41- New Information
Skydivejg
had from a presentation I gave at the PIA Symposium in 1995 when I was Director
of Safety & Training for USPA. I had a list with all the recorded US fatalities
from 1963. The list is copied below with updated totals from 1995 through 1997.
The worst year for fatalities was 1981 with 56 skydivers killed. 1976 & 1979
were next worst with 55 killed each of those years. 1977 had a total of 50.
These were the only years with 50 or more fatalities. The period from the early
70's to the early 80's was one of the worst periods in the history of the
sport. This has been attributed to the rapid influx of new gear to the sport
during this period and the changes it required jumpers to go through with our
emergency procedures. There was also a learning curve that the manufacturers
went through as they refined their new designs and learned from the problems
that jumpers encountered.
Total USA Fatalities by Year
63- 34
64- 31
65- 25
66- 28
67- 31
68- 23
69- 39
70- 30
71- 39
72- 34
73- 44
75- 41
76- 55
77- 50
78- 48
79- 55
80- 47
81- 56
82- 29
83- 29
84- 35
85- 27
86- 31
87- 28
88- 23
89- 36
90- 23
91- 30
92- 27
93- 41
94- 30
95- 26
96-37
97- 32
The total number of skydivers killed from 1963 through 1997 is 1226!!!! That's
a lot of our friends who aren't out there any more. There seems to be a new
trend in increased fatalities in recent years. What can we attribute this to?
Think about what you can do to help keep the numbers down.
And hey, let's be careful out there!!!!!
Jack Gregory
mave...@hal-pc.org
training and gear changes (like maybe AFF, and SOS, and more docile
squares) that took place between 80 and 83:
80- 47
81- 56
82- 29
83- 29
That's quite a drop there, and from the following years, it really looks
like it reset the baseline. Any college students taking statistics out
there? This is pretty juicy stuff.
Wendy W.
Dave Briegs
total jumps in each of the years mentioned.
Dave
http://www.concentric.net/~bodyflyr/
ynot...@my-dejanews.com
skyd...@aol.com (Skydivejg) gave an excellent synopsis of historical
skydiving fatalities in the U.S.
Thanks so much, Jack, for putting these things in an historical perspective.
Does anybody have similar information regarding the skydiving community
worldwide?
Thanks,
tony
______________________________________________________________________
"The marvels of today's modern technology include the development of a
soda can which, when discarded, will last forever ... and a $27,000 car
which when properly cared for will rust out in two or three years."
-----------== Posted via Deja News, The Discussion Network ==----------
http://www.dejanews.com/ Search, Read, Discuss, or Start Your Own
bil...@qualcomm.com
mave...@hal-pc.org wrote:
> Looking at those numbers, I'd be REAL interested in some of the
> training and gear changes (like maybe AFF, and SOS, and more docile
> squares) that took place between 80 and 83:
>
> 80- 47
> 81- 56
> 82- 29
> 83- 29
be careful drawing conclusions from data like this. it may have just been
raining a lot that year (fewer jumps, more clear and pulls, softer dirt.)
-bill von
Scott Schnabel
fatalities in 1981 and made a concerted effort for better training and
increased awareness. As I recall an inordinate amount of fatalities were
students and people with under 100 jumps. Looks like the same effort is
needed now in regards to canopy control problems. Anyone remember the
recurring incident where the student would pull the cutaway, realize he
screwed the pooch and dump the main, only to have it leave him at line
stretch?
Scott
D-10977
mave...@hal-pc.org wrote in message <363633...@hal-pc.org>...
>Looking at those numbers, I'd be REAL interested in some of the
>training and gear changes (like maybe AFF, and SOS, and more docile
>squares) that took place between 80 and 83:
>
>80- 47
>81- 56
>82- 29
>83- 29
>
ynot...@my-dejanews.com
mave...@hal-pc.org suggested a statistical analysis of the data.
This, I think, is a good idea, but it's going to take a lot of work that isn't
apparent when you see a simple time series of data. There are so many other
things (independent variables) that could effect the numbers.
One would certainly want to know as much detail as possible about each
fatality, with some indication as to probable cause such as gear, operator
error, weather, medical status, etc. Not all of the fatality reports that we
read even today have this information clearly stated, sometimes because it's
not clearly known. I hate paperwork, but a standardized fatality checklist
might prove useful for this sort of thing in the long run.
Consider that an earlier cutaway and reserve deployment may have prevented a
fatality. Did the presence of Capewells affect the cutaway time, or did the
operator just blow it? Did the increasing prevalence of Bill Booth's 3-ring
release system affect the fatality rate in the early 80's, or did stricter DZ
enforcement of BSR's minimum pack-opening altitude have anything to do with
it? None of these things are mutually exclusive, either. Perhaps you can see
that we may be able to consider enough independent variables to attempt to
describe a statistical solution "space", instead of a solution "line", with
number of fatalities on the Y axis and time on the X axis.
To get this historical information would be a tremendous feat. An intelligent
analysis of it would be an even greater thing. It would be a paramount naiveté
to speculate in its absence, but if we can't get those data, or that kind of
analysis, then speculation *will* happen. You bet.
One question, though: What would we actually gain from this excercise?
tony
__________________________________________________________________________
"Today is the tomorrow you worried about yesterday."
Liv4fear
>total jumps in each of the years mentioned
Great point Dave! The percentage would certainly be much more informative. Not
only that, there's the issue of the cause of death. There have probably been
trends based on changes in gear and other factors.
Another point of interest is the the demographics of those who are dying. For
example, the average age for '97 was ~42 with the average number of jumps
~1000. It would be interesting to see how such variables have changed over
time.
The analysis would be easy. Hell, I have the stats program. It's the gathering
of data that would suck. I dunno if USPA has always kept such good records of
such incidents...
Oh well, back to work...
Blue Skies,
Mark
mave...@hal-pc.org
>
> Actually, the jumping community in general was shocked by the number of
> fatalities in 1981 and made a concerted effort for better training and
> increased awareness. As I recall an inordinate amount of fatalities were
> students and people with under 100 jumps. Looks like the same effort is
> needed now in regards to canopy control problems. Anyone remember the
> recurring incident where the student would pull the cutaway, realize he
> screwed the pooch and dump the main, only to have it leave him at line
> stretch?
>
> Scott
Yes; actually, I was an instructor in those days (though I'm no longer
active), and I remember the shock, and the fatality report from that
year. It would take a lot more than what we have hear combined with a
few guesses to get a real statistical analysis, but, from memory, I can
assure everyone that the number of jumps in the 80's didn't go down that
much from the 70's, and based on the numbers Jack posted, it was
steadily climbing through the 60's and 70's, then reset at a lower
level, and has been climbing generally more slowly since, but hasn't
really reached its previous level.
We no longer have belly bands on old gear to twist; we no longer have
ripcords on B4 containers with cones to get hard pulls from; we no
longer put students out on cheapos into tight DZs (though the number of
steering fatalities was never high), etc. etc.
Wendy W.
Dave Briegs
mave...@hal-pc.org wrote:
>
> Yes; actually, I was an instructor in those days (though I'm no longer
> active), and I remember the shock, and the fatality report from that
> year. It would take a lot more than what we have hear combined with a
> few guesses to get a real statistical analysis, but, from memory, I can
> assure everyone that the number of jumps in the 80's didn't go down that
> much from the 70's, and based on the numbers Jack posted, it was
> steadily climbing through the 60's and 70's, then reset at a lower
> level, and has been climbing generally more slowly since, but hasn't
> really reached its previous level.
<snip>
wait a minute. are you saying that there are LESS jumps made per year today
(around 3.5 million or so) than there were in the 70's? i find that hard to
believe. am i psycho?
Dave
http://www.concentric.net/~bodyflyr/
Skydivejg
>(around 3.5 million or so) than there were in the 70's? i find that hard to
>believe. am i psycho?
>
>Dave
I think many people are missing the point about my original post. I didn't post
the fatality figures for the last 35 years to provide a basis for statistical
analysis. I posted them to give everyone an understanding of the dangers of the
sport and the need for constant vigilance in the area of safety. Don't get lost
in how many jumpers were killed per how many jumps. The bottom line is that
from 1963 to 1997 there were 1226 sport skydivers recorded as killed in the
USA. Think about that for a moment. 1226 Killed!! That's the statistic that
really matters. Instead of talking about what % of this and what % of that,
please remember the closing words of my first post-
Dany Brooks
13,500 USPA members, but I assume that back then there were far less
members in relation to total jumpers than there are now. It's steadily
grown from there to the current membership total of 33,000, but it's also
estimated that there were 310,000 first jumps last year with few of them
actually going on to be members (tandems?).
Some other years for membership: 1986 = 16,400; 1990 = 20,000; 1995 =
30,000 (huge growth here).
Dany
Dave Briegs <sky...@erols.com> wrote in article
<363638E2...@erols.com>...
> All these numbers are kind of useless without the corresponding number of
> total jumps in each of the years mentioned.
>
> Dave
> http://www.concentric.net/~bodyflyr/
>
>
Mary Jo Robinson
the majority of the deaths were related to each year and possibly the minority to
see what trends have progressed, recessed, subsided and re-surfaced. JMHO.
Fly free...
Paul A-30933
Skydivejg wrote:
> 80- 47
> 81- 56
> 82- 29
> 83- 29
> 84- 35
> 85- 27
> 86- 31
> 87- 28
> 88- 23
> 89- 36
> 90- 23
> 91- 30
> 92- 27
> 93- 41
> 94- 30
> 95- 26
> 96-37
> 97- 32
>
> The total number of skydivers killed from 1963 through 1997 is 1226!!!! That's
> a lot of our friends who aren't out there any more. There seems to be a new
> trend in increased fatalities in recent years. What can we attribute this to?
Kevin Keenan
because of multi-occupant jump plane crashes, some with fairly sizeable loads. While
the argument can be made that these too are "skydiving fatalities", I think it's
obvious that they are different from a freefall or canopy-related accident.
The numbers really do take a bit of studying before drawing conclusions. However the
fact remains that "shit happens" and we must all take care that it not happen as
much.
kevin keenan
Spad
plane crash fatalities were NOT classified as skydiving fatalities. My
personal opinion is that they ( the plance crash victims) WERE skydiving
fatalities. Why were they in the plane? If a dz operator fails to maintain
the aircraft or a jump pilot makes a mistake in judgment or a skydiver fails
to pull in time, the result is the same and WE need to address the ISSUES
that relate to US, not the damn semantics or "stats."
Numbers are numbers, bodies are bodies. The FAA is going to get tired of the
latter and then the "gig" is up. You DON"T have a right to use the airspace.
Spad
D-3904
Kevin Keenan wrote in message <363802CE...@atlantic.net>...
Rita
> wait a minute. are you saying that there are LESS jumps made per year today
> (around 3.5 million or so) than there were in the 70's? i find that hard to
> believe. am i psycho?
You're not psycho ... but I would imagine there is a factor today that would
have to be taken into consideration. You've probably got a lot more jumps
being made today, but loads of them are tandem jumps. The student has a
highly experienced canopy pilot along ... hence, there should actually be
less fatalities today than there were in previous years ... unless my
reasoning is off?
Blue ones!
--rita
Rita
>
> Don't have any totals other than membership numbers. In 1976 there were
> 13,500 USPA members, but I assume that back then there were far less
> members in relation to total jumpers than there are now. It's steadily
> grown from there to the current membership total of 33,000, but it's also
> estimated that there were 310,000 first jumps last year with few of them
> actually going on to be members (tandems?).
Shouldn't use the tandems, though, in calculating the fatality statistics.
The same "odds" would not be present for death with a tandem. Sure you could
die, no question, but your chances for a "successful" jump are far greater
with tandem than with another type of jump.
Students and newbies make mistakes ... because of their inexperience. We make
faulty judgments. Tandems don't have that inexperience factor to deal with.
Tandem students are also jumping with some of the most experienced canopy
pilots on the dz. Hence, at least to my reasoning, their overall risk would
be lower.
Anyone care to tell me how much of an asshole I am? :)
Blue ones!
--rita
Skydivejg
>One thing to consider is this; Several of those years with large totals got
>that way
>because of multi-occupant jump plane crashes, some with fairly sizeable
>loads.
The fatality figures I supplied did not include any skydivers killed in jump
plane crashes. These were all people who had jumped from the aircraft.
Jack Gregory
Thomas Kerler
> die, no question, but your chances for a "successful" jump are far greater
> with tandem than with another type of jump.
that's probably not quite true ... at the least the "far" part.
In average there are about two tandem fatalities per year
(I think this year there were even four or so)
Obviously those number are (fortunately) too small to make
a meaningful statistical analysis - but with maybe a couple
of 100,000 tandem per year the risks seem to be in the same
ball park as with other jumps.
Tandem rigs are a good deal more complicated than standard
rigs and the configuration of two people in the same harness
opens up a few more possibilities to screw up. And even the
most experienced tandem masters are fallible.
--
Blue Skies, Thomas D-20874 Muff#2160
http://www.osu.edu/students/skydiving/
http://www.skydiveohio.com/
Some Basic RW-Skills:"... To be more to the point: If you
don't have anything positive to say, then say nothing.
If you cannot come across with good vibes, then keep the
negative things to yourself. Don't spit in the soup that
you'll have to eat later. ..."
Pat Works (The Art of Freefall Relative Work)
Elladan
Rita wrote:
> Dany Brooks wrote:
> >
> > Don't have any totals other than membership numbers. In 1976 there were
> > 13,500 USPA members, but I assume that back then there were far less
> > members in relation to total jumpers than there are now. It's steadily
> > grown from there to the current membership total of 33,000, but it's also
> > estimated that there were 310,000 first jumps last year with few of them
> > actually going on to be members (tandems?).
>
> Shouldn't use the tandems, though, in calculating the fatality statistics.
before anyone jumps to the same conclusion i almost did, she means using tandem
jumps in calculating jumps/death.
> The same "odds" would not be present for death with a tandem. Sure you could
> die, no question, but your chances for a "successful" jump are far greater
> with tandem than with another type of jump.
we could argue that point to death, but let's make it easy.if it's twice as safe
the ratio of jumps/death hasn't changed so ya tandem jumps should count
>
>
> Students and newbies make mistakes ... because of their inexperience. We make
> faulty judgments. Tandems don't have that inexperience factor to deal with.
> Tandem students are also jumping with some of the most experienced canopy
> pilots on the dz. Hence, at least to my reasoning, their overall risk would
> be lower.
>
> Anyone care to tell me how much of an asshole I am? :)
no way, you're our FRIENDwe want you to be HAPPY
we wouldn't want you to be SAD
CRW skies
paul j
Spad
necessarily safer) when you're on student status and are accompanied by an
Instructor and/or Jumpmaster, no doubt about it. What about later when
you're off of student status? In other words, exactly WHO is getting killed,
from WHICH program, WHEN and under WHAT CIRCUMSTANCES (what part of the
skydive)?
AFF is a good program, in theory, but from what I've seen in it's
application there is a tendency to emphasize the "here, now, fun & FREEFALL"
part of skydiving and to de-emphasize the
survival-skills-once-the-canopy-opens part. The latter is time consuming to
teach and only adds to the "ground" time that produces little or no income?
It's not a "positive" learning experience or curriculum (until you need
it!)? It's "negative" in it's approach to the sport and tends to turn people
away? I've heard all these questions & arguments at one time or another when
the survival skills are "short-changed" in a dz's instructional program.
What's my point?
I believe 46-50% of the fatalities this year (and maybe in other years?)
occurred after deployment of the canopy. Is there a common thread as to the
source of these tragedies that relates to a method of instruction and what's
not being taught or emphasized? Is there a pattern in the percentage curve
that relates to the introduction of particular discipline or change of
rules, techniques and/or presentation within a discipline?
This isn't a "slam" on any particular discipline. It's a question I would
like to see explored in depth without anyone getting defensive or paranoid.
Since injuries go largely unreported, we may get a skewed picture doing what
I have suggested using only fatalities but you have to start somewhere. IF
those who have the needed information will provide it, then we might, just
might, fix it.
The MOST dangerous part of this sport has ALWAYS been AFTER you open your
canopy. Ask any life or health insurance underwriter. They've got some
really "hard core" statistics but it costs to get companies to search &
provide them.
"If the student failed to learn, the Instructor failed to teach." That
"retro" phrase needs to be touted again more often. That's for sure.
Thanks to Jack Gregory for providing the numbers (although each number
represents a tragedy, some personal to me) and I hope he keeps providing us
with more of his past insights.
Ya know? This is the best "thread" I'v seen on the NG in a long time! :-)
Spad
D 3904
Dave Briegs wrote in message <36387D72...@erols.com>...
If I recall correctly, someone examined the numbers in the last year or
two and found that "statistically", tandem jumps are NOT the safest student
jumps. AFF's are.
dave
Dave Briegs
Spad wrote:
> Are AFF's safer? Depends upon the perspective, I think. AFF's are safe (not
> necessarily safer)
In the stats that I remember reading (sure wish I had them!), perspective had
nothing to do with it. AFF had a lower rate of fatalities per given amount of
jumps than tandem did, at least during the period examined.
Peace,
Dave
http://www.concentric.net/~Bodyflyr/
Rita
> In other words, exactly WHO is getting killed,
> from WHICH program, WHEN and under WHAT CIRCUMSTANCES (what part of the
> skydive)?
> AFF is a good program, in theory, but from what I've seen in it's
> application there is a tendency to emphasize the "here, now, fun & FREEFALL"
> part of skydiving and to de-emphasize the
> survival-skills-once-the-canopy-opens part.
This is exactly what I was referring to ... what happens once the canopy
opens?
True, in freefall I can well imagine that AFF jumps would probably be safest.
The student is wearing a pretty much standard sport rig ... a rig that is
relatively simple in design. He has two highly trained and qualified
jumpmasters in the air with him ... jumpmasters who can and will do
everything they can to assure an "uneventful" skydive, up to and including
pulling the ripcord for him. With tandem, you have a tandem master, also
highly trained, physically attached to the student in freefall, but you have
a rig far greater in its complexity. Sure, I can well understand AFF being
the safer of the two methods.
But now you have deployment. Canopy is open and it's a good one. Alright!
Now, where is the greatest margin of safety? The tandem student shares that
canopy with a professional tandem pilot ... someone who has done this sort of
thing many times before and has probably attained a certain level of
expertise. The student receives hands-on coaching ... how to read the winds
... when to hold ... when to crab ... when to run with the wind. He has help
with setting up the landing and help with the flare.
Now take the AFF student. He's under canopy ... maybe for the first time.
Hopefully, he's got assistance on the radio ... someone to "guide" him ...
note the term "guide." That person can only offer advice *if* the student
chooses to take it. What if the student brainlocks? What if the student
panicks ... runs away from the dz ... doesn't listen to instructions? What
if the radio goes totally dead? Now, it's a different issue ... and one that
could spell problems. Just how much canopy control training is offered in
that first jump course? I hate to say it, but other than being told that you
pull the right toggle to go right, and the left one to go left, both to flare
... I don't recall getting very much. Directly before our jump, we were
taken out onto the field ... shown the direction of the winds, and advised as
to our holding position ... our "play" area.
Canopy control training is something that is rather difficult to teach. It's
something that has to be learned ... over the course of many jumps ... many
experiences, and perhaps even some mishaps. Some students pick it up
rapidly, while others (like myself) learn more slowly.
I often wonder if it is even possible to give students a good canopy control
course, other than perhaps putting them into some sort of VR Simulator, such
as Cleveland Parachute Center has. There would seem to me to be no really
reliable simulation of a canopy control experience other than actually doing
it. So the options are either some tandems until the student demonstrates
sufficient knowledge, or a simulator. So perhaps dz's are simply doing the
best they can to teach this skill in the forum of an FJC.
It is for this reason ... the canopy control experience ... that I kind of
preferred my time spent in a static line student progression. All those
canopy descents ... all that knowledge gleaned ... with descents monitored by
a jm on the field, who could not only assist via radio, but could then
critique my landing after the jump.
That is why I still say ... order of potential danger:
Most dangerous: S/L (not because of landing, but because of presence
of the actual static line)
Middle Road: AFF (solo canopy descent ... same as S/L)
Least Dangerous: Tandem (due to the assisted canopy descent)
Blue ones!
--rita
Peter Liemberg
>One question, though: What would we actually gain from this excercise?
--------------------------------------------
Knowledge ?
("Is this policy or did we think it over ?" - anonymous civil servant)
*********************************************
Tandem JM/I DZO; "When I get my Z licence
I probably don't need a parachute anymore..."
Peter Liemberg
>I think many people are missing the point about my original post. I didn't post
>the fatality figures for the last 35 years to provide a basis for statistical
>analysis. I posted them to give everyone an understanding of the dangers of the
>sport and the need for constant vigilance in the area of safety. Don't get lost
>in how many jumpers were killed per how many jumps. The bottom line is that
>from 1963 to 1997 there were 1226 sport skydivers recorded as killed in the
>USA. Think about that for a moment. 1226 Killed!! That's the statistic that
>really matters. Instead of talking about what % of this and what % of that,
>please remember the closing words of my first post-
>
>Think about what you can do to help keep the numbers down.
>
>And hey, let's be careful out there!!!!!
-----------------------------------------------------------
Now I'm lost, I'm afraid.
In order to be careful (as you require me to be), I must fully
understand where the danger lies, which is a bit more complicated than
'the ground is hard and slamming it will take your vital organs
apart'. Since we have first jump courses it is apparent that all the
facts of life while skydiving are not that obvious; why give a course
if we could tell everybody that 'being careful' is an excellent idea &
here's a rig, there's the plane - just wait for the green light...?
Nowhere in the world such a 'course' exists.
Courses are given with at least the collective experience of people
who were there / done that & bought the T-shirt - even as early as in
1963.
Many national organisations are collecting & passing on data in order
to give this 'collective experience' hand & feet.
Sad as it is, the only way to do this is by analysing the data - like
finding out about erratic jump history, borrowed / foreign gear etc.
(lightly loaded Nova's anyone?)
Now to do this in a meaningfull way, you need some sort of statistical
analyses which prefarably includes the number of jumps where all the
mishaps were collected from. In the Netherlands we had one fatality
until now (this year 1998). In the USA...well, we know the website
dont we?
No one is his right mind would suggest from this "you have more
accidents than we" alone that we in the Netherlands are safer than you
in the USA - our collective number of jumps is around 60.000 per year,
which should give you some perspective.
The fact that sound statistical methodology isn't common knowledge for
most of us, doesn't mean that it doesn't exist.
There's an interesting tool called the 'Poisson distribution' for
small fractions (like: most walk away after landing but we see some
fractures...)
There's also a math professor overhere on the newsgroup
(Thomas Kerler <ker...@math.ohio-state.edu>) who probably knows a lot
more about it than I do, but one thing I know about this 'Poisson
distribution' is that it is totaly apliccable to statistical analyses
of many aspects of skydiving and student training; a DZ as well as a
national organisation might end up with reliable figures that tell
them wether this was 'just a bad luck year' or they should 'clean up
their act' in (f.i.) student training, packing requirements or tandem
licencing.
(the mathematician Poisson originaly started out with the question
what the risk was for the stablemen in the Prussian army to get kicked
by a horse while collecting manure...i.e. statistical analysis of
small fractions in large populations or "should the manure be
collected without horses in the stable & is this practically
feasible?")
The first thing you should know of course is the amount of jumps where
all this misery stems from. But make no mistake; it isn't neccesary to
go around and ask all the DZ's how many jumps were made. You can get
quite reliable figures with some sophisticated estimation. (Even
comparing total USPA membership with fatalities is a start, albeit a
rough one...)
We could find out if pull outs are better than throw outs, if AFF is
superior to static-line - even in the long run, wether AAD's are 'good
for everyone' and what to do with a PC in tow.
After all that we could finaly use the newsgroup exclusivly for what
it was invented for...
(sheep jokes anyone ?)
;-)
BTW: Of course it goes without saying that we had one to many this
year...
Jonathan Goodman
thought it was eventually going to morph into something else as threads
often do. However, there genuinely appear to be some people interested
not only in the results of a close examination of whatever data have
accumulated, but also in the methodology of such an enterprise as well.
Perhaps I can aid in bringing to focus such an investigation.
First off, a Poisson distribution is probably not of any use here. When
I explain what this mathematical construct is used for, I bring up an
analogy wherein the bottom of an elevator shaft is covered with small
dishes (some number N) and a man at the top of the elevator shaft dumps
a bag of marbles (some number M). The marbles will fall randomly and
distribute themselves into the various dishes at the bottom. The
question is how many cups does one expect to contain zero marbles? ...
one marble? ... two marbles? ...etc. The Poisson distribution is a
model that describes this type of random process in terms of N and M.
For this reason, I don't see how it can be much use in our analysis
here. I used the Poisson distribution last in counseling medical school
applicants. I used the sheer volume of applicants versus the number of
open slots in US medical schools to determine what percent of people
would obtain zero acceptance offers. (under the assumption that the
distribution of acceptances is random, a false assumption though the
process is so capricious that it SEEMS true)
In fact, the methods of analysis in this case are not going to be
terribly different than the type of research that I'm currently involved
in. I am currently studying the results of a particular type of
reconstructive surgery for a particular type of congenital malformation
at a particular children's hospital. This is a retrospective type of
study; the data have been gathered and my job is to sift through the
material and come up with certain correlations or the lack thereof.
The trouble with our discussion here so far is that it has been treated
as if all one has to do is "put the data in the analysis program" and
out come "answers". This is the same type of unclear thinking that I
observed not too long ago in a protestor who was picketing in front of a
research lab that undoubtedly used some sort of animal model. She
proudly displayed the emblem of "PETA" (people for the ethical treatment
of animals) and several other animal rights oganizations. Being the
pompous asshole that I am, I decided to start a curbside debate with
her. "If medical researchers shouldn't use animal models to study
disease, then what would you propose we do with human suffering?" I
asked. "They can use computer models!" she responded. "Well if they
don't know how the real system works in the first place, (which is why
they study it)
how are they supposed to create an appropriate model in a computer??" I
asked incredulously. "They can program a computer to do ANYTHING now...
" she concluded. I left it at that! She obviously had her mind made
up.
Anyone who reads this NG is obviously versed enough in the use of
computers to know that its just a soldier carrying out orders. The same
is true of a statistical analysis. In order to engage this project
(assuming that the data have been recorded and are correct) one has to
ask a well formed "research question". The more specific the question,
the more likely the data can give you something meaningful. I won't
bore you with the details of deciding "statistical significance", but in
general the more nebulous the question, the more data is needed in order
to answer it.
I think a reasonable first pass at a "research question" here is:
Are the relative numbers of serious accidents per total number of jumps
increasing in any of the following categories over the data gathering
period:
Group 1: skydivers with less than 20 jumps
Group 2: skydivers with between 21 and 50 jumps
Group 3: skydivers with between 51 and 100 jumps
Group 4: skydivers with between 101 and 200 jumps
Group 5: skydivers with between 201 and 500 jumps
Group 6: skydivers with between 501 and 1000 jumps
Group 7: skydivers with greater than 1000 jumps
Thats looks like a nice simple research question that arrives at a yes
or no answer. However, LOOK AT THE SIZE OF THE UNDERTAKING!!! We
would have to collect information about how many jumps were being
performed in each of those categories every year under study! Those
numbers would have to be estimates, leading to a first line of
uncertainty. The collected info about accidents probably contains
numbers of jumps per accident victim (it has since I've been in the
sport) but does it contain data about serious incidents that were
survived? What should be the inclusion criteria in "serious accident"
?? Does one have to die? In trauma research, one has to set out
specific guidelines for inclusion in the study question. For example we
had a study going for a while which involved bullet wounds to the "true
abdomen" defined in terms of bullet entry beneath the ribs, above the
umbilicus, etc., etc. It seemed well defined, but then we'd see patients
shot IN the umbilicus. How do you play that one?
These are the nightmares of a statistical analysis. Especially in a
retrospective study design, where one is left constantly saying "DOH! If
only we'd been gathering THIS number or THAT number for the past 13
years..." one is never left feeling that a satisfactory result is
obtained.
If over time the data base has been kept in a format anything like the
last couple of years (the time I've been in the sport and following the
magazines), then we have information about the time and location of a
fatality, along with the rig used, the victim's licensure status and
logbook total, and a categorization of the type of accident. I doubt
there is anything about serious career ending non fatalities or
significant non-fatal hospitalizations.
We may be able to get something like a USPA official estimate of the
total numbers of jumps per year performed at US drop zones as well.
Additionally there may be estimates of numbers of first jump courses or
tandem jumps by whuffos who never jump again that we could factor in and
out (just for fun) and there is obviously record of numbers of people
advancing through licensure status which could be followed as an
estimate of the community's "jump maturity" (ill defined as that is).
What I propose is something LESS than an analysis. Instead of actually
asking a question of the numbers, lets just look at them in the form of
a chart or graph. We've spent as much time on this thread in terms of
person-hours as it would take to key in the results to a spreadsheet
from the magazines alone, and I'm willing to bet that theres a USPA file
on this that will save us the time. Once we're looking at the chart, we
can start to get fancy "oh look! theres a rise in the number of
accidents due to folks with high jump numbers who hook turn, and it
begins here in the graph... isn't that about when ellipticals became
popular?" (bing! a research question pops out).
The point is, until we're looking at the graphics, this whole enterprise
is a bunch of playground shouting. If theres a USPA rep with the
knowledge of where to find this info (Mr. Gregory? care to save me some
time?) I'd be more than happy to key it in and plot it. If I don't hear
anything through the NG over the weekend, I'll contact USPA on Monday
regarding this.
Recess is over ;) ...Let's do this right.
Jonathan Goodman, MSIV
USC School of Medicine
<jogo...@hsc.usc.edu>
Rita
> The point is, until we're looking at the graphics, this whole enterprise
> is a bunch of playground shouting. If theres a USPA rep with the
> knowledge of where to find this info (Mr. Gregory? care to save me some
> time?) I'd be more than happy to key it in and plot it. If I don't hear
> anything through the NG over the weekend, I'll contact USPA on Monday
> regarding this.
Wow! I am truly impressed here ... no kidding.
Look, I don't pretend to have the intelligence or the background to be of
much help in this kind of an undertaking, but for what it's worth ... if
you need any help with keying or other administrative tasks, just let me
know.
I would love to see the results of a study such as this.
I think we can learn quite a great deal from them too!
Blue ones!
--rita
Spad
Erik
>Perhaps I can aid in bringing to focus such an investigation.
Thank you for your insight. This sport NEEDs serious statistical
research. If numbers from one drop zone can help, let me know what
those numbers are.
Erik Johnson