OP
I believe you missed the context there - significant difference between the two gas mixtures . . .
Deco For Divers, Ed 2, And Speed Of Helium Off-gas
- Thread starter Jax
- Start date
-
- Tags
- helium off-gassing trimix
Please register or login
Welcome to ScubaBoard, the world's largest scuba diving community. Registration is not required to read the forums, but we encourage you to join. Joining has its benefits and enables you to participate in the discussions.
Benefits of registering include
- Ability to post and comment on topics and discussions.
- A Free photo gallery to share your dive photos with the world.
- You can make this box go away
I believe you missed the context there - significant difference between the two gas mixtures . . .
In this context:
The bolded portion could be taken quite literally, as you have done.
Or, it could be taken as actually meaning "no significant difference in risk of DCS between the two gas mixtures."
As the gases being compared were O2/He vs O2/N2, I don't see how anyone would mean it in the literal sense. An O2/He mix is clearly significantly different than an O2/N2 mix.
Also, the context of the post is whether using He to replace some or all of the N2 in a breathing as affects the required deco, I feel like the only reasonable way to interpret the bolded part of @David Doolette's post is the latter way that I wrote. I.e. there was no difference in risk of DCS found, in the study he referenced.
I certainly could be wrong. I'm just saying that that is how I did interpret it and I am still inclined to interpret it that way until/unless he comes back and clarifies.
And, if I am correct in my interpretation then, as I said, I am surprised that 1:32 and 6:32 are considered to be not significantly different.
I understood the same as you Stuart, what is then the criteria of significant ?
Sorry for my ignorance but what I can see written, and that is what I see, not saying that it is what happen ( as I was not there when those tests were done ), but it seems that after 3/16 they didn't want to go for another 16 dives to complete the 32 done with He, that way you are not comparing equal numbers, by shortening the numbers of those dives, It appear they created a hypothesis and not a fact.
Sorry for my ignorance but what I can see written, and that is what I see, not saying that it is what happen ( as I was not there when those tests were done ), but it seems that after 3/16 they didn't want to go for another 16 dives to complete the 32 done with He, that way you are not comparing equal numbers, by shortening the numbers of those dives, It appear they created a hypothesis and not a fact.
Last edited:
huzenhagen
Registered
US Navy uses heliox so the test is set up to see if using nitrogen would result in superior decompression (as bulhmann's model would suggest). Once they reached a certain number of events in their trial (dcs) you can statistically determine that there continuing further will not show that your test intervention (nitrox) is superior to your control (heliox).
While it may have been interesting to us if they'd carried on the study it achieved the aims it was designed to do and so was halted to avoid any further harm to the participants.
On the 1/32 and 3/16 while that looks like a hugely different event rate given the small number of participants the study has lower power and so that difference may not be statistically significant
huzenhagen
Registered
So this is where the statistics used in scientific studies and 'common sense' often disagree. This was a small study, a 20% difference with several hundred participants would be highly significant. Because it is small there is a greater probability any observed difference is due to chance.
Take flipping a coin. We would assume that in a fair test heads should show up half the time but because there is an element of chance you sometimes get runs of the same side. If you flip a coin 10 times and get 7 heads you'd probably say that was just random chance despite it being 30% more than tails but if it's 70 out of a hundred (or 700/1000) you're going to start wandering just how the coin has been loaded.
Most studies use whats called a p value. Roughly speaking it says what is the probability that this magnitude of difference in this sample has been seen when there actually is no difference in the general population. Most studies take a value of less then 5% (p<0.05) as being statistically significant.
You can calculate it for this study here: GraphPad QuickCalcs: Analyze a 2x2 contingency table.
you'll se the p value is 0.1014 so double the minimum to be considered significant
For all I understood, then we don't know a fuc.... thing what is better over the other, while one say it was not significant, then one understand that it is Ok or almost the same, while apples to apples where not compared.
Key words you use, model would suggest, sometimes, assume, probably, probability.
And statistics had been proved not to be 100% accurate.
Non of them translate to FACT.
We still don't know facts and probably will never know, as no sane group of divers scientifically proved similar individuals will offer them self to play lottery with DSC, to much money to much risk, we still are wandering in grey waters.
We can use one algorithm with success as well as a mix but you still can get DSC, but at least one try to follow the Algorithm and mix choices that had less incidents and learn from others bad experiences, that is part of self preservation, it will be kind of stupid to go in the opposite direction, because of maybes, assumptions, probabilities.
You cross a river using 2x2 timbers and it had broken 5 times in 20 crossings, then you find 4x4 timbers and it broke 1 time in 20 crossings, what does your common sense tell you which one is better over the other and what will be your preferred choice ???
Key words you use, model would suggest, sometimes, assume, probably, probability.
And statistics had been proved not to be 100% accurate.
Non of them translate to FACT.
We still don't know facts and probably will never know, as no sane group of divers scientifically proved similar individuals will offer them self to play lottery with DSC, to much money to much risk, we still are wandering in grey waters.
We can use one algorithm with success as well as a mix but you still can get DSC, but at least one try to follow the Algorithm and mix choices that had less incidents and learn from others bad experiences, that is part of self preservation, it will be kind of stupid to go in the opposite direction, because of maybes, assumptions, probabilities.
You cross a river using 2x2 timbers and it had broken 5 times in 20 crossings, then you find 4x4 timbers and it broke 1 time in 20 crossings, what does your common sense tell you which one is better over the other and what will be your preferred choice ???
Of course not. Statistics is a tool to help make sense of variability in a data set. Biological systems (e.g. humans decompressing) are messy, variable systems. The statistically non-significant difference between 1/32 and 3/16 means that these results are highly unlikely to be different from any random outcome. For example, the results could have easily been 3/16 versus 1/32 in the other direction had the study been repeated.
Want to make your statistics tighter? Find 200 people who are willing to join each of the treatment groups and risk bending themselves. Then get them to repeat this 10 times. That would give us a lot better information, but Remy as you mentioned, good luck designing such a study.
...so we do the best we have with the available information.
archer1960
Contributor
You'd need a few dozen volunteers to get anywhere near statistically meaningful numbers and I very much doubt a chamber big enough even exists.
Similar threads
