Byfield's posts for the month of December. We sorted them into four groups, based on their "Yoshieness" (presence or absence of the name "Yoshie", or forwarded text from Yoshie), and their "interrogativeness" (presence or absence of a non-rhetorical interrogative sentence). Borderline cases were randomly allocated by flipping a coin, a procedure which should not bias the results.
Since there were 130 active participants in December [1], we would expect each participant to be addressed by 0.315 of a post by T Byfield. However, there is obvious variance in the cross-sectional frequency of addressing -- some posters typically talk to each other more often than others, while the majority only occasionally post, so this estimate may be seriously biased in an unknown direction. However, we decided to use it as the most scientific estimator, pending a more detailed study.
RESULTS:
The Byfield posts broke down as follows
Yoshie, interrogative: -- 4.00 Yoshie, non-interrogative -- 7.00 Non-Yoshie, non-interrogative -- 25.00 Non-Yoshie, interrogative -- 5.00 [2]
Thus, total Yoshie-directed posts were equal to 26.8% of the sample, with interrogative Yoshie posts equal to 9.75%. Furthermore, the interrogative/non-interrogative ratio for Yoshie posts was 57.1%, while the interrogative/non-interrogative ratio for non-Yoshie posts was 20.0%. It would seem that Yoshie is more likely to be posted to by T Byfield than a randomly selected non-Yoshie, and having been posted to, is more likely to recieve an interrogation.
SIGNIFICANCE OF RESULTS:
Without standard deviation estimates, it is difficult to know whether these results have statistically significant. It is also not possible to use the normal probability tables, as t byfield's posts were not distributed across recipients in accord with the normal probability density function. Nevertheless, we estimated a standard deviation of 2 [3], and treated the non-normal data as if they were in fact normal. From this process, we gained the result that the frequency of posting was significant at the 5% level and the frequency of interrogation significant at the 1% level. The "coefficient of stalkage", equal to (2*interrogative)+(0.75*posts) [4], was judged significant at the 5% level.
CONCLUSION
The evidence we have seems to suggest that stalking is indeed being observed. However, numerous methodological weaknesses may mean that our results may not be robust.
Comments are invited, and should be sent to dev/null at flemings.com
[1] Strictly, 131, corrected for double counting of "kelley" and
[2] Including "What's your point?" in a context in which the rhetorical status of the question was not immediately apparent.
[3] Estimated numerically using our bootstraps by the methodology detailed in Guto Penderyn's working paper "Inconvenient Assumptions and How to Avoid Them", Board of Trustees of the Social Security Trust Fund, 1928 (reprinted in "Great Welsh Statisticians of the Interwar Years", University College of Llangefni Press, Huw Price ed.)
[4] Taken from "Stalking and Concealed Weapons: A Free Market Solution" by Ieuan Griffiths, Welsh Journal of Law & Economics, Aberystwyth University Press, 1999
___________________________________________________________________________
_____
---------------------------------------------------------
This email is confidential to the ordinary user of the
e-mail address to which it was addressed. If you are not
the intended recipient, please notify the sender
immediately on (44) 171 638 5858 and delete the message
from all locations in your computer. You should not copy
this email or use it for any purpose, or disclose its
contents to any person : to do so may be unlawful.
Email is an informal method of communication and is
subject to possible data corruption, either accidentally
or on purpose. Flemings is unable to exercise control
over the content of information contained in
transmissions made via the Internet. For these reasons
it will normally be inappropriate to rely on information
contained on email without obtaining written confirmation
of it.
----------------------------------------------------------