Beesource Beekeeping Forums banner

What bee logical fallacies do you hear the most?

2356 58
I think I hear the "correlation does not imply causation" one. It might be why people put extra stuff in hives.
  • Like
Reactions: Oldtimer
1 - 20 of 59 Posts

·
Registered
Joined
·
3,019 Posts
I don't think I have heard that the expression as "Correlation does not imply causation."

I always hear it as "Correlation does not mean causation."
If it'll help ...

Cum hoc ergo propter hoc - 'with this, therefore because of this'.

The phrase "correlation does not imply causation" refers to the inability to legitimately deduce a cause-and-effect relationship between two events or variables solely on the basis of an observed association or correlation between them. The idea that "correlation implies causation" is an example of a questionable-cause logical fallacy, in which two events occurring together are taken to have established a cause-and-effect relationship.
LJ
 

·
Registered
Joined
·
3,019 Posts
Bees prefer to live in trees (with thick walls, insulation++ etc) in cavities high up off the ground - never proven.

Horizontal hives are 'unnatural' (based on those trees again).

Bees prefer vertical colony expansion/contraction to horizontal expansion/contraction.

Swarms prefer 40 litre cavities - never demonstrated, let alone proven.

Bees (intentionally) heat the box, not just the cluster ... :)

There is such a thing as The Perfect Beehive.

Lorenzo Langstroth designed the modern-day 'Langstroth' Hive.

Georges De Layens designed the 'Layens' Hive.

I'm sure there are others ...
LJ
 

·
Registered
Joined
·
1,889 Posts
I got this from Gray Haun (some of you may of heard of him) back in 1987 when we were discussing swarm traps img20221109_08440168.pdf .
 

Attachments

  • Helpful
Reactions: Mandrex

·
Registered
Joined
·
765 Posts
If it'll help ...

Cum hoc ergo propter hoc - 'with this, therefore because of this'.

The phrase "correlation does not imply causation" refers to the inability to legitimately deduce a cause-and-effect relationship between two events or variables solely on the basis of an observed association or correlation between them. The idea that "correlation implies causation" is an example of a questionable-cause logical fallacy, in which two events occurring together are taken to have established a cause-and-effect relationship.
LJ
I think it depends on the understanding of "implies".
If implies is understood to mean "suggests the possibility of", then It is true that correlation - especially strong correlation - implies causation.

For example, in large observational studies there is a robust correlation between ibuprofen use and cardiac events. This is generally accepted to imply causation - that Ibuprofen causes cardiac events. However, this has not been demonstrated by random controlled trials. It is also possible that use of ibuprofen is a marker for inflammation, as it is an anti-inflammatory. If that were the case, then use of ibuprofen and cardiac events could have a common cause - systemic inflammation, possibly related to diet or lifestyle factors.

Even random controlled trials at best only strongly imply causation because the mechanism of causation is not demonstrated by merely statistical means.

However, whenever we see a correlation, we immediately are faced with the question - is this correlation due to causation.

An simple example is: The number of medical doctors in a community correlates directly to the number of murders in a community. This is a very robust correlation. This implies that Medical doctors cause murders, or that murders cause medical doctors. We respond to that implication by rejecting it, as it appears illogical. We look for another explanation. Most murders happen in large cities. most medical centers are located in large cities. We conclude this is a coincidence. We reject the implication. Thus the implication exists.

It would be far less ambiguous to state correlation does not demonstrate causation, or correlation does not prove causation. Those statements are clearly true. Causation requires a causative relationship. Absent a plausible cause and effect narrative, we tend to reject the implication of causation.

It is also true that with a strong enough cause and effect relationship, we often accept a hypothesis without evidence of correlation. For example, it is well known that exposure to radon causes a large number of cases of lung cancer, and that risk of developing lung cancer is dose related. However, no observational studies in the general population have been able to demonstrate a correlation between residential radon exposure and lung cancer rates. You have to actually look at the details of the studies to see that, as publishing a conclusion that there is no correlation (or a weak correlation in the opposite direction) is not a career move as it is contrary to accepted science. This is consistent over studies by government and academia in Europe and north America. Does that mean causation doesn't exist? No. It means existing studies are unable to demonstrate it, or it doesn't exist. Science chooses the former. Not wanting to debate Radon. It probably isn't good for you, and if it is, it isn't very good for you. Just talking about the limits of observational science and the way consensus is formed.

However if you parse the word "implies" differently, you might come to a different conclusion.
 

·
Registered
Joined
·
190 Posts
I don't think I have heard that the expression as "Correlation does not imply causation."

I always hear it as "Correlation does not mean causation."
What I've heard is "Correlation does not EQUAL causation."

To "imply" is only to suggest. "Mean" is certainly closer to equal.

To the topic at hand, the thought that my bees recognize me.
 

·
Registered
Joined
·
1,889 Posts
hot bees will have you steppin' an a fetchin' like your heads on fire and your arse is a catchin'

Is this what we are going for here?? I am not real sure about the topic, above my pay grade.
 

·
Registered
Joined
·
3,019 Posts
What I've heard is "Correlation does not EQUAL causation."
To "imply" is only to suggest. "Mean" is certainly closer to equal.
If we remove the dismissive word 'only', then we have: "To imply is to suggest" - which I think is exactly the problem we are faced with. Take Seeley's 'Nest of the Honey Bee' paper as an example - as this is a totally observational study, it's contents can only ever be 100% correlation - and yet so many people cite that study which does indeed 'suggest' cause and effect relationships. It is precisely the seductive nature of such suggestion (i.e. the psychological mechanism) which lies at the heart of this extremely common logical mistake:
LJ
 

·
Registered
Joined
·
3,019 Posts
Hi Graham - yes, I'm familiar with that paper. It's interesting enough, and based on one of Seeley's observational studies together with an experiment he conducted on a treeless island which was intended to ascertain the ideal volume for a bait hive. Unfortunately, that experiment was flawed.
The idea was to offer swarms a choice of 3 box volumes. The first swarm which issued chose neither of them, but instead chose to set up home in the chimney of the only dwelling on that island. It was only after access to the chimney was denied to the bees that they began to choose the experimental boxes instead.

More importantly, as Seeley was conducting a classic single-variable (i.e. box volume) experiment, all other variables were held constant. So - the colonies were raised in standard 40L boxes, transported to the island in 40L boxes, and in due course swarms then issued from the same 40L boxes. Seeley then analysed the data and concluded that swarms have a preference for 40L boxes.

Had he concluded that colonies raised and transported within 40L boxes, from which swarms of broadly similar size issue(*) and then go on to preferentially select the same size boxes, then I would voice no criticism - but in the real, non-experimental world, swarms issue from all types and sizes of cavities, and thus their population sizes will likewise vary considerably.

Since Seeley's experiment, it has been firmly established that approx. 75% of a colony's bees leave as a prime swarm - thus my suggestion is that such swarms will tend to choose a cavity the size of which approximates to that from which they left - i.e. one which is not overly large, but still has room for a modest expansion.
Only last year I had one swarm arrive (on the ground !) which I initially housed within a single standard brood box (circa 40L), only to observe that a very large number of the bees were unable to fit within it - the swarm promptly left the following day.

Seeley's explanation for such large swarms is that multiple swarms probably combine when on the wing - but an alternative explanation is that the size of swarms is directly proportional to the colony size, which in turn is largely determined by the volume of the box housing that colony. So - with hindsight, the above event would suggest that the swarm had most probably issued from a double-brood-box beehive.

Within the cited paper, Seeley appears to be associating the high success rate observed (c.80%) with the height at which those swarm boxes were located, but unless a choice of box heights had been offered, one cannot reasonably infer that any preference had actually been expressed, with the one influencing the other - and thus to highlight what after all is only an observation (i.e. a correlation rather than causation), is really rather misleading.
LJ

(*) In the legend to Fig.2, the author writes: "We found no relation between the population of the swarm and the size box selected."
Well - I can see a possible relationship. :) When the 40L option was removed, "some swarms accepted the large box (100L), but none accepted the 10L box." That observation supports my 'hypothesis' that swarms issuing from the 40L boxes were most probably similar in size, such that the 100L box would have been reluctantly accepted (as being too large - but with no other option available), whereas the 10L box would simply have been far too small and was thus always rejected.
 

·
Super Moderator
North Georgia USA Zone 7b Mutts and Carniolans in Langs.
Joined
·
844 Posts
When the 40L option was removed, "some swarms accepted the large box (100L), but none accepted the 10L box." That observation supports my 'hypothesis' that swarms issuing from the 40L boxes were most probably similar in size, such that the 100L box would have been reluctantly accepted (as being too large - but with no other option available),
May 11th had a swarm (presumably my own) find a way into a stored stack of three deeps. Can I infer that no more suitable option was available? Or is that too a logical fallacie ;)

Fun read for any who have never seen this site,Spurious Correlations
 

·
Registered
Joined
·
1,889 Posts
So when bees have a cavity of an almost unlimited size, what kind of hive did they come from?
This hive was in the ceiling of the balcony on the second floor, did the bees come from a hive that was 8 feet off the floor or 18 feet off the ground?

Wood Building Tints and shades Natural material Beam
Wood Natural material Pattern Metal Rock
 

·
Registered
5 ,8 ,10 frame, and long Lang
Joined
·
5,545 Posts
So when bees have a cavity of an almost unlimited size, what kind of hive did they come from?
This hive was in the ceiling of the balcony on the second floor, did the bees come from a hive that was 8 feet off the floor or 18 feet off the ground?

View attachment 71723 View attachment 71724
looking at that wavy comb I'd say they came from a beach house.
12 combs so 12 feet up is my guess.

:)

GG

one is straight and one wavy was it 2 different colonies?
 
1 - 20 of 59 Posts
Top