[Prev][Next][Index][Thread]

(meteorobs) Statistics & Meteor Science [long]



OK OK.. one last try... here is my complete post...
    I just have to respond to Malcolm's statement    "Why do Americans
like statistics so much,  and treat them as precise figures rather than
an approximation to the truth".  Is that first part intended on being a
criticism?  Whatever is was,  the second part of the statement is
partially false and misleading.   Give us folks on the other side of the
pond more credit than that.  Don't judge us Americans by what you see on
via American TV, ol' boy.  we're not all that dumb ;->
     Statistics and Probability Theory was what I studied as an under
grad... I don't know who 'most Americans' are,  but lets set the record
straight so that list readers are not mislead:
     Statistics are not ALWAYS approximations or estimates.  Yes, when
Dr Jenniskens present data on the slope of meteor activity curve (B) or
Marco gives us his calcs on zenith hourly rates using his '97 LEONID
data... well, yes,  these is a statistical estimates and should be
treated as such... they are the best  approximations (unbiased and with
minimum variance) of true unknown population parameters.   (In fact,  I
have recently seen the traditional ZHR given as "EZHR"..  ESTIMATED
Zenith Hourly Rates.  Good.  That is what it should be notated as
because that what it is)
     On "precision"... quite to the contrary Malcolm,  statistical
estimators ARE  precisely calculated figures... Means, medians, standard
deviations, modes, interquartile range, etc etc.. Aren't these precisely
defined quantities?   yes, they are.   We then use these precisely
defined quantities (or estimates) to make inferences about related
(meteorical) population parameters.   the "truth".as Malcolm states..
Actually, if we want to be fussy...  means, medians etc are technically
referred to as "estimators" (mathematical functions).   And we then use
empirically obtained data (e.g.,  visually-obtained meteor data) to plug
into these estimator fucntions and, viola!. we  generate a (statistical)
estimates...  Now thats being precise..
   In Bobs case, he was giving us rounded off data on his actual counts
so that non regular meteor observers could get an idea/impression on the
frequency with which a regular meteor observer sees stationary
meteors... A very important distinction should be made right here.
Bob's numbers are PRECISELY TABULATED FREQUENCY COUNTS, not statistical
estimates of meteorical population parameters  (although he may use them
to ultimately make such inferences) ...  This follows from most
fundamental distinction between two types of statistics, descriptive and
inferential.  Lets not confuse to two.
   It sure seems to me Europeans like statistics very much as well. Take
a look a the elegant web site of the Dutch Meteor Society managed by
Casper et  al.  Following a meteor shower, site visitors are treated to
a wealth of beautifully arranged statistical tables of numbers giving
estimates of standard meteor parameters.  The IMO presents wonderful
activity curve data based on world-wide visual observations.   Meteor
statistics galore.. dont ya just luv it!
    In closing,  some friends of mine make fun of me when we observe a
meteor shower ("Oh there goes Joseph, of course, he has to COUNT the
meteors").  They are non-scientific types who don't understand why
Joseph can't just lay back a enjoy the phenomena.  Well, I happen to
enjoy science and the scientific method.  Where would meteor science be
today if Dr
Denison Olmstead didn't "like statistics so much"  and  didn't compile
and formally analyze the numerous accounts and counts of the great 1833
Leonid Storm?

     That is all

     Best regards to all,

    Joseph Assmus
    Office of Sleep Studies, UCSD
    San Diego, CA  USA


ps.. Mark, I sure hope those Padres pound the heck out of the Braves...
I think we Pad's have the better statistics