fisherdvm:
I am sorry, I had only 3 statistics courses, statistics for engineer, statistics and epidemiology, and a 500 level statistic class.
I bow to you, Josh...
OK. Good. Lets not however assume that other people also do not have similar experience.
Now, the original "cut and paste" was comparing the *relative risk* of activities. I agree that no values were given (man hours/days weeks), and no odds ratios were given either, as were no absolute values. In this case, however, we *can* assume that, as its a comparison, the "units" were the same. There is no reason to assume that diving was calculated in "seconds diving" and being a passanger as "man years in a car", otherwise it *would* be impossible to compare. So, for us to assume that the calculation was done in terms of "time per activity" is fair and sound (in the absence of further data).
Now, the analysis of risk (and relative risk) is VERY well studied, and the statistics of calculations of this is also very well know. In terms of a rigorous analysis,then yes, the cut and paste is perhaps only slightly better than worthless. It would be interesting for us, to get a feel for the data to know the sample sizes, the definitions of "risk" and "safer", how the data was collected (retrospectively, prospectively) what are the odds ratios etc etc etc. We would also like to know the source of the study being quoted. However, what we got was what we got.
That being said, however, the main problem in this thread is the "throwing the baby out with the bath water." What *solid* data there is, gets cited when it supports our arguement, and trashed when it does not. Thalassamania says that as we dont have *full* data, then we can't trust *any* data:
Thalassamania:
Fisherdvm has it right. The problem with diving fatality/accident stats has always been that the numerator is always an incomplete value and the denominator has always been totally unknown. I trust an expert, "I guess that ..."
Is that correct? Well, no. Of course it is not. Statistsics is, largely, about analysis of incompete data sets to enable us to draw conclusions about the whole data set with a certain degree of assurance (remember that bit fisherdvm?).
So, what data *do* we have?. We have, with a certain degree of assurance
1. The number of reported accidents per year. Now is this *all* accidents per year? No of course not. Is it a "selected set" of data? To a large degree, yes. Between years, is there *any* evidence that the percentage of all accidents reported, or the *type* of accidents reported varies significantly? No, there is no evidence that the percentage of all accidents, or the type of accidents reported changes. So we *can* use this data as at least *representative* of the number/type of accidents in diving each year.
2. The second type of data we have is the number of certifications each year. Do we have *exact* data on all certifications every year, and does anyone collate the various organizations together to provide a total number of certifications each year? No. Can we use, say, *one* organization as a guide? Yes, again, in the absence of data that says that that organization does something *markedly* different from other organizations that affects year to year reporting, we could use the data from one organization to give us an *idea* of where the whole market segment is going.
OK. We have two *representative* pieces of data. Neither set is *complete*, and both sets are merely representative of the whole of the diving industry. Both sets *are* however relatively "solid" numbers.
From those two data sets it is clear, despite any thing else going on that diving is getting SAFER. This is explicitly stated by DAN, and there is no evidence *anywhere* to the contrary.
NOw, Mike says:
MikeFerrara:
You're still focusing on accidents.
Yes, mike. To an extent. Why? Becuase its the only area where there is *some* REAL data.
and then:
MikeFerrara:
About 100 reported to DAN, but there's the thousand injuries. Then there all those that aren't reported to DAN. Then as has already been pointed out, we have no idea how many active divers there are or how many dives are being done
As said above mike, you do not need a "full" or "complete" data set to draw valid conclusions about the whole data set.
Now thalassamania also said that:
Thalassamania:
I trust an expert, "I guess that ..." over the self serving estimates of the industry (that unfortunately includes DAN in this matter).
That, frankly is just sad. Some one who bases their opinion on no data, or upon no factual basis is not an "expert". They are, at best, a good guesser and at worst a blowhard. To provide a convincing arguement, one that will convince the doubtful - you need DATA. Real data, and not "I guess".....or "I think......." or "I believe....". As a side point, it would be interesting to know the basis by which Thalasamania proposes that DAN are "self-serving". They are in the buisness of selling dive insurance. It would be more in their interest to make diving out to be more dangerous (to sell more insurance) rather than *less* dangerous.
OK. When it was pointed out that DAN clearly and explicitly stated that diving was safer nowadays, Mike switched his point and said that low standards are evident by murking up the bottom and damage to the coral. Fine. Only problem is, we have shifted from an "objective" assessement to a "subjective" assesement. The second problem with that, is unless those students and divers that Mike see's have thier dive affiliation blazoned across their wet suit, its difficult to see how mike can start to point the finger at specific organizations. However, be that as it may, however experienced mike is, however "right" he is, he is still basing it on a personal opinion. Again, we come back to the difference between "objective" and "subjective". If Mike can find some way to *quantify* what he is saying, then perhaps more people will listen to him seriously. I am not saying Mike does not have a point (and the mid water skills is one that I personally am taking on board), but, without some form of quantitation its just one mans (or more) hobby horse.
Now Mike also says explicitly:
Thalassamania:
Part one of the problem ...Around here our local dive sites are so silted out and torn up when divers are in the water that it isn't worth diving. I'd like to clean that up so we can all enjoy it more.
Part two...Many of the areas that the divers rototill through are covered in spawning nests for large percentage of the diving season. Those nests and the eggs in them get destroyed by the divers wallowing in the bottom.
OK, you make two points. lets take point 2, as thats the easiest to tackle. Ecologically, that habitat destruction due to divers is gonna be relatively easy to quantify - if you are prepared to put some time and effort into it. Get some 1 meter length wooden slats, make them into squares (PVC pipes would do as well). Pick two areas, one frequented by the type of diver you think are a problem, one not (for whatever reason). You want those two areas to be as similar as possible, with the only variable being the divers in one. Put the square down, and *count* the number of your indicator species (nests/eggs whatever). Do that over as long as you can, as regularly as you can. Two season? Three? Now, at the end of your study, *compare* your two data sets. Now you can start to say "divers in this area are impacting significantly, indicator species show a XY% reduction in ..... blah blah blah". You could video actual divers impacting upon the dive environment to enfoce your point. Get your divers (and students) involved. Underwater photography, videography, marine concervation, underwater naturalist etc etc all rolled into one neat little ball. From what I recall, Thalassamania is supposed to be some sort of Marine biologist or other, he could be giving you far more pertinent advice on how to *document* the habitat destruction than I can. If you start from a position of *real* data, people *are* going to listen to you. Ranting on boards like this may be fun, but, in the final analysis, its rather pointless. So, Mike, do you wanna blather on about standards, or do you want to *do* something? Your call.
-j-