Ilya Maclean
charlatan
Excellent content, Ilya. Thanks. But on the basis...
Burden of proof lies with whatever is to be proved. If you want to prove existence, that's one thing. There are ways to prove (with some confidence level) that the bird is not present, and in fact this tends to be the better approach to finding whether they are there or not. That's my real point.
For example on the remote cameras; if you wanted to show with some conclusiveness that IBWO is not present, you might do some statistical analysis on what coverage would be necessary, over what time frame, etc, and come up with a plan to more thoroughly show they are not there with some rigorous statistical confidence. Then suggest the searches do this to really find the bird. If they fail, you've got a strong case the bird is really not there. What you might find, however, is that there is no way to get enough cameras in place to do this. And if that can be shown, then you are demonstrating that searchers should probably stop wasting their time with them.
Personally, I think audio has potential to bring this whole deal to conclusion, not video.
You have some good comments on the items, but they are not yet in the direction of proving them not possible.
At the risk of being psuedo-intellectual...
No, you cannot prove the probability of not an IBWO not being present using remote cameras. You can prove the probability of seeing one, and at face value the probability of not seeing one might appear to be one minus the probability of seeing one, but it isn’t and that is the fundamental reason why science proceeds by rejecting null hypothesis and not the other way round! The whole fallacy of proving a negative yet again!!
Typically when analysing sets of data to detect e.g. a difference, you use statistics to work out the probability of rejecting the null hypothesis - i.e. that the datasets are not significantly different from one another. This is not the same as the probability that they are the same. You could get a high probability value because they are similar, but you could equally get a high p-value because your dataset is too small. Thus, a probability value of 0.05 does indicate that you can be 95% sure they are different, but a p-value of 0.95, does not indicate that you can be 95% sure they are the same.
Edit: Having thought about this a bit longer, I realise what I said before is actually bollix. The probability of non-existance is actually the same as the probability of theoritical occurance, but then not seeing. What can't be proven is the probability of occurance, given non-sightings.
Doesn't negate the fundemental point though. While theoretically the burdon of brief lies where ever you want it to be, the accepted norm is that it lies with those making a claim, and for very good reasons.
Last edited: