www.ditl.org

Mikey wrote:

Deepcrush wrote:

Wow, a ban over this is like using a nuke because two guys bumped and dropped their beers.

Well, what kind of beer?

Sam Adams

Meh. Not worth it. @Shran - I don't think these guys have much choice.

Tyyr wrote:

IanKennedy wrote:It's not pedantic to explain what something is to someone who clearly demonstrated he new s**t all about the topic he was preaching on. If you keep up that attitude around here you won't be here much longer.

Ah, so one population figure where I didn't go into the exhaustive detail in explaining confidence intervals as well as the implications there of, where I instead just truncated things so I don't have to write another two paragraphs explaining something you can find with two minutes of searching and I don't know s**t about this? As opposed to what, all the other highly informed posts we get on this subject and others?

Stating the % of the CI without actually giving the values is totally meaningless. Really, utterly meaningless. For example I could say that the mean was 1,000,000 and the 95% CI was 800,000 to 1,200,000. But equally it could be 1,000,000 with a 90% CI of -1,000,000 to 100,000,000. You can see from the first example that they have a reasonable degree of accuracy. The second one they've no idea at all what the value is. I say again if you don't specify the values you are grossly misleading people. The reason I pointed out the error in the first place is that not everyone here knows stats that well, and you where using the 90/95% figure to try and bolster peoples opinions of your numbers. Giving the percentage is utterly useless, it tells you nothing. The percentage used is completely arbitrary, typical convention is to use 90 or 95%. You either give the range or give nothing there's no in between.

That's what you're threatening to ban me for? Or is this threat because I didn't just knuckle under cause your name's in red?

As for threatening to ban you I didn't I merely pointed out that there is a rule here for not being jerk. You where being a jerk. After all this forum is a private area owned by Graham and myself. Membership here is privilege not a right. As the owners yes I do expect a little respect from the members, just as you would from visitors to your home. Is that unreasonable? Do I expect you to knuckle under, No, do I expect you to not be a jerk, emphatically yes. We have already banned one person for that and I would rather not have to ban another.

IanKennedy wrote:Just one question. If you know what a CI is why would you make a statement like this which is clearly meaningless.

tyyr wrote:Actually the population figures I'm getting, from several sources (one of them the IWC itself) have confidence intervals of 90 to 95%. In other words the scientific community is pretty confident in their estimations.

The estimations I've been finding have typically been given with the 90 and 95 confidence intervals and the intervals are fairly tight. Yeah, I didn't go to all the trouble of writing out the intervals, explaining their significance, and doing all the percentages with the intervals and everything else. I cut it down and simplified it for use in a message board discussion. If someone wants to do the research themselves it takes two minutes of searching, if you're a slow typist. And yeah, if you come into a discussion like this and the only thing you've got is to bitch about a definition without even asking why it was stated that way and without adding anything to the discussion itself at all then yeah, that's what I call being a pedantic twit.

If you had said the values are this and they have fairly tight CIs I would have had no problem at all. It's the fact that you seemed (and perhaps still do for all I know) to think that the fact that these CIs were 90 or 95% had any meaning. Nor does it mean that the 'scientific community is pretty confident in their estimations' only the actual CI values can tell you that. Is that not correct?

Deepcrush wrote:

Wow, a ban over this is like using a nuke because two guys bumped and dropped their beers.

Yes, thanks. You are very helpful.

IanKennedy wrote:

Deepcrush wrote:

Wow, a ban over this is like using a nuke because two guys bumped and dropped their beers.

Yes, thanks. You are very helpful.

Hey, I do my part. Besides, you two were the ones that got stupid. I just had a little fun with it.

IanKennedy wrote:Stating the % of the CI without actually giving the values is totally meaningless. Really, utterly meaningless. For example I could say that the mean was 1,000,000 and the 95% CI was 800,000 to 1,200,000. But equally it could be 1,000,000 with a 90% CI of -1,000,000 to 100,000,000. You can see from the first example that they have a reasonable degree of accuracy. The second one they've no idea at all what the value is. I say again if you don't specify the values you are grossly misleading people. The reason I pointed out the error in the first place is that not everyone here knows stats that well, and you where using the 90/95% figure to try and bolster peoples opinions of your numbers. Giving the percentage is utterly useless, it tells you nothing. The percentage used is completely arbitrary, typical convention is to use 90 or 95%. You either give the range or give nothing there's no in between.

Yes, I generalized and cut it down. I admit that and that it was not the proper use of the stat. I was in a hurry and didn't feel like explaining it all. As for the accuracy of the numbers, yes I was using them to bolster confidence in my numbers. I did not fail to include the CI's in an intentional attempt to mislead however. In the case of the sperm whales the 90% CI was 1,950,000 to 2,450,000 whales. The minke's the 95% CI was I believe about 590,000 to 740,000 whales. So yes, the population estimates are quite sound. Did I use the CI incorrectly? Yes, and I admit that, I was in a hurry. Was I trying to cover up very weak numbers? No, I wasn't. My whole goal in this has been to try and bring some actual facts to light in what typically appears to be a purely emotional argument. In this case I cut a corner.

As for threatening to ban you I didn't I merely pointed out that there is a rule here for not being jerk. You where being a jerk. After all this forum is a private area owned by Graham and myself. Membership here is privilege not a right. As the owners yes I do expect a little respect from the members, just as you would from visitors to your home. Is that unreasonable? Do I expect you to knuckle under, No, do I expect you to not be a jerk, emphatically yes. We have already banned one person for that and I would rather not have to ban another.

Fair enough, don't think that I don't appreciate the board I really do. However my reaction to being told that I "don't know shit" about what I'm talking about is never going to be a good one. Probably about as well as you reacted to being called a "pedantic twit," which is rather mild on the scale of being a jerk I have seen slung around in more heated arguments here.

IanKennedy wrote:If you had said the values are this and they have fairly tight CIs I would have had no problem at all. It's the fact that you seemed (and perhaps still do for all I know) to think that the fact that these CIs were 90 or 95% had any meaning. Nor does it mean that the 'scientific community is pretty confident in their estimations' only the actual CI values can tell you that. Is that not correct?

Ian, I am very well aware of how a CI works and that you need the range and the interval to mean something. I always have. I freely admit to cutting a corner in this regard but it doesn't mean I don't know what I'm talking about. Two of the intervals I can recall are already posted. The other's were trending about the same. At no point was I hiding weak data or was I even wrong in my statement "scientific community is pretty confident in their estimations" I simply failed to post all the supporting evidence.

Tyyr wrote:

IanKennedy wrote:Stating the % of the CI without actually giving the values is totally meaningless. Really, utterly meaningless. For example I could say that the mean was 1,000,000 and the 95% CI was 800,000 to 1,200,000. But equally it could be 1,000,000 with a 90% CI of -1,000,000 to 100,000,000. You can see from the first example that they have a reasonable degree of accuracy. The second one they've no idea at all what the value is. I say again if you don't specify the values you are grossly misleading people. The reason I pointed out the error in the first place is that not everyone here knows stats that well, and you where using the 90/95% figure to try and bolster peoples opinions of your numbers. Giving the percentage is utterly useless, it tells you nothing. The percentage used is completely arbitrary, typical convention is to use 90 or 95%. You either give the range or give nothing there's no in between.

Yes, I generalized and cut it down. I admit that and that it was not the proper use of the stat. I was in a hurry and didn't feel like explaining it all. As for the accuracy of the numbers, yes I was using them to bolster confidence in my numbers. I did not fail to include the CI's in an intentional attempt to mislead however. In the case of the sperm whales the 90% CI was 1,950,000 to 2,450,000 whales. The minke's the 95% CI was I believe about 590,000 to 740,000 whales. So yes, the population estimates are quite sound. Did I use the CI incorrectly? Yes, and I admit that, I was in a hurry. Was I trying to cover up very weak numbers? No, I wasn't. My whole goal in this has been to try and bring some actual facts to light in what typically appears to be a purely emotional argument. In this case I cut a corner.

As for threatening to ban you I didn't I merely pointed out that there is a rule here for not being jerk. You where being a jerk. After all this forum is a private area owned by Graham and myself. Membership here is privilege not a right. As the owners yes I do expect a little respect from the members, just as you would from visitors to your home. Is that unreasonable? Do I expect you to knuckle under, No, do I expect you to not be a jerk, emphatically yes. We have already banned one person for that and I would rather not have to ban another.

Fair enough, don't think that I don't appreciate the board I really do. However my reaction to being told that I "don't know s**t" about what I'm talking about is never going to be a good one. Probably about as well as you reacted to being called a "pedantic twit," which is rather mild on the scale of being a jerk I have seen slung around in more heated arguments here.

Fine, we'll leave it at that then.

IanKennedy wrote:If you had said the values are this and they have fairly tight CIs I would have had no problem at all. It's the fact that you seemed (and perhaps still do for all I know) to think that the fact that these CIs were 90 or 95% had any meaning. Nor does it mean that the 'scientific community is pretty confident in their estimations' only the actual CI values can tell you that. Is that not correct?

Ian, I am very well aware of how a CI works and that you need the range and the interval to mean something. I always have. I freely admit to cutting a corner in this regard but it doesn't mean I don't know what I'm talking about. Two of the intervals I can recall are already posted. The other's were trending about the same. At no point was I hiding weak data or was I even wrong in my statement "scientific community is pretty confident in their estimations" I simply failed to post all the supporting evidence.

I didn't care a thing about the actual numbers or the question at hand. If you check you will see I wasn't active in the original topic in the slightest. I merely found it highly dubious to put your statements together. If had the smell of someone trying to pull the wool over others eyes by using terms like confidence interval, and on top of that completely failing to use the term with any meaningful way. That I do always take offense at. If that was not your intent then I will take your word for it. As for hiding weak data my main problem was we had no way of knowing if that was the case, until now. I would have had no problem if you had simple said 'and the 95% confidence intervals are fairly tight', this isn't a scientific paper after all.