Great White Shark Attack Paddle vs Pedal Numbers - Kayak Fishing Adventures on Big Water’s Edge

pbb · 10-08-2014, 11:27 AM

As someone with a stats background (Ph.D. in Sociology with a quantitative focus, M.A. in Demography), I would agree that there is nothing here that suggests any conclusive "statistically significant" differences by propulsion method.

The OP is correct that you would really need to know how many people are out there in these types of kayaks at risk of getting attacked to do the analysis correctly.

In fact, you would ideally need to know not only what kind of kayaks are out there in consumers' hands, but how much they use them. In other words, if you had 1000 people with Hobie mirage kayaks and 1000 people with paddle kayaks, and the mirage drive kayakers spend twice as much time on the water as people with paddle kayaks, you would expect them to experience twice as the number of shark attacks even if the rate of shark attacks was exactly the same.

PAL · 10-08-2014, 11:57 AM

Quote:

Originally Posted by pbb

As someone with a stats background (Ph.D. in Sociology with a quantitative focus, M.A. in Demography), I would agree that there is nothing here that suggests any conclusive "statistically significant" differences by propulsion method.

The OP is correct that you would really need to know how many people are out there in these types of kayaks at risk of getting attacked to do the analysis correctly.

In fact, you would ideally need to know not only what kind of kayaks are out there in consumers' hands, but how much they use them. In other words, if you had 1000 people with Hobie mirage kayaks and 1000 people with paddle kayaks, and the mirage drive kayakers spend twice as much time on the water as people with paddle kayaks, you would expect them to experience twice as the number of shark attacks even if the rate of shark attacks was exactly the same.

Great stuff! I'd like to quote you in an update to the original story. Please PM your name, and add any other statistical observations you'd care to share. Maybe something about the relatively small sample size?

GregAndrew · 10-08-2014, 04:54 PM

Quote:

Originally Posted by PAL

Great stuff! I'd like to quote you in an update to the original story. Please PM your name, and add any other statistical observations you'd care to share. Maybe something about the relatively small sample size?

I said essentially the same thing as a comment on the page the article appears on. It is the total number of hours spent in the "Zone" by the different types of kayaks divided by the number of attacks on each. That will give you the attacks per, probably, hundreds of thousands of hours in the "Zone".

Sample size is not a term that applies here. What you have is a small number of occurrences. Sample size is the number of data points that you survey in order to make a prediction about a population. Like the number of Hobie owners you question in order to find out if they have been attacked by a shark.

PAL · 10-08-2014, 05:05 PM

Quote:

Originally Posted by GregAndrew

I said essentially the same thing as a comment on the page the article appears on. It is the total number of hours spent in the "Zone" by the different types of kayaks divided by the number of attacks on each. That will give you the attacks per, probably, hundreds of thousands of hours in the "Zone".

Sample size is not a term that applies here. What you have is a small number of occurrences. Sample size is the number of data points that you survey in order to make a prediction about a population. Like the number of Hobie owners you question in order to find out if they have been attacked by a shark.

I hear you, but who has the time (or the data) for that?

That is likely why the risk of suffering a shark attack is typically expressed in terms of the overall population. For instance, you have a better chance of getting killed by a falling coconut than you do a shark... blah, blah, blah. That's no help other than in simplistically illustrating the low overall occurrence of death by shark.

Someone who never dips a toe in the ocean has no chance of succumbing, not even if they are Austin Powers (just maybe if the sharks have fricken laser beams).

Life is one big judgement call.

GregAndrew · 10-08-2014, 05:38 PM

Quote:

Originally Posted by PAL

I hear you, but who has the time (or the data) for that?

That is likely why the risk of suffering a shark attack is typically expressed in terms of the overall population. For instance, you have a better chance of getting killed by a falling coconut than you do a shark... blah, blah, blah. That's no help other than in simplistically illustrating the low overall occurrence of death by shark.

Someone who never dips a toe in the ocean has no chance of succumbing, not even if they are Austin Powers (just maybe if the sharks have fricken laser beams).

Life is one big judgement call.

Sharks don't have laser beams? Sorry, my pet peeve is people using data as statistics, especially politicians and environmentalists.

PAL · 10-09-2014, 07:34 AM

NCKA poster Northern Boy posted this. I think he's right. If all attacks had involved Hobies or none had, the trend would be convincing:

Quote:

I'm not so sure about that. Statistics is an exact science, but the design and interpretation of statistical tests is not.

There are a lot of factors to be considered in the latter - the importance of those depends on the strength of the effect - to give a simplistic example - if the 12 attacks had all been on Hobies, then I don't think anyone would be worrying about whether Hobies spend more time on the water or what color they were.

However, as it is, the descriptive statistics available are indeed pretty muddled (I had not seen those before, they are very informative) and thus the small sample size is a problem, as are the other, potentially confounding, factors like the number of Hobies vs paddle etc. Hopefully not a 'problem' that will be fixed anytime soon.

NB

pbb · 10-09-2014, 03:30 PM

I agree you have a limited number of occurrences. Most likely it is the full universe of occurrences (at least for recent years). And this is different from the sample size. I'm working on a study of occupational injuries where about 13000 people were screened for (asked about) occupational injuries and less than 200 such injuries were identified. You sample size in that case is the 13000.

Here again, we really don't know how many hours people are spending on the ocean, so we have neither a sample nor a denominator for a rate of attacks.

But given that you have a similar number of occurrences, you would have to believe that the hours spent at risk (i.e., on the ocean) by the two groups was very different to conclude that you might have a higher risk for one group.

My sense (and here we depart from anything statistical or scientific) is that there aren't orders of magnitude lower person hours spent in Hobie Mirage kayaks on the open ocean compared to paddle kayaks, hence I don't see anything in the data that I would find to be convincing that there is a difference.

It wouldn't be scientific, but you could try to get a sense by creating a "poll" on various California kayak fishing forums about how many hours users fish or kayak in the ocean during an average month in mirage drive or paddle kayaks. That might give some more information, but it wouldn't be conclusive since it would be a convenience sample rather than a probability sample of kayak anglers. But I can't think of any feasible way of doing the latter, except possibly asking DFG to add that to their CRFS survey when they survey kayakers. (I guess you could ask DFG to let you do a mail or email survey to everyone with a fishing license and and ocean stamp, if you had a lot of money to burn).

GregAndrew · 10-08-2014, 05:07 PM

More than the color of the kayak, I would like to know the visibility of the water at the surface in each instance. If it is generally less than say 15' (at the speeds they are probably traveling) then you might infer that the sharks are attacking a silhouette more than anything else. That could also cloud the data if the sharks can't tell until last second that a Hobie has fins. However, if the sharks are seen before the attack or the water is generally very clear, then all bets are off on that.

10-08-2014, 11:27 AM	#1
pbb Senior Member Join Date: Jul 2013 Location: San Diego Posts: 149	As someone with a stats background (Ph.D. in Sociology with a quantitative focus, M.A. in Demography), I would agree that there is nothing here that suggests any conclusive "statistically significant" differences by propulsion method. The OP is correct that you would really need to know how many people are out there in these types of kayaks at risk of getting attacked to do the analysis correctly. In fact, you would ideally need to know not only what kind of kayaks are out there in consumers' hands, but how much they use them. In other words, if you had 1000 people with Hobie mirage kayaks and 1000 people with paddle kayaks, and the mirage drive kayakers spend twice as much time on the water as people with paddle kayaks, you would expect them to experience twice as the number of shark attacks even if the rate of shark attacks was exactly the same. __________________ 2008 Olive Hobie Revolution 13

10-09-2014, 03:30 PM	#7
pbb Senior Member Join Date: Jul 2013 Location: San Diego Posts: 149	I agree you have a limited number of occurrences. Most likely it is the full universe of occurrences (at least for recent years). And this is different from the sample size. I'm working on a study of occupational injuries where about 13000 people were screened for (asked about) occupational injuries and less than 200 such injuries were identified. You sample size in that case is the 13000. Here again, we really don't know how many hours people are spending on the ocean, so we have neither a sample nor a denominator for a rate of attacks. But given that you have a similar number of occurrences, you would have to believe that the hours spent at risk (i.e., on the ocean) by the two groups was very different to conclude that you might have a higher risk for one group. My sense (and here we depart from anything statistical or scientific) is that there aren't orders of magnitude lower person hours spent in Hobie Mirage kayaks on the open ocean compared to paddle kayaks, hence I don't see anything in the data that I would find to be convincing that there is a difference. It wouldn't be scientific, but you could try to get a sense by creating a "poll" on various California kayak fishing forums about how many hours users fish or kayak in the ocean during an average month in mirage drive or paddle kayaks. That might give some more information, but it wouldn't be conclusive since it would be a convenience sample rather than a probability sample of kayak anglers. But I can't think of any feasible way of doing the latter, except possibly asking DFG to add that to their CRFS survey when they survey kayakers. (I guess you could ask DFG to let you do a mail or email survey to everyone with a fishing license and and ocean stamp, if you had a lot of money to burn). __________________ 2008 Olive Hobie Revolution 13

10-08-2014, 05:07 PM	#8
GregAndrew Senior Member Join Date: Mar 2009 Posts: 2,384	More than the color of the kayak, I would like to know the visibility of the water at the surface in each instance. If it is generally less than say 15' (at the speeds they are probably traveling) then you might infer that the sharks are attacking a silhouette more than anything else. That could also cloud the data if the sharks can't tell until last second that a Hobie has fins. However, if the sharks are seen before the attack or the water is generally very clear, then all bets are off on that.